1 Thursday, 27 October 2022 2 (10.17 am) 3 MR BLAKE: Good morning, sir. 4 SIR WYN WILLIAMS: Good morning. 5 MR BLAKE: Today's first witness is Terence Austin. 6 SIR WYN WILLIAMS: Yes. 7 TERENCE AUSTIN (affirmed) 8 Questioned by MR BLAKE 9 MR BLAKE: Thank you very much. Could you give your full 10 name, please. 11 A. Terence Paul Austin. 12 Q. Mr Austin, in front of you should have a witness 13 statement? 14 A. Yes, I do. 15 Q. Could I ask you to turn to the final page, that's 16 page 19. Can you confirm that that is your signature? 17 A. Yes, it is. 18 Q. It's dated 13 September of this year; is that right? 19 A. That's correct. 20 Q. Is that statement true to the best of your knowledge and 21 belief? 22 A. It is. 23 Q. Thank you, Mr Austin. 24 For the purpose of the transcript the statement is 25 WITN04190100. Mr Austin, that statement and its 1 1 exhibits will go into evidence, so I'm not going to 2 repeat lots of matters that are contained therein. 3 We're going to spend about half a day going through 4 various matters. I'm going to start with your 5 background. 6 You spent 40 years in the IT sector; is that right? 7 A. That's correct. 8 Q. You joined ICL as a programme director in 1995? 9 A. Correct. 10 Q. You were employed specifically to deliver the benefit 11 card and the Electronic Point of Sale System; is that 12 right? 13 A. The whole solution with regard to the ITT. 14 Q. So I think you said you were responsible for delivering 15 the IT Pathways solution; is that right? 16 A. Yes, correct. 17 Q. You became systems director at some point. 18 A. Yes, there was a period of time, several months after we 19 started, after award of contract, where it was obvious 20 that the requirement was expanding tremendously and it 21 was felt that we should reorganise the team so that 22 I would then take responsibility for the solution and 23 that one of my colleagues would come in and take over 24 the position as a programme director, so I relinquished 25 areas such as implementation, training, business 2 1 requirements, et cetera, so that they went elsewhere 2 within the ICL Pathway organisation. 3 Q. You left in October 2000? 4 A. I did. 5 Q. Would it be right to say that you were the senior 6 technical director of the Horizon programme? 7 A. No, it wouldn't. 8 Q. Were you the most senior individual who had 9 responsibility for the IT Pathway solution within ICL? 10 A. Yes. 11 Q. Were you part of the management structure? 12 A. I was. 13 Q. Were you happy with how ICL was being managed at the 14 time? 15 A. Yes. I was recruited specifically for the ICL Pathway 16 project. 17 Q. Did you have faith in its internal controls and internal 18 processes as a company? 19 A. I was not familiar with the internal processes because 20 I was brought on specifically to look at the ITT for the 21 ICL Pathway programme. 22 Q. But as your career developed, did you have faith in 23 the -- 24 A. Within the ICL Pathway? 25 Q. Yes. 3 1 A. Yes. 2 Q. Did you think it was a good company at scrutinising 3 itself? 4 A. Yes. 5 Q. Did it have an effective audit process, for example? 6 A. ICL itself? 7 Q. Yes. 8 A. I wasn't aware of it, but I assume there was. We 9 introduced our own within ICL Pathway. 10 Q. So within ICL Pathway, were you happy with the audit 11 process? 12 A. Yes. 13 Q. Can we look at FUJ00080690. This is the EPOSS PinICL 14 task force report. You can ignore, I think, the date on 15 the top right-hand corner because the EPOSS PinICL task 16 force was, as it says on the front page, in place 17 between 19 August and 18 September 1998. Do you 18 remember this document? 19 A. I do. 20 Q. Do you remember seeing the document at the time? 21 A. I do. 22 Q. It has your name on the distribution. Can you tell us 23 about the authors -- the author, if we just look below 24 that, it has J Holmes and D McDonnell. J Holmes is, 25 I think, Jan Holmes, I think we have been pronouncing it 4 1 Jan but it is Jan Holmes? 2 A. It is. 3 Q. He was the audit manager; is that right? 4 A. He was. 5 Q. Who was D McDonnell? 6 A. That was another person within the audit team, as far as 7 I can recall. 8 Q. Do you think he was the deputy development manager? Is 9 that a title you remember? 10 A. I remember the title, but I don't remember the name, I'm 11 afraid. 12 Q. From what you have said about the audit process at 13 ICL Pathway, presumably you had faith in them in 14 carrying out that task? 15 A. Well, I was the instigator of the task force. 16 Q. Yes, so you appointed the right individuals to carry out 17 that task? 18 A. Yes, which was Jan. 19 Q. Of drafting the report. 20 A. Yes. 21 Q. Did you have any reason not to believe what they said in 22 the report? 23 A. The reason I hesitate is that auditors are not technical 24 people. They interview people and they draw conclusions 25 and they then summarise those conclusions and their 5 1 recommendations. 2 Q. But I think Jan Holmes was the auditor but Mr McDonnell 3 was -- 4 A. A deputy development -- 5 Q. -- a technical man. Can we look at page 4 of that 6 document, please. This is the introduction and I'm 7 going to just read from the second paragraph for the 8 purpose of the transcript. It says: 9 "This report presents the outcome of the Task Force 10 activity and identifies factors which prevented the 11 original target (zero or near to zero residual PinICLs) 12 being met." 13 Now, PinICLs are error logs or -- 14 A. Defects, yes. 15 Q. "During the course of the Task Force it became clear 16 that there are significant deficiencies in the EPOSS 17 product, its code and design, and these are also 18 presented in this report. Finally the report contains 19 recommendations from the authors which we believe should 20 be implemented by the programme to address the 21 shortcomings identified." 22 So that's the very first section of this 20-page 23 document and it is highlighting there that there were, 24 in their view, significant deficiencies in the EPOSS 25 product, its code and design; do you agree with that? 6 1 A. Absolutely, yes. 2 Q. Can we turn over the page to page 5, please, and the top 3 half of that page. In summary, it says there that they 4 had spent five weeks trying to get the PinICLs down to 5 zero and it is the second paragraph there: 6 "The position at 1300 hours on 18th September is 7 that 166 PinICLs have been fixed and closed and 165 8 remain in WIP." 9 Is that "work in progress"? 10 A. Yes. 11 Q. "This indicates that the Task Force has failed to meet 12 its primary objective." 13 So they closed 166, but 155 (sic) remained, 14 presumably that indicates a significant remaining 15 problem, despite the task force having -- 16 A. It does. 17 Q. -- closed a number of PinICLs? 18 A. It does. 19 Q. Moving over to page 7 of this report, there is a section 20 on EPOSS code and, again, for the purpose of the 21 transcript, I appreciate it's a relatively long passage, 22 but I'm going to read it, it says: 23 "It is clear that senior members of the Task Force 24 are extremely concerned about the quality of code in the 25 EPOSS product. Earlier this year the EPOSS code was 7 1 re-engineered by Escher and the expectation is that the 2 work carried out in Boston was to a high standard and of 3 good quality. Since then many hundreds of PinICL fixes 4 have been applied to the code and the fear is that code 5 decay will, assuming it hasn't already, cause the 6 product to become unstable. This present [I think that 7 means 'presents'] a situation where there is no 8 guarantee that a PinICL fix or additional functionality 9 can be made without adversely affect [I think it means 10 'affecting'] another part of the system. 11 "However, a more worrying concern from the 12 Programme's perspective should be the reliance on the 13 EPOSS product in its current state as a basis for 14 planning and delivery. During the Task Force there was 15 relatively little testing that directly impacted EPOSS 16 and yet [over] 200 PinICLs, roughly 50 per week, were 17 raised. Immediately following the conclusion of the 18 Task Force it is intended to re-run System Test Main 19 Pass and various other test streams. While I am 20 confident that the fixes delivered by the Task Force 21 will prove to be reliable, I fully expect the PinICL 22 rate to increase as further testing is carried out." 23 Now, concerns that were expressed there include that 24 the code will decay; do you remember that? 25 A. I do. 8 1 Q. That's, I think, a term that's used to describe where 2 there's a reduction in reliability and effectiveness -- 3 A. That's what it is implying, yes. 4 Q. -- over time and that could, for example, affect things 5 like the cash and counter performance, potentially. 6 A. Potentially, yes. 7 Q. Can we look at page 17 of the same document please, 8 paragraph 7.3, "Existing Code": 9 "NB: This section has been produced with the 10 assistance of Dave McDonnell and Martin Smith and their 11 combined experience of structured programming." 12 Do you remember who Martin Smith was at all? 13 A. He would be one of the team, one of the programmers. 14 Q. So, again, two technical experts? 15 A. Yes. 16 Q. They say: 17 "Although parts of the EPOSS code are well written, 18 significant sections are a combination of poor technical 19 design, bad programming and ill thought out bug fixes. 20 The negative impact of these factors will continue and 21 spread as long as the PinICL fixing culture continues. 22 This is partly due to the nature/size of the bug-fixing 23 task and partly due to the quality and professionalism 24 of certain individuals within the team." 25 Then over the page: 9 1 "Whoever wrote this code clearly has no 2 understanding of elementary mathematics or the most 3 basic rules of programming." 4 That's all pretty damning, isn't it? 5 A. Very much so. 6 Q. Presumably, that would have been quite well-known 7 amongst the team at the time? 8 A. Maybe amongst the team but not amongst the management 9 part of the team. 10 Q. So I think if we turn to the first page of this document 11 it shows the distribution. There is yourself there, 12 Mr Bennett, Mr McDonnell and then it says "Library". Do 13 you know what that was a reference to? 14 A. It will be stored in the library of the audit team, 15 I believe. 16 Q. Would the information from this report have been -- you 17 say it was known to the team. Who would have known 18 about -- 19 A. Well, the reason for me asking for the task force in the 20 first place was that I wasn't very happy with the way 21 the product was, so I felt let down by the people who 22 had developed it, who were supposed to be experts in 23 their field. So because I was getting reports that the 24 product was not stable and that it was not behaving in 25 a way that we would expect, I called the task force. 10 1 So I called the task force together and brought in 2 the people who I thought were the most competent people 3 within my team to be able to look into this and see what 4 was going on, and it was as a result of that that Jan 5 writes a report at the end of their activity and when 6 I received this report, which was, as you quite rightly 7 say, "very damning", and a massive worry to me and the 8 rest of ICL Pathway, we had to take every part of it 9 extremely seriously. 10 So we took -- I called in all the members of the 11 task force and my lead designer and we went through 12 every element of it to find out what we could do and 13 what options we had to do. 14 We identified who the people were that were 15 responsible for producing the product in its particular 16 state and they were removed from the team, so that we 17 were starting off with a different team that was looking 18 at it and designing it and managing it. 19 Q. Do you remember who was removed from the team at all? 20 A. No, I'm sorry, I can't. I genuinely can't. 21 Q. So this was 1998? 22 A. Yes. 23 Q. We then have another report in 1999 that I would like to 24 take you to, that's -- 25 A. First of all -- sorry to interrupt, I would just like to 11 1 say that, as a result of the various meetings following 2 this, we went into a corrective action plan. So we 3 decided what actions we needed to take in order to get 4 the product into the state that we were happy with. 5 Q. So let's look what happened a year later, 1999, and that 6 is FUJ00079782. This is the "CSR+ Development Audit". 7 Very briefly, can you tell us what CSR+ was? 8 A. Core System Release, I believe. 9 Q. That was also conducted by Jan Holmes, if we scroll down 10 a little bit; so the audit manager conducted this audit. 11 A. Yes. 12 Q. Can we look at page 6, please, and scroll down to the 13 bottom of the page. There's a section on "Audit 14 Conduct" and how the audit was conducted. It explains 15 there that there were some 35 interviews over 16 a four-week period between September and October 1999, 17 so quite a comprehensive audit? 18 A. Absolutely, yes. 19 Q. As I said, this is a year on from the Holmes/McDonnell 20 report that we just saw. 21 A. But this is an audit of the entire release, so it's the 22 entire product, not just the EPOSS product. 23 Q. Yes. Can we look at page 7, please, and 2.5. This 24 addresses "Next Steps", and it says there, in the first 25 paragraph, that there was an opportunity during the week 12 1 of 8 November to challenge any recommendations that were 2 made in that report. 3 It is page 19 -- as you say, there are plenty of 4 pages that address other aspects of CSR+, and it is 5 page 19 that addresses EPOSS. Again, I'm going to read 6 a fair bit out, I'm afraid. It says: 7 "From the CSR+ perspective the development of the 8 EPOSS product has been successful with software drops 9 being made according to planned schedules and confidence 10 in the team that future drops will also be achieved on 11 time. 12 "Unfortunately EPOSS continues to be resource hungry 13 in dealing with live problems associated with CSR and in 14 ensuring that these fixes are brought forward and 15 incorporated into the CSR+ product. 16 "The EPOSS task force report [which we have already 17 seen] raised the question of the maintainability and 18 resilience of the EPOSS code following the 6 week PinICL 19 blitz where some 550 PinICLs were processed. Since then 20 [so a year later] a further [approximately] 996 PinICLs 21 have been raised -- using the 'Product = EPOSS and 22 Target Release = IR-CSR or PDR-CSR' search criteria -- 23 and these can only have had a detrimental effect on the 24 quality of the code. In particular the maintainability, 25 resilience and potential for change aspects must be 13 1 subject to doubt. The report also identified many 2 instances of poor programming technique and application 3 of coding standards and while CSR+ changes have been 4 reviewed by the Team Leader, no attempts have been made 5 to address the significant body of code not affected. 6 There is also anecdotal evidence that EPOSS components 7 used by other applications are fragile and cause 8 problems for the calling application, Print Server was 9 mentioned by both LFS and APS Counter teams." 10 So "resource hungry", it sounds as though it 11 required a lot of attention; is that right? 12 A. That's correct. 13 Q. It has been -- well, there are 996 more PinICLs that 14 have been raised, so it's got worse since that EPOSS 15 task force were carrying out their job, in terms of the 16 number of PinICLs at least; would you agree with that? 17 A. That's correct. 18 Q. "PinICL fixes can only have a detrimental effect on the 19 code", presumably, again, that's a reference to code 20 decay or something similar to that? 21 A. Correct. 22 Q. It says: 23 "... CSR+ changes have been reviewed by the Team 24 Leader ..." 25 Who was that "Team Leader" a reference to, do you 14 1 know? 2 A. I would imagine that would be Steve Warwick. 3 Q. Can we go over the page, please: 4 "The figures indicate that the problems facing EPOSS 5 during the Task Force period have not diminished. Of 6 greater concern are the non-EPOSS PinICLs within the 7 group suggesting that there are still serious quality 8 problems in this vital, customer facing element of the 9 system." 10 Then there's a box there and the box says: 11 "The EPOSS Solutions Report made specific 12 recommendations to consider the redesign and rewrite of 13 EPOSS, in part or in whole, to address the then known 14 shortcomings. In light of the continued evidence of 15 poor product quality these recommendations should be 16 reconsidered." 17 Now, is that box a recommendation? 18 A. It's saying that we should reconsider the rewrite 19 option. 20 Q. But I think we saw at the beginning of this report that 21 there was an opportunity to object to particular 22 recommendations by 8 November. I think it was talking 23 about those kinds of things, wasn't it, as 24 a recommendation? 25 A. Yes. Actually, the fact -- what I'm finding -- yes, I'm 15 1 struggling because this was a massive issue for me and 2 I discussed it with all my technical team in-depth. 3 Rewriting a product is not necessarily a solution in 4 itself because you can reintroduce problems, or you can 5 have people who don't understand the requirement as 6 well, and you can actually end up with a product that's 7 even maybe a little better but not necessarily any 8 better. 9 My preferred solution at the time, having spoken to 10 all the people, is that the decay was in certain parts 11 of the product and we should focus our efforts and see 12 if we can stabilise those parts that were causing the 13 majority of the PinICL. In fact, this was a view 14 strongly felt by Steve Warwick, who I had a lot of 15 respect for. 16 Q. We will come to see how your view was that you weren't 17 in favour of a rewrite? 18 A. Well, not initially. It was still an option on the 19 table. I hadn't dismissed it. I just felt it wasn't 20 necessarily -- we should explore other avenues first 21 before taking that pretty drastic course of action. 22 Q. But it was a recommendation from ICL Pathway's auditor 23 Jan Holmes? 24 A. It was a recommendation. 25 Q. Yes. Can we look at page 47 of this document, please. 16 1 These are the terms of reference of the audit and can we 2 look over the page to "Reporting". It says there: 3 "A final report will be produced and distributed to 4 the Director and Senior Managers of the Development 5 Directorate, as well as the Managing, Deputy Managing 6 and QRM directors." 7 Then it has a list of the distribution -- at least 8 of the terms of reference and you are listed there 9 first. Now, it's not in alphabetical order but you're 10 first. Is that because you were the most senior or that 11 it was most appropriate to you? 12 A. No, Mike Coombs was the most senior. It was because it 13 was most appropriate to me. 14 Q. Thank you. Can we please go to FUJ00079783, please. 15 This was a month later and, following the audit, there 16 were a list of corrective actions that needed to take 17 place and this sets those out. So I think, effectively, 18 it sets out the recommendations and what's being done 19 about them. 20 You are listed as a recipient and throughout this 21 report you are, I think, TPA; is that right? 22 A. That's correct. 23 Q. Can we look at page 3, please, and down to the "Key to 24 plan". Now, there are various shorthand terms 25 throughout this document and one of them is "Owner", 17 1 which is "The identified owner of the corrective 2 action", and the other is "MTM", "Management Team Member 3 to whom the CA Owner reports". So there's a distinction 4 between somebody who effectively takes ownership of the 5 recommendation and the person that manages the relevant 6 team or relevant person; is that right? 7 A. Yes, yes. 8 Q. Could we scroll down, is that possible, over to the next 9 page. So these are various recommendations. We see 10 there 3.2, 3.3, I think 3.4, also, were recommendations 11 relating to various documentation that needed to be 12 actioned and you are there as the MTM, so there you're 13 taking the management responsibility for those issues. 14 A. Yes. 15 Q. Can we keep on scrolling to page 6, please. That's 16 4.2.1, thank you very much. There you are the owner of 17 this particular issue; is that right? 18 A. That's correct. 19 Q. Can we just look at the left-hand side. It says: 20 "The audit identified that EPOSS continues to be 21 unstable. PinICL evidence illustrated the numbers of 22 PinICLs raised since the 1998 Task Force and the rate of 23 their being raised. 24 "The EPOSS Solutions Report made specific 25 recommendations to consider the redesign and rewrite of 18 1 EPOSS, in part or in whole, to address the then known 2 shortcomings. In light of the continued evidence of 3 poor product quality these recommendations should be 4 reconsidered." 5 So that's effectively what we saw in that audit from 6 Jan Holmes and you are down there as the owner of that 7 particular recommendation. 8 A. The recommendation was to consider a rewrite. 9 Q. Yes. Can we look on the right-hand side. Is it 10 possible to blow up that right-hand side? There are 11 certain things there that I think I'm going to need your 12 help with because I don't quite understand. Let's see 13 where we get to. 17 November: 14 "This action falls within Development but requires 15 higher level drive. Has links with CS and BD." 16 Do you remember what that means at all? 17 A. "CS", I think, is customer services and "BD" is business 18 development. 19 Q. Thank you: 20 "MJBC to speak with TPA direct." 21 Is that -- 22 A. Mike Coombs. 23 Q. Mike Coombs, thank you very much. 25 November: 24 "Work on AI298 identified that majority of problems 25 ([approximately] 80%) were to do with error and printer 19 1 error handling. Daily meetings had been instigated. 2 TPA of view that while original code had not been good 3 it would be difficult to justify the case for rewriting 4 now." 5 So I think it's right to say that in November, 6 towards the end of November, you were of the view that 7 a rewrite wasn't your preference? 8 A. It was a very risky thing to do and if the judgement was 9 that 80 per cent of the errors were down to error 10 handling and printer handling, printer error handling, 11 then we should attack that part of the Code and probably 12 rewrite that. 13 Q. I mean, that would still leave 20 per cent, of course? 14 A. Yes, it would. 15 Q. There's an email issued by yourself, I think, and that 16 says: 17 "We have not formally closed down the recommendation 18 that we re-engineer the EPOSS application due to its 19 inherent instability. Since this recommendation was 20 made, a number of events/actions have taken place. We 21 embarked upon a major maintenance exercise for LT2 which 22 targeted several known stability issues. In parallel, 23 we carried out a defensive testing activity which 24 identified a number of faults which were addressed. The 25 intensive exercise designed to remove Acceptance 20 1 Incident 298 resulted in many substantial improvements 2 to the error handling, messaging and printing aspects of 3 this product." 4 That's the 80 per cent that we talked about just 5 now: 6 "We finally introduced improved unit and link 7 testing and more disciplined configuration control. 8 Finally, the maintainability and enhanceability of the 9 product has been proven by the speed and quality of the 10 SIP16 and EPOSS reconciliation developments. 11 "We will of course continue to monitor the PinICL 12 stack for the next few months and if necessary 13 re-evaluate this decision. Would Jan please close this 14 issue formally using the rationale described." 15 So what you are doing there is asking Jan Holmes to 16 close the recommendation because, in your view, it 17 effectively didn't need to be written at that time? 18 A. At that stage, I was suggesting that the evidence was 19 showing us that it was stabilising and that the number 20 of problems we were experiencing was reducing and 21 I didn't believe that it justified a rewrite, but it's 22 not my decision alone. That would have been discussed 23 with Mike Coombs and the board in general -- not the 24 board in general, the ICL Pathway management team. 25 Q. Who would that be? 21 1 A. That would be Mike Coombs, John Bennett and the other 2 people in the room: Tony Oppenheim, Mike Bennett, all 3 the people in the ICL Pathway management team. 4 Q. Can we go back to the page itself. So we have the 5 management team members there but you are down as the 6 owner there? 7 A. I was down as the owner but that's not a decision 8 I could have made alone. 9 Q. If we keep on scrolling on this particular document 10 there are more tasks there, more recommendations, but 11 you will see that your name is in the management level 12 there rather than the owner, so that's -- I think it's 13 fair to say the key corrective action or recommendation 14 that you were the owner of was that one about the 15 PinICL, the EPOSS system? 16 A. Yes, without doubt. There were others -- there's one at 17 the bottom of that page which was down to me. 18 Q. Yes. Can we look at WITN04600104 please. This is the 19 same report but a bit later. It is version 2.0 and it 20 is dated 10 May 2000, so it's six months later. Can we 21 go to page 6 and do precisely what we did on the last 22 document, which was scroll through. You will see there, 23 3.2, for example, by that stage had been closed. We can 24 scroll to the next page. Those early documentation 25 ones, I think they were all closed. 22 1 Then it is page 9 where the 4.2.1 appears, and 2 that's what we saw last time, but if we scroll over the 3 page, it has been updated. Thank you very much. Again, 4 I'm going to need your help with a bit of interpreting 5 here, I think. If we could look at the right-hand side 6 of the screen, perhaps we could blow that side up. 7 8 December: 8 "JH [that's Jan Holmes] requested statistics on 9 fixes delivered to live from RM." 10 Who was "RM"? Might that have been Royal Mail, 11 I wasn't sure? 12 A. No, no. It could be release management. 13 Q. "Also informed [yourself] that requires agreement of 14 [Mr Coombs] before this can be closed." 15 A. Absolutely. 16 Q. So Jan Holmes there is asking for statistics on fixes 17 before he can be satisfied that it should be closed. 18 A. Yes. 19 Q. Then we look at 8 December: 20 "[Mr Coombs] confirmed that unless [maybe release 21 management] statistics contradicted reports provided by 22 PJ ..." 23 Is that Mr Jeram? 24 A. Peter Jeram, yes. 25 Q. "... the recommendation could be closed." 23 1 Nothing is disclosed in December and nothing is 2 closed in January, February, March and we're in April 3 now, 7 April. There's, again, an email to Mr Coombs, 4 yourself and Mr Jeram: 5 "... providing details of [release management] EPOSS 6 fixes to live. Asked for confirmation that matched PJ 7 reports. If does then will close." 8 So was confirmation given before 10 May of the sort 9 that Jan Holmes was requesting? 10 A. If Mike closed it, then yes. 11 Q. Well, we will see the basis on which it was closed but 12 it does seem like quite a few months have passed and the 13 kinds of statistics that were being asked for there 14 weren't produced, or there seems to be some sort of 15 issue because we go through, as I say, December, 16 January, February, March, April -- 17 A. No, I don't think it's -- excuse me, it's not -- I don't 18 think it's suggesting that. I think it's suggesting 19 that, through the observation period, while further 20 testing was going on, that the statistics didn't -- 21 demonstrated that the product had stabilised and was no 22 longer producing the kind of problems it was before. So 23 it's a case of saying that -- Mike's saying that "If 24 I can see statistics from release management that 25 support the recommendation, then I will authorise the 24 1 action to be closed". 2 Q. It certainly seems as though, throughout early 2000, at 3 least, he wasn't getting those statistics, would you 4 agree with that? 5 A. I -- there is an inference of that, but I can't recall 6 that. 7 Q. We're going to return though document but perhaps we can 8 just go to FUJ00058190, please. This is the ICL Pathway 9 Monthly Progress Report for February 2000. Is this the 10 kind of document that you would have seen at the time? 11 A. I used to write one of the sections of it. 12 Q. Can we look at page 5 of that report. Thank you very 13 much. So rollout is on track by that stage: 14 "We have now exceeded 4,000 post offices and are 15 achieving the targeted 300+ implementations per week. 16 This is a tremendous performance ..." 17 So at that point, February 2000, quite rapid 18 acceleration of the rollout. 19 Can we look at page 6, please. There is a heading 20 "New Business": 21 "Now that Acceptance has been achieved and National 22 Rollout and Customer Service are seen by [Post Office] 23 as going well, there are positive engagements now 24 starting on new business." 25 So it seems as though there is movement towards 25 1 focusing on new business, by that stage. Would you 2 agree with that? 3 A. That's what it implies, yes. 4 Q. Can we look at page 24 of this report, and please do 5 tell me if there's something that you recognise that you 6 wrote -- if this page 24, for example, is something that 7 you wrote, then please do let us know? 8 A. No, it's not. My section is the one headed 9 "Development". 10 Q. Okay, so this is "Acceptance Loose Ends". Do I take it 11 from that that those are certain things that haven't yet 12 been resolved? 13 A. It would appear that way, yes. 14 Q. Can I read to you that second bullet point. It says: 15 "We have dealt with queries from POCL concerning 16 AI376." 17 That was the lack of data integrity AI: 18 "One formal letter has been responded to attempting 19 to avoid the conclusion that we had not found EPOSS 20 reconciliation incidents that we should have found or 21 that we have not reported those we did find. In reality 22 CS are greatly hampered in 'spotting the incident' 23 because the reports have not had fixes implemented and 24 report large amounts of do-nothing information. We have 25 attended the Release Management Forum and proposed some 26 1 reordering of the fix backlog, but it will be at least 2 until the first week of March before this situation 3 improves." 4 So, I mean, the impression that you get there is 5 that, despite the rollout going and progressing rapidly, 6 there is some dispute about under-reporting from ICL. 7 Do you remember that, or is that a fair observation in 8 relation to that paragraph? 9 A. "CS" is customer service and what we were doing at this 10 stage was it was moving from a development kind of 11 project into a customer service project, so it was 12 a transformation. This is what often happens in IT 13 programmes. Once they have gone through an intensive 14 development phase they have to move into a system 15 support service management kind of environment. 16 To do that, you need a completely different 17 organisational structure and you need different skills 18 and such-like, and Stephen Muchow who was the service 19 manager, this is his report, and it's -- it would appear 20 to be doing -- he is reporting against the helpdesk and 21 the flow from the helpdesk through and how fixes are 22 being identified. 23 I was right at the end of the chain, if you like, so 24 when you get the helpdesk, which was a mixture of 25 experts on both POCL and ICL Pathway, and that's where 27 1 some of my team ended up moving towards second line and 2 third line within that, and we would also work on the 3 helpdesk occasionally, so they were no longer reporting 4 into me, they had moved across into a service management 5 environment. So what this is doing, it seems to be 6 suggesting that Stephen at that point is not happy with 7 that process. 8 Q. But does it seem that in early 2000 there were some 9 allegations from the Post Office that ICL Pathway 10 weren't reporting as many incidents as perhaps they 11 should have? 12 A. I don't believe that to be true. We reported 13 everything. 14 Q. But do you remember an allegation of that sort? 15 A. No, I don't, actually. 16 Q. "CS [customer service] are greatly hampered in 'spotting 17 the incident' ..." 18 I mean, were there issues in early 2000 with really 19 spotting incidents amongst your team? 20 A. Well, that's written in a way -- it says customer 21 services "are greatly hampered in 'spotting the 22 incident'", and I don't understand that statement. 23 I genuinely don't understand it. I can only guess and 24 speculate that it is to do with -- well, it's saying 25 "spotting the incident". The incident would be reported 28 1 to the helpdesk, so I think it could be saying that the 2 team that was in place at the time were struggling to 3 identify where the problem is. 4 Q. Can we go back to WITN04600104 and to page 10, please. 5 Thank you very much. Sorry, could we go to one page 6 before that -- and over the page. Sorry, other way. 7 Perfect, thank you very much. 8 So I put to you earlier about late 1999/early 2000 9 is seems as though -- the inference, as you agree, from 10 that document is that there's some sort of problem in 11 producing the statistics that have been asked for by Jan 12 Holmes and, set against what I have just shown you, 13 I mean, would you accept that it looks as though there's 14 some sort of issue going on in early 2000 about 15 providing accurate information about the number of 16 incidents? 17 A. Yes, they should -- there was, there was. It would 18 definitely imply that there was an issue there. I can't 19 recall what the issue was, no I can't. 20 Q. Then there was a reminder on 3 May and then 10 May you 21 have this, and this is a response received from 22 Mr Coombs: 23 "As discussed this should be closed. Effectively as 24 a management team we have accepted the ongoing cost of 25 maintenance rather than the cost of a rewrite. Rewrites 29 1 of the product will only be considered if we need to 2 reopen the code to introduce significant changes in 3 functionality. We will continue to monitor the code 4 quality (based on product defects) as we progress 5 through the final passes of testing and the introduction 6 of the modified ..." 7 Is it C14 or CI4? It's a fix of some sort? 8 A. It's a release of some kind, yes. 9 Q. Yes. Let's say CI4: 10 "... codeset into live usage in the network. PJ can 11 we make sure this is specifically covered in our reviews 12 of the B&TC test cycles." 13 Then it is closed on the 10th. So it says 14 "Effectively as a management team" you have decided -- 15 who was the management team? 16 A. Well, the management team was the people in ICL Pathway. 17 That would be people like Peter Jeram, myself, Mike 18 Coombs and Stephen Muchow and various other people. 19 Q. Now, by referencing the management team there, by the 20 sound of it the decision of management might have been 21 taken contrary to those lower down the chain. Would you 22 accept that? Were there, for example, people within the 23 team who were really saying at that stage "We really 24 need to rewrite this product"? 25 A. There could be some programmers, yes. Yes, there was -- 30 1 there was a difference of opinion, without doubt. 2 Q. That was a different of opinion between management and 3 those -- 4 A. No, there was a difference of opinion within the 5 technicians, so the problem I had as a manager is I was 6 getting contradictory information. I was getting a view 7 that was -- from the PinICL viewpoint I could see the 8 product was unstable and when I'm trying to identify 9 what the issue is and what we're going to do about it 10 and talking to the various people, there was two 11 different views: there was those in the team that felt 12 it should be rewritten and those in the team that felt 13 that we should focus our efforts in certain aspects of 14 the products. 15 Q. Do you think those who felt it should be rewritten were 16 in the majority? 17 A. I don't know, but they were just equal on the people 18 I spoke to. 19 Q. I mean, the reference there to management team suggests 20 at least that it was management who thought it shouldn't 21 be, but perhaps those below thought it should be? 22 A. As I said originally, there would have been people, 23 programmers, who may have felt that it was the right 24 thing to do. 25 Q. Can we look at your witness statement, that's 31 1 WITN04190100. Side by side if we can, but if we can't 2 that's not a problem. It is page 12 of your witness 3 statement, paragraph 26. Thank you. So it's about 4 halfway down that paragraph you say: 5 "The option was debated at length by senior members 6 of the ICL Pathway management and technical teams and 7 the outcome was that we should embark upon a major 8 exercise to target the specific areas known to be the 9 source of most (circa 80%) of the issues identified 10 which were error handling and printing. If this 11 approach was unsuccessful, then a rewrite would be the 12 only option available. However, the product did become 13 stable and the number of outstanding defects did fall 14 within the levels defined in the acceptance criteria." 15 It may be suggested that that is slightly 16 inconsistent with what's being communicated by 17 Mr Coombs. I will run you through where those 18 inconsistencies may or may not lie. 19 You have said in your statement, for example, it was 20 debated by members of management and technical teams 21 and, as I say, Mr Coombs seems to focus on the 22 management team. Again, do you think it is possible 23 there that it was the technical teams who were in favour 24 of the rewrite and management who weren't? 25 A. No, no. The technical teams means the people in my 32 1 development and design -- architectural design and 2 development, who I had the most confidence in and we 3 would thrash, looking at the facts and decide what was 4 the best option forward. So they were the -- comprised 5 of people like Alan Ward who was my chief architect. 6 They would comprise of people like John Hunt who was one 7 of the consultants on the team. Steve Warwick who was 8 the expert in EPOSS and Pete Jeram who was the 9 development manager at the time. So that's what I mean 10 by the "technical team" is that we would have pulled in 11 people that we had confidence in and we would thrash out 12 what we thought was the best way forward. 13 Q. Amongst those names that you have mentioned, were they 14 all in favour of a rewrite? 15 A. No, none of them were. 16 Q. Your explanation in the witness statement seems to focus 17 on what we know as AI298, that's the overall stability 18 Acceptance Incident. Would you accept that there were 19 other issues, such as cash account imbalances, whether 20 caused by what we know as AI376 or something else at 21 that time, which were still occurring and still related 22 to the EPOSS product? 23 A. That's right. What I'm implying in my witness statement 24 is that, if you like, we were judged on acceptance and 25 acceptance was whether the product, the overall product 33 1 met the business requirements as stated in the 2 functional specifications, the business requirements 3 specifications, and whether it met the criteria for the 4 number of defects and errors that were still available 5 in the product. 6 If we achieved that, which was -- if I recall was 7 zero As and ten Bs or -- and no restriction on Cs at 8 that point, then from that perspective we have met the 9 criteria on two fronts. 10 Now, if the decision we had made was -- wasn't the 11 right one, then it would have shown up in that and we 12 would have failed acceptance, so that's what I'm 13 implying in my witness statement, is that we wouldn't 14 have achieved acceptance. 15 Q. So would you accept -- and I think others such as Tony 16 Oppenheim have said this -- that there would still be 17 some circumstances where it wouldn't be possible to 18 identify what has gone wrong? 19 A. One of the biggest difficulties with the instability 20 issues was, in the 1990s, that you -- some of you, 21 forgive me, may recall in the '90s if you had a PC that 22 actually it was very subjected to blue screens and lock 23 ups. 24 I mean, nowadays you never see a blue screen, but 25 Windows NT in the '90s, you would get a blue screen and 34 1 trying to track down what had caused that in such 2 a complex system as Pathway was extremely difficult 3 because a lot of the blue screens we were experiencing 4 were at the counter, frustrating the postmasters 5 considerably, understandably, but trying to understand 6 what was happening at the time -- because the blue 7 screen would then -- you would reboot it and it would go 8 away and, as most of you know now, even a common 9 solution to a printer or a PC or a laptop is just to 10 switch it off and switch it back on again, but when 11 you're in a distribution system, switching it off and 12 switching it back on again, which is a reboot, we want 13 to find out what the problem is. 14 We have no way of seeing the counter. We can't get 15 onto the counter to have a look at it and we can't take 16 what we used to call "dumps" or "print-outs" of the 17 store, the message store to try and see what caused that 18 issue. 19 So that was a constant thorn for us to try and -- we 20 would get a lock-up or a freeze and all you could 21 suggest to the people was that "We will try and find" -- 22 and we did find quite a few of them and it was down to 23 two or three very, very clever individuals that managed 24 to track down what some of these were and they were 25 quite obscure. So that's what I'm referring to is that 35 1 the stability one was to do mainly with blue screens and 2 freeze -- and lock ups and the AI376 was to do with 3 balancing the cash account. 4 Q. Absolutely, so your focus in that particular passage as 5 well is on AI -- what we know as AI298, that's the 6 overall -- 7 A. It's on the stability one, yes. 8 Q. But you clearly accept at that stage there were other 9 issues, especially with the EPOSS product, that were 10 continuing, even if, as you say, it may have been circa 11 80 per cent that was the overall stability issue? 12 A. Yes, but we -- during the period from there onwards we 13 focused on PinICLs. I called -- I introduced a -- 14 during the -- when these PinICLs were being raised and 15 the problems on reconciliation and freezes and lock ups 16 were being identified, we had what we called "morning 17 prayers" and every morning at 8.30 am all the top people 18 in my team would meet and we would go through the latest 19 incidents and the ones that we had -- previous 20 incidents, find out what progress we had made on the 21 ones that we had identified previously, and any new ones 22 overnight -- the previous day and then we would decide, 23 during that meeting, the course of action and then we 24 would meet the next day. We did that day after day 25 after day for many weeks. 36 1 Q. Absolutely. I mean we're here in -- this is May 2000 2 so -- 3 A. This is closing down something in May 2000. It's not -- 4 and it's as a result of what's been going on and what 5 Michael is saying there is that there was no evidence to 6 suggest that it shouldn't be closed. 7 Q. Your evidence is that, after closure, there were still 8 a number of incidents still continuing and you had 9 dedicated -- 10 A. No, no, no, that was before that. 11 Q. So after May 2000 there weren't incidents? 12 A. No, there would still be incidents, there's always 13 incidents in a system of that size but they were being 14 monitored through the helpdesk and down through the 15 support channels. 16 Q. But by May 2000 your focus had been on AI298 in 17 particular? 18 A. And -- well, 376 as well. 19 Q. But 376 you would accept continued? 20 A. We were looking at all issues identified by -- coming 21 down through all PinICLs, all incidents that had been 22 raised on the EPOSS product at that point and we were 23 focusing on every one of them. So we didn't leave these 24 to one side, or -- every one that appeared like it 25 looked like it was EPOSS -- and quite often it wasn't 37 1 EPOSS, it was either to do with the TIP interface or it 2 was to do with the processes or it was to do with 3 reference data, or it was to do with migration or some 4 kind. Just because it was identified, that was where 5 the source of the problem existed, that didn't 6 necessarily -- that's where the source of the fix 7 existed. 8 So I would bring everybody into my office and every 9 morning, witnessed by senior Fujitsu test and diagnostic 10 experts that came over from Japan and were allocated by 11 Fujitsu who sat in my meetings to watch me and decide 12 whether they were happy with the process. So I was 13 under a lot of scrutiny there and I personally wanted to 14 see and get this sorted out. It mattered an awful lot 15 to me to get it sorted out. 16 Q. So are you saying that throughout the year 2000 it was 17 well-known in Fujitsu, including people who came over 18 from Japan, that there were -- 19 A. We've got a timescale problem here. May 2000 is at the 20 end of that period not the beginning. I'm talking about 21 the period prior to that when we were doing model office 22 testing, end to end testing, acceptance testing and live 23 trial. 24 Q. So by May 2000 it all suddenly stopped? 25 A. No, it -- EPOSS was not providing the kind of problems 38 1 that would justify rewriting the code. That's what 2 Mike's saying. 3 Q. So it wasn't providing sufficient problems to justify 4 rewriting the code? 5 A. Correct. 6 Q. But it was still providing problems? 7 A. As were other parts of the system, yes. 8 Q. But you would accept that the EPOSS was still an issue 9 post May 2000? 10 A. Well, it depends what you mean by an issue. There were 11 problems, defects which were being dealt with in the 12 normal support way. 13 Q. I mean given all the information that you received over 14 those two years, so starting from that Holmes/McDonnell 15 report in 1998 and then the report in 1999 and the 16 issues in early 2000, didn't you think that it might 17 have been the time to start thinking about rewriting 18 EPOSS? 19 A. I did, I did, several times think about it and I was 20 persuaded by the technicians working on the product that 21 they felt they could sort it out. 22 Q. Were there any particular technicians who you felt 23 persuaded you? 24 A. Steve Warwick, in particular, because he knew the 25 product better than anybody else. 39 1 Q. Let's look at FUJ00079333, please. Now this -- the top 2 email, the top two emails, are emails of 10 May 2000 in 3 the evening, 6.36 pm, 6.28 pm in the evening of 10 May, 4 so it looks, certainly from this -- from the time, that 5 that was after the issue had been closed; would you 6 agree with that? Do you remember was there a meeting in 7 the daytime, or discussions in the daytime, on the 10th 8 to close the recommendation? 9 A. I can't recall, but it -- this seems to be after that, 10 yes. 11 Q. Now, let's look at that second email on the screen, so 12 this is an email from you. In fact can we go over the 13 page. We will start with the original request, so this 14 is actually before the closure, so this starts in April, 15 27 April. Do you remember who Pat Lywood was at all? 16 A. No. 17 Q. So these are current issues -- 18 A. Oh, I think Pat Lywood was someone that went -- that was 19 in the customer services support line. 20 Q. It says "Current issues on", and that's I think either 21 C14 or CI4 EPOSS, and can we look -- halfway down the 22 screen we can see there "Balancing process overheads": 23 "After migration to CI4 a new process is introduced 24 to the cash account process. Every office will be 25 required to declare non-value stock. If the office 40 1 fails to do this process he will not be able to balance 2 or complete the cash account." 3 Then it says that: 4 "Paul Westfield and I will ensure this is included 5 in the backfill training provided to the existing 6 offices." 7 Then further down there's reference to "Risk of code 8 regression", that's another heading: 9 "There may be fixes that have been produced and 10 delivered into CI3 that have been missed from CI4. 11 "I will take this up with Dave Royle and ask for 12 assurance that all clone PinICLs have been tested. 13 I will supply a list of the PinICLs that we have tested 14 in CI3R release." 15 So that's the start of the chain and, if we go to 16 page 1, Stephen -- is it Muchow or? 17 A. Muchow. 18 Q. At the bottom of the page, sorry. He sends you 19 an email, again on 27 April, so before the closure of 20 the EPOSS issue, and he says: 21 "I am particularly concerned with the risks of 22 degraded counter and cash account performance and of 23 code regression between CI3 and CI4." 24 Then you respond in the email above that -- and one 25 thing that may become relevant in due course, Gareth 41 1 Jenkins is copied into that email -- and you say at 6.36 2 on 10 May, so presumably after the recommendation has 3 been closed: 4 "Steve, I share your concerns regarding counter 5 performance and code regression. To that end we are 6 focusing on those areas of functionality where we appear 7 to be experiencing performance degradation and 8 attempting to establish where the problem lies. I have 9 been personally aware of these problems for several 10 weeks and would not expect CS to authorise CI4 unless 11 these issues were resolved. I have raised the issue of 12 extra work during weekly balancing with Mike who will be 13 discussing it with Dave Smith. This has been introduced 14 by POCL to support LFS. 15 "I cannot give you a 100% guarantee that code 16 regression will not occur at CI4 because, by its very 17 nature, it is not fully automated and never will be. 18 However, our end to end processes are designed to reduce 19 the possibility of this occurring to an absolute minimum 20 and I have recently requested a reconciliation where it 21 is possible to do so." 22 So you have said there that there's a problem that 23 you have been aware of for several weeks and you cannot 24 give 100 per cent guarantee that code regression won't 25 occur, keep it to a minimum. Code regression, that's 42 1 similar to code decay, is it, but it means, I think, 2 that by fixing one thing, it could break something else 3 that was working before? 4 A. Yes. It's not decay, no. Code regression -- we used to 5 have a testing sequence called regression testing and 6 whenever you put a fix into a product of any kind you 7 will subject it to regression testing to see that you 8 haven't undermined or introduced another problem, or 9 affected something you had already done, hence the 10 reason it is called "regression testing", to ensure that 11 you have not regressed the problem. 12 So the problem is that, when you're fixing faults, 13 you cannot guarantee that you haven't caused 14 a regression because it's technically -- if a programmer 15 puts the fix in, he does it, he tests it to his ability, 16 it then goes into regression testing and regression 17 testing says it's okay. It then goes into the live 18 environment and because you have assembled far more of 19 the system at that point, there's so many more moving 20 parts, then you may -- another error may crop up. So 21 some regression may -- that may well happen and that 22 will be -- that is the case with every IT system I'm 23 aware of. 24 Q. Is it a bit like Whack A Mole, where one problem comes 25 up, you try and fix it and then something else pops up 43 1 somewhere else? 2 A. You may have inadvertently -- by fixing the problem you 3 may have re-introduced something else or you may have 4 knocked on to some other part of the programme that the 5 programmer didn't release. 6 Q. As I say, the evening that that recommendation about 7 EPOSS was closed, I mean, there may be some people 8 asking how could you close that recommendation knowing 9 all that is contained within that email? 10 A. This is to do with performance. This is to do with 11 performance degradation. This is not to do with EPOSS 12 degradation. It's another issue. What's happening is 13 that we're noticing that the time it's taking on -- the 14 counter performance on its response times and the time 15 it's taking to do the cash account is degrading and that 16 is a performance issue. What -- we don't know and we're 17 trying to find out what's causing that, so that's why 18 I was aware of it, is because I had been aware that it 19 was happening but we were finding it extremely difficult 20 to reproduce it in the laboratories. 21 Q. But if we look at the subject, it's -- C14 was 22 a particular fix on EPOSS, so it was a problem that was 23 fundamentally on EPOSS? 24 A. It was on EPOSS but it was a performance element. What 25 it says is regarding counter performance and code 44 1 regression so was there something happening where the 2 code was regressing which was causing the performance to 3 get worse? So that's the question that I couldn't 4 guarantee and what Steve is asking the question. 5 Q. Presumably, those issues could have real life 6 implications for those who were trying to balance, for 7 example? 8 A. Yes, yes. I mean, I was very concerned that we -- it 9 was taking longer to do the account balancing at the end 10 of the day than it should. 11 Q. But do you think that these kinds of issues should have 12 been raised with Jan Holmes before closing that incident 13 on that day? 14 A. We didn't have -- we didn't have any PinICLs. It was 15 something that I was aware of that what appeared to be 16 happening -- there was no proof, there was no evidence. 17 All that we were getting was a feeling that the counter 18 seemed to be degrading in performance while we were 19 going through this work, and this is -- Steve's making 20 the same point, so it -- I get where you're coming from, 21 but we would have seen these as two different issues. 22 In hindsight, maybe -- I can accept the point you're 23 making is that maybe the fact that these were starting 24 to occur, we should have perhaps raised them during 25 that, but we weren't -- at the time that Jan wrote the 45 1 original report and the time that Mike -- we had no 2 evidence that it was -- I'm not being very clear here. 3 What I'm trying to say is we couldn't -- we didn't 4 know what was causing the problem and we didn't know 5 what extent the problem was either. 6 Q. Do you think that might have been something to tell Jan 7 Holmes? 8 A. (Pause) The reason I'm hesitating is I think you're 9 probably right. In hindsight, you're probably right, 10 but counter performance and the time it was taking to do 11 something we used to look at in a different way. We 12 didn't -- it could have been for a variety of reasons. 13 It may not have been anything to do with the EPOSS 14 product. It may have been to do with the way that the 15 counter was booting in the morning. It could have been 16 all sorts of things that may have been causing that to 17 happen, so it wouldn't -- it wouldn't necessarily be 18 pointing at EPOSS. 19 Q. But you're talking there of code regression as well as 20 counter performance so -- 21 A. Yes, but that could be code regression anywhere in the 22 counter. He is talking -- he is saying EPOSS -- and 23 Steve wasn't a technical guy, so what -- it's 24 terminology being used there. I have responded in 25 saying -- because I know what he is referring to -- that 46 1 it was to do with counter performance and potential code 2 regression, but that's not necessarily just EPOSS. 3 Q. One of your solutions to the EPOSS problem though was to 4 implement fixes along the way, rather than rewriting, so 5 presumably every fix, there's potential for code 6 regression? 7 A. Yes, see, I know that we found out what was causing this 8 and it wasn't EPOSS. So that's the point I'm trying to 9 make is there's several elements -- if -- you may have 10 already seen, if you have seen a technical environment 11 description of the system, there are several elements 12 that sit within the counter and one of the issues in the 13 '90s was trying to get any system to boot up with all 14 the mass of software that had to initialise in there, 15 and things like the counter slowing down could have 16 been -- as I said, it could have been to do with any 17 product within the counter that was causing that, not 18 necessarily EPOSS, and we did resolve this problem. As 19 far as I can remember, we did resolve the problem. 20 Q. I want to take you to a PinICL. It will only take 10 to 21 15 minutes. Would you like to break now or can we break 22 after I have taken you to that, because that's the end 23 of my questioning on EPOSS issues? 24 A. No, we can carry on. 25 Q. Thank you very much. Can we look at FUJ00067416, 47 1 please. It's a PEAK rather than a PinICL. For those 2 who have been following this Inquiry, this also appears 3 within the expert report at page 157. 4 Now, I'm going to need your help quite a lot with 5 this because it's not very easy to understand, but can 6 we look at 16 May, so the first entry. This is six days 7 after that EPOSS issue was closed. It says: 8 "The host generated cash account line comparisons 9 report dated [15 May] where post office 169207 has 10 a difference in the receipts and payments total for cap 11 [that's cash accounting period] 06. Please 12 investigate." 13 Now, the third entry there, can we just scroll down 14 slightly, so it's 19 May -- actually it's the fifth 15 entry, sorry: 16 "This office has had big problems with its receipts 17 and payments. [Cash account periods] 5, 6 and 7 did not 18 match. The differences are ..." 19 It gives the difference: 20 "The office has already reported problems balancing 21 which are being investigated by development ..." 22 So big problems being identified and I think if you 23 look at two entries down there's another report on 24 19 May, receipts and payments issue in CAP7. 25 Can we go to 24 May, that's over the page, page 2. 48 1 You so it says there: 2 "The cause of the problems in all three [Cash 3 Accounting Periods] at this outlet was the fact that 4 Stock Unit DD's rollover records from CAP5 to CAP6 5 represented a 'nil' balance (the total stock holding was 6 nil, no receipts or payment transactions were recorded) 7 despite the fact that the stock unit had been trading 8 normally during this period. This issue was raised in 9 PinICL 43811 and is still under investigation within the 10 EPOSS development team." 11 Was that your team, the EPOSS development team? 12 A. No, not at that stage. As I said to you, it had moved 13 over into a support environment and I wasn't responsible 14 for the team at that point. 15 Q. But there was a specific team dealing with problems with 16 EPOSS on 24 May -- 17 A. Yes. 18 Q. -- and they dealt with issues to do with -- 19 A. Absolutely. 20 Q. -- balancing issues. 21 A. Absolutely. 22 Q. Can we look at the entry on 30 May, that's one entry 23 down. There's more information there about further 24 investigation. I'm not going to read that out. Perhaps 25 if you could just take a short period just to have 49 1 a quick look at what it says there. 2 (Pause) 3 Then if we go down to the bottom of the page there's 4 an entry of 4 July. Can we just scroll over to the next 5 page. At the top of the next page seems to be 6 an explanation: 7 "This problem is the same that already resolved in 8 PinICLs 37884 & 37663, namely that of DataServer not 9 tree building & populating correctly. A diagnostic has 10 been put into DataServer to detect any such problems." 11 That's the explanation. 12 Can we look at 12 July. This is where it becomes 13 difficult to understand and I'm going to need your help 14 if you're able to, on page 4, the entries from 12 July. 15 Can I ask you just to read to yourself those entries 16 briefly. 17 (Pause) 18 If you're able -- if we could carry on scrolling 19 perhaps, because there's another entry on 12 July at 20 12.29. Can I ask you to read that final entry. 21 (Pause) 22 A. Right, okay. 23 Q. Can you help us, it seems as though what they are trying 24 to do is work out how to reproduce the cash account as 25 it should have been prior to an error; is that right? 50 1 A. That's true, but they also believed that they understand 2 what has caused the problem, which is pretty obscure. 3 Q. Yes. Can we have a look at the next page, 8 August, 4 halfway down the page 2.35 in the afternoon on 8 August: 5 "I have spoken to Martin McConnell who advised call 6 to be routed to EPOSS ..." 7 That's, again, that EPOSS team, is it? 8 A. I don't know what FP stands for because, at that point, 9 as I said, it's not within my chain. 10 Q. Can we go over the page to 13 September, please. There 11 are two entries on the 13th. If that first one could 12 be -- perfect, thank you very much. Again, I'm going to 13 ask you to read those to yourself. I mean, I will read 14 just very briefly the first one. It says: 15 "It proved to be very difficult to resurrect the 16 cash account data for week 5. Steve Warwick's analysis 17 tool showed that not only was stock unit DD corrupt but 18 also stock unit XXX. EPOSS nodes ... were missing and 19 had to be resurrected." 20 If I could ask you just to have a quick look at that 21 and then also the one below that. 22 (Pause) 23 A. Yes. 24 Q. And the one below, sorry. 25 (Pause) 51 1 A. Yes. 2 Q. Are you able to briefly describe what you think is going 3 on there? 4 A. There was -- it would appear that they have come to the 5 conclusion that something had kicked in on archiving 6 from the counter, which caused the problem, and it is 7 that -- as I said before, that's obscure for that to 8 have happened and, as a result of that, it's caused this 9 impact and then it would appear that they are struggling 10 to be able to reproduce the cash accounts and the 11 figures that would balance them off and they're 12 suggesting the way in which they can move forward on it. 13 Q. Can we go to the final entry on that page, and it goes 14 over the page, so this is 14 September 2000. I will 15 read this one. This is an entry that says: 16 "Thanks for all the effort. For the time being 17 I have agreed that reconstructed cash accounts will not 18 be needed all the time, but only by special request of 19 POCL. 20 "I have already issued the final BIM report. 21 "As such please close this call, and hope for the 22 best with the CI4 code which should make this type of 23 incident very rare." 24 So, I mean, here we are in September 2000 and the 25 approach is "Let's hope for the best". I mean, is "hope 52 1 for the best", is that what you decided to do in terms 2 of the rewrite of -- 3 A. I was not responsible for this -- 4 Q. No, but "hope for the best" was that the kind of 5 attitude that was taken in respect of the EPOSS product 6 back in May: hope for the best, hopefully it will be 7 very rare? 8 A. No, no. The data that had been received prior to May, 9 which resulted in Mike suggesting that it should be 10 closed, didn't suggest -- it suggested that we had 11 resolved the issue. There were still going to be 12 problems because of the sheer nature and scale and 13 complexity of the project, but the original issue which 14 was to do with error and printer handling and cash 15 account balancing, we believed that in the majority of 16 instances, 99 per cent of the -- we had managed to sort 17 it. That's what Mike was referring to. 18 We knew -- well, we didn't know but, as a result of 19 this, which I had no involvement with whatsoever, but 20 a very obscure incidence of where an archiving programme 21 kicked in, which caused this problem -- it shouldn't do, 22 but it did and it seems to be something to do -- between 23 Riposte and the archiving suite that caused this. And 24 these are the kind of problems you get in large systems 25 and the idea is to try and track it down. 53 1 It would appear that what they're saying is this 2 shouldn't happen and it may happen again, but if it does 3 it would be another PinICL. We would have to look 4 straight into it straightaway, if it happened, is what 5 they're saying. "Hope for the best" is it should sort 6 itself out in C14, is what he is suggesting. It's not 7 terminology I would have used but -- 8 Q. "It should sort itself out" is, in fact, a phrase we 9 have heard quite a lot of, especially during the human 10 impact stages. Was the general feeling that things will 11 sort themselves out? 12 A. No, no. Okay, my terminology wasn't very helpful there, 13 but what I'm trying to say is somebody -- I wasn't 14 responsible for C14 or -- I don't know what it was, 15 sorry, so I can't be helpful on that. I don't know what 16 it involved. I don't know what was in there. There may 17 have been a lot of fixes put in there. 18 Q. Can we look -- and this is the final thing before the 19 break -- at FUJ00080690. It is the first document that 20 I took you to today. Can we look at page 7, please, and 21 at the top of page 7. 22 So in 1998 you were being told that there were 23 hundreds of fixes, code decay, the system was unstable, 24 no guarantee it won't adversely affect another part of 25 the system. I mean, looking back at that paragraph and 54 1 knowing what you know now, do you think you should have 2 agreed to rewrite EPOSS back in 1998 or 1999? 3 A. The best advice I was getting at the time was that, if 4 we were to rewrite, did we have the -- did we have the 5 people to do it, did we have the expertise to do it and, 6 by doing so, would we run the risk of just creating 7 another problem because one of the reasons why this got 8 into this situation is that we were forced to do rapid 9 application development and, by doing that, you haven't 10 got a functional specification, you've got what we 11 call -- you're talking to people and you're trying to 12 get EPOSS to sit along -- POCL people to sit alongside 13 you and the problem was that we had people who were 14 working in that environment that weren't very 15 professional and weren't very good at their job. 16 So I was convinced by the people that were giving me 17 the evidence that it was a certain part of the EPOSS 18 product that was responsible for a very large 19 per cent -- we say 80 per cent but it could have been 20 90 per cent. I was measuring whether that was a good 21 decision by the number and type of PinICLs that we were 22 getting, come the live trial, and that's how I measured, 23 personally, whether that was the right decision to do. 24 If the product had been as bad and Steve, for 25 example, and others had been wrong, then we would never 55 1 have got to that stage. We would never have got to the 2 acceptance situation and number of PinICLs. We would 3 have had a product that we would have had to have 4 rewritten. 5 Q. Do you think that you listened enough to the members of 6 the team who were urging a rewrite? 7 A. The reason I'm hesitating is that I believed I did. 8 I believed -- to their argument. I listened to their 9 argument and so did some of the people who were more 10 technical than I was. I believed that everybody was 11 given an opportunity to give their view of what should 12 happen. They believed that we should rewrite it and 13 there were other people that thought that we shouldn't. 14 Q. Once you had taken the decision not to rewrite, did you 15 sideline those who were urging a rewrite? 16 A. Not as far as I'm aware. No, not -- I don't believe so. 17 Q. By May 2000, so that was the date when the 18 recommendation was finally closed, presumably that was 19 far too late then to start thinking about a rewrite, 20 given the number of Post Offices that already had the 21 system in place? 22 A. No, it wasn't too late. I mean, you could rewrite 23 a product over a period of time to match the user 24 interface that the postmasters were used to. You could 25 replicate that functionality in parallel and then 56 1 release it at a later date if you felt that was the 2 right thing to do. 3 Q. At no point did you feel that that was the right thing 4 to do? 5 A. Not during my time. I didn't have -- I'd have 6 sufficient evidence to suggest that that was the right 7 thing to do. The ones that you pointed out that were 8 picked up later were very obscure situations. There was 9 nothing there to suggest it was a rewrite. 10 MR BLAKE: Thank you. 11 Sir, I think that's an appropriate time for 12 a ten-minute break. 13 SIR WYN WILLIAMS: Yes, fine. What time shall we resume? 14 MR BLAKE: I think we can actually -- we can come back at 15 midday. 16 SIR WYN WILLIAMS: Midday, all right. Have a break, 17 Mr Austin, and don't talk about your evidence, although 18 I'm sure you won't think of it at any rate. 19 A. Thank you. 20 (11.40 am) 21 (Short Break) 22 (12.00 pm) 23 MR BLAKE: Sir, we can see you now. 24 SIR WYN WILLIAMS: Good. 25 MR BLAKE: Mr Austin, I only have about 15 more minutes and 57 1 then I'm going to hand over to Mr Stein and Mr Henry to 2 ask you a few further questions. 3 Briefly, while we're on EPOSS, I want to ask you 4 about RAD, rapid application development. 5 We know from a report that was sent to the Post 6 Office by project mentors -- the reference there is 7 POL00038829 but we don't need to bring the document 8 up -- that Pathway started with rapid application 9 development methodology, but that appears to have been 10 discontinued after a staff member left the project. Do 11 you know anything about that at all? 12 A. Yes. Yes. It was not proving to be very successful. 13 Not only was -- one of the staff members left but also 14 it did depend -- it was very dependent -- had dependence 15 on POCL providing experts that to define their 16 requirement, so -- and that was very time-consuming for 17 POCL and POCL didn't feel that they had enough people to 18 fulfil that role. So we decided to reverse engineer and 19 produce a document that then could be used to use 20 a traditional waterfall approach. 21 Q. So I was going to ask, an advantage of RAD is that it 22 can get something working as quickly as possible, but it 23 relies on later iterating and replacing -- 24 A. Absolutely, and that -- it was also obvious that POCL, 25 as a customer, were not comfortable with that because it 58 1 meant that it wouldn't be fully functional. So you 2 would be putting out a series of releases over a period 3 of time and they wanted a fully functional system and 4 RAD was not the right way to do that. 5 Q. I think the basic prototype framework in an RAD isn't 6 used, ultimately, in the main build of the system; is 7 that right? 8 A. Well, nowadays it is, but I can't recall -- because RAD, 9 at that time, was quite an immature methodology, I was 10 uncomfortable with it as an individual, as a development 11 person, but I felt that it was still immature and 12 I didn't really understand how you got from A to B and 13 how you got a system that represented what the user 14 wanted. 15 I understood the mechanics but I didn't understand 16 how you achieved a product that would deliver to the 17 customers what they wanted. 18 Q. Am I right to infer from some of your earlier evidence 19 that you didn't have some faith in some of the technical 20 members of the team? 21 A. That's how it -- when I saw -- after requesting the task 22 force, I was extremely disappointed and upset that we 23 had ended up with a product which, on the face of it, 24 looked like it was pretty bad. So yes, that -- I was 25 then put into a very, very difficult position because, 59 1 going back to what we said earlier, it's easy to say 2 "rewrite the product", that's the easy option because 3 you can just stand back and say "rewrite it". That 4 doesn't necessarily mean you're going to end up with 5 something that's better, it just means that you have 6 said that's the right thing to do. 7 Actually trying to get a product and fix it and make 8 sure that it's stable -- and I genuinely believed that 9 we had done that, and so I think that I was comfortable 10 that the fact that we had taken that approach and we 11 succeeded in getting the product -- it had been a very 12 rocky ride, but we had achieved that objective, 13 I believed, genuinely, at the time. 14 Q. So what had happened to that original RAD product? 15 A. Well, that RAD product was then enhanced. It was quite 16 some way off what the functional requirement was needed 17 to be, so we took it on from that viewpoint onwards and 18 used it internally within the team. 19 Q. Did you have a final signed off design? 20 A. Yes, in the end, yes. We had to, as I said, reverse 21 engineer and we had to get some business requirements 22 that were signed off by POCL, and a design that was 23 signed off by POCL. 24 Q. I want to ask you about Post Office's awareness of 25 various issues. You have said at paragraph 32 of your 60 1 witness statement that POCL were made aware of every 2 defect in the ICL Pathway's solution? 3 A. As far as I was aware, they were. 4 Q. When you say "defect", do you mean Acceptance Incident 5 or do you mean more than that? 6 A. No, I mean more than that. Every PinICL that we were 7 going through that we had received, my understanding 8 was -- and I have no reason to think otherwise -- is 9 that we went through every one with POCL 10 representatives. 11 Q. So every PinICL was -- 12 A. Except for the Cs, a lot of the Cs, but even some of 13 those we went through as well, because every one the 14 category had to be agreed. 15 Q. The PinICLs or the Acceptance issues? I mean, PinICLs 16 are -- 17 A. PinICLs -- 18 Q. -- every incident -- 19 A. Yes, if it's agreed that it's an error, then the 20 category of that error needs to be agreed with POCL. 21 Q. I mean, the kind of document that we saw, the PEAK, was 22 your impression that those were being seen by POCL? 23 A. Well, they were POCL people that were represented within 24 the helpdesk, is my understanding. 25 Q. So it was through the helpdesk that you understood POCL 61 1 obtained their information on problems with the system? 2 A. Well, yes, being raised by the postmasters and any 3 incident -- and by POCL. Any incident raised by POCL, 4 whether it be the postmasters or the managers elsewhere 5 within POCL, would come and be raised as a PinICL 6 through the helpdesk. 7 Q. Was there some sort of policy in place between POCL and 8 ICL Pathway as to what level of information should be 9 formally shared or informally shared? 10 A. I'm not sure. 11 Q. Did you have any concerns about the sharing of 12 information with POCL? 13 A. In the early days, yes, because it was a PFI and, 14 therefore, we had our solution and, therefore, we were 15 there to deliver the solution in line with our -- the 16 requirements, so we wouldn't necessarily share that 17 information at that point. 18 Q. I will ask you about that stage shortly but, in terms of 19 when it was all up and running, it was your belief that 20 POCL had the level of detail that was contained in, for 21 example, the PinICLs? 22 A. Yes, yes. That was my understanding, yes. I had no 23 worries about them seeing the breakdown of the helpdesk 24 and the comments, and I believed that POCL were involved 25 in that process. 62 1 Q. What about the other way round? You have said that 2 Pathway wasn't aware of POCL's own systems? 3 A. No, not at all. 4 Q. Did that cause you difficulty? 5 A. I wouldn't say it caused me difficulty. It was a bit 6 like that -- if we had a problem with the TIP interface, 7 we -- we didn't know whether there were any issues with 8 TIP that was causing that. We could see -- when it was 9 our side, we could see what it was saying and we would 10 have to investigate and, quite often, we would find it 11 was an error with TIP. The same would apply to CAP in 12 the Benefits Agency, that we could see through the 13 interface that there was an issue. We didn't know what 14 the issue was, we just see we're having problems with 15 the interface. 16 Q. I mean, you were designing an end-to-end product so, 17 presumably, that was pretty crucial? 18 A. Agreed. 19 Q. Where did that failure lie then? 20 A. With the customers. 21 Q. Did you ask? 22 A. Many times. 23 Q. And what was their response? 24 A. They didn't feel that it was appropriate. 25 Q. Let's look at the invitation to tender. Your role in 63 1 the procurement process, I think you have said in your 2 statement, was to evaluate the system's requirements in 3 the invitation to tender to determine the resources that 4 were required; is that right? 5 A. That's correct. 6 Q. Part of that was presumably working out how that 7 end-to-end system could be achieved? 8 A. Yes. So I would have to look at the business 9 requirements, as defined in the ITT, and, again, it was 10 a PFI, so we had already developed a solution and were 11 the process of developing the solution to meet those 12 business requirements, and so my task was to make sure 13 that the product that we had was -- the solution that 14 Pathway had would be delivered within that time. That 15 was what my job was. 16 Q. We have seen -- I will give the reference but we don't 17 need to bring it up on screen, it is POL00031117 -- 18 Keith Todd produced a position paper in 1998, which said 19 that POCL couldn't reasonably have believed that the 20 Post Office's premises were fit for automation. Was 21 that your view? 22 A. Yes. It wasn't just my view, but it was a view that 23 I shared. 24 Q. Wasn't it part of your job at that stage, that 25 invitation to tender stage, to check things like that? 64 1 A. Well, that's how we found out, is by going in and 2 then -- with the implementation to go into the post 3 offices and do the surveys and it was as a result of 4 doing the surveys that we found that out. 5 Q. Wouldn't it have been obvious from quite an early stage 6 in the 1990s that many Post Office branches wouldn't 7 have had even computers? 8 A. No, it wasn't that, it was to do with things like power 9 points, it was to do with desks that were suitable, it 10 was to do with all the aspects of -- as far as the 11 surveys, the surveys were showing that a lot of the 12 premises were not fit to be able to automate in the 13 state they were in and required quite a lot of work in 14 order to make them such. 15 Q. So some blame has been placed on the Post Office for 16 that but it sounds as though you, ICL, were, in fact, 17 carrying out your own investigations as to the state of 18 fitness of the post offices? 19 A. Well, we had to do surveys, that was part of -- because 20 if there was any additional work to be done in the post 21 office, we were obliged to do it, so we went in to 22 survey and then we commissioned the work that was 23 necessary. 24 Q. Was that before the agreements had been signed with the 25 Post Office? 65 1 A. No, that was after. 2 Q. So wouldn't it have made sense to do them before? 3 A. We weren't allowed to. 4 Q. Were you involved in preparing readiness before the 5 contracts were signed and establishing what requirements 6 might be required? 7 A. We were -- we had no reason to suspect that where the 8 ITT wanted the system developed would be capable of 9 accepting that. We didn't find out until after the 10 award of contract when we started to commit resources to 11 the surveys and then the surveys were showing that quite 12 a few of the premises needed a lot of work doing to 13 them. 14 Q. Were surveys really -- I mean, couldn't you have looked 15 at some post offices yourselves before signing the 16 contract? 17 A. Well, we did, but there were 19,500 of them, so you 18 could only do a snapshot, if you were allowed to do so. 19 Q. And did you do a snapshot? 20 A. We didn't, no. As far as I know, we didn't, but I can't 21 recall. 22 Q. Why not? 23 A. I'm pretty sure we weren't allowed to. 24 Q. Could they have prevented you from going into post 25 offices? 66 1 A. No, they couldn't have done but they wouldn't have been 2 very happy if we had gone there. We would have had to 3 have sought permission from POCL to do that. 4 Q. Keith Todd said that there would have been a view taken 5 of the suitability of the post offices and he referred 6 repeatedly to a full-time team that ICL had. Were you 7 part of that team? 8 A. Well, the implementation, yes -- the implementation team 9 would have been the team that was looking at that, yes. 10 Q. Who was head of that team? 11 A. I can't recall his name because they changed two or 12 three times. 13 Q. At that stage, were you a senior member of that team? 14 A. I was the Programme Manager, of which that team was 15 part. 16 Q. Did you give thought at that stage to the fact that many 17 post offices would use a telephone line connection? 18 A. Yes, we -- that wasn't -- if that was the case, we knew 19 that wouldn't work. We had to find alternative 20 solutions to that. 21 Q. So did you foresee issues with ordinary phone lines 22 being used, such as interference, et cetera? 23 A. Absolutely. 24 Q. Did you raise those -- 25 A. Yes, we did, yes. We had to come up with different 67 1 solutions for those post offices where that was the 2 case. 3 Q. Can you tell us a little bit just what you recall of 4 consideration being given to those kinds of issues at 5 an early stage? 6 A. Well, we had to come up with a satellite option. We had 7 to come up with a different -- we had to get ISDN lines 8 into those post offices and we were given the impression 9 by BT that we could get ISDN lines into any of the 10 locations we needed to. It was only much later on that 11 we found that BT were struggling to do that. 12 Q. Were you concerned about those that didn't have ISDN 13 lines? 14 A. Absolutely, yes, and we had to find a way of trying to 15 automate them. 16 Q. Because what kind of problems would it cause if they 17 weren't? 18 A. Well, the connection wouldn't be a good one. It would 19 be -- it just wasn't suitable for a distributed IT 20 system. 21 Q. Peter Copping told us that ICL had told PA Consulting 22 that they had seriously underestimated the amount of 23 work that was involved. Would you agree with that? 24 A. Yes. 25 Q. Some witnesses have said that training was also 68 1 underestimated by ICL. Is that something you would 2 agree with? 3 A. I would. 4 Q. Again, I mean, training was something that presumably 5 you were considering at that invitation to tender stage? 6 A. Yes, it was. 7 Q. What steps did you take at the invitation to tender and 8 procurement stage to consider the level of training that 9 was required? 10 A. Well, we put forward our training. One of the issues 11 that we had with training was that -- and it remained 12 a problem -- personal problem, even though I wasn't 13 responsible for it in the latter part of the programme, 14 but I couldn't -- I couldn't see how you could roll out 15 a system at the pace that we were rolling it out and 16 also train people that had such a wide-ranging 17 capability and I could appreciate that some of the 18 postmasters and postmistresses would be horrified by the 19 implementation of an IT system, especially that one day 20 you're manual and the next day you're fully automated. 21 And I know that we looked at training these people 22 for a long period of time to give them as much 23 information as we could, but the issue was that, by the 24 time they were scheduled to be rolled out, they may have 25 forgotten it, which is quite reasonable. So we had to 69 1 bring the training so that it got very close to their 2 implementation, within a week or two, as my 3 understanding because I wasn't really into it -- 4 responsible for it, in order to overcome that issue. 5 But there was still the problem of basic training of 6 IT, never mind the system which was quite sophisticated 7 in what it did, and so how that was overcome and whether 8 it actually was overcome, I don't know. 9 Q. Do you think that training was underestimated at the 10 invitation to tender and procurement stages by ICL? 11 A. Yes. 12 Q. You have said in your statement at paragraph 5 that DSS 13 and the Post Office were blissfully unaware of how 14 unrealistic the timetable was. 15 A. Well, they put an indicative timeframe and the only way 16 that could be met was if the solution that we had got 17 and we had defined and we wrote it up and we put it in 18 a fully functional -- functional specification and said 19 "That's the system that we're going to deliver in that 20 timeframe". At that time there were 280 something 21 agreements to agree and we believed that CAPS was ready 22 to go and on that basis we demonstrated what our system 23 was capable of doing and we defined it in a functional 24 requirement. 25 What transpired was that that functional requirement 70 1 was never approved by BA or POCL. 2 Q. It has been alleged that you obstructed the PDA in 3 getting hold of certain information. It has been said 4 that you had said that the contract prevented you from 5 providing them with certain information. 6 A. Well, that was the PFI. I mean I was -- I was, if you 7 like -- not instructed but certainly advised by my peers 8 and by my managers and seniors that a PFI contract meant 9 that we were to be left to develop the system, that was 10 our risk, our responsibility and whether we were 11 successful or not would be proved by acceptance. 12 Q. And you have said at paragraph 10 of your statement that 13 DSS and POCL were not ready, managerially or 14 technically. Why do you think that? 15 A. Because they couldn't answer the questions we were 16 asking. So if they had been ready, technically -- they 17 knew that we were interfacing with their systems so they 18 should have people in place ready to front those 19 questions and answer them and if we were having 20 difficulties in saying -- we needed to define an 21 interface specification between ourselves and several 22 other systems that we were interfacing with. With the 23 ones with TIP and the ones with CAP we never got that 24 specification to the level of detail that we needed and 25 when we had issues we didn't have people that could help 71 1 us to resolve them. 2 Q. Do you think it would have helped to have shared more 3 information with the Post Office? 4 A. That was not in the nature of the contract, with 5 respect. You don't -- when you're carrying the risk 6 then the customer is taking a view that they wish to 7 transfer the risk to a supplier and that supplier then 8 defines the system in the way that they want to define 9 and therefore it's not a waterfall approach, it's not 10 a standard contract. 11 You needed a fully functional requirement in order 12 to do that. 13 Q. It may not have been part of the contract, but knowing 14 now what we know presumably more information sharing at 15 an earlier stage would have been beneficial, wouldn't 16 it? 17 A. Whether I believed it or not, it's not the nature of the 18 contract and I did have to adhere to my peers in that 19 way. It was not something that I could choose to do. 20 Q. Finally, you have also spoken about tensions between the 21 DSS and Post Office. Where were you getting that 22 information from and how did that impact on the work 23 that you were doing? 24 A. It was in meetings where they were -- both parties were 25 present and in memos and letters, if you like, in terms 72 1 of emails and such. You could detect that there was 2 a difference in their objectives and I suspected that 3 was because the Benefits Agency were looking for 4 alternative ways other than the Post Office for solving 5 their problems and I think the Post Office were aware 6 that that was a possibility. 7 Q. How did that impact the work that was going on at ICL? 8 A. At my level it probably didn't. It meant that I was 9 being pushed in a certain direction by BA, if you like, 10 as the senior partner, even though that may not have 11 been in the best interests of POCL and I had to try and 12 balance the two because as far as I was concerned I'm 13 there to try and satisfy both parties. 14 MR BLAKE: Thank you, Mr Austin. I am going to hand over 15 now to Mr Stein and Mr Henry. 16 Sir, do you have any questions before I do that? 17 SIR WYN WILLIAMS: No, no thank you. 18 Questioned by MR STEIN 19 MR STEIN: Mr Austin, I have a number of questions for you 20 that relates to the operation of the system and the 21 start up for it. My name is Sam Stein. I represent 22 a large number of subpostmasters, mistresses and 23 managers. 24 You were asked a number of questions by Mr Blake 25 that touched on the question of whether the branch 73 1 offices were going to be ready for automation. Forgive 2 me for perhaps being rather direct about this: wasn't it 3 blindingly obvious that some rural branch offices would 4 be totally incapable of automation from the beginning of 5 all of this? 6 A. Not to the extent that they were. 7 Q. Well, have you ever been to a countryside branch Post 8 Office in your entire life before you started working on 9 this project? 10 A. Well, they have a counter, don't they? 11 Q. Yes, but did you ever actually think about what the 12 implications were for small rural places that are going 13 to need to go from paper based into an automated base? 14 A. Absolutely. 15 Q. Right, so why was that not brought into the thinking in 16 relation to this project? 17 A. Well, it was as far as I was concerned. It was a major 18 issue. 19 Q. There appears to have been surprise from ICL that POCL 20 weren't aware of the possible demands of automation. If 21 you were aware of it and you were not surprised by it 22 and this was a matter that was concerning you, why 23 didn't you tell them to start off with? 24 A. We did. 25 Q. Right, so you're saying that you pointed out that some 74 1 of these offices were not going to be suitable and you 2 are saying that POCL just didn't listen; is that right? 3 A. No, sorry, I'm saying that training -- I thought you 4 were on training. 5 Q. No, I was asking about automation, I thought that was 6 clear. 7 Well, what's the answer? Are you saying that you 8 were perfectly aware of potential problems with branch 9 offices but you didn't bring it to the attention of POCL 10 or are you saying -- 11 A. Oh, yes, we did. Yes, we did. 12 Q. Right, so was that before contract was signed or after 13 contract was signed? 14 A. After the contract was signed. 15 Q. Right, why not before? 16 A. Because I wasn't given -- we weren't given the 17 opportunity to do that, as far as I can recall, and 18 I have to say we may well have done, I can't recall 19 that. I have to be honest -- 20 Q. This is a major issue, Mr Austin, isn't it? The 21 question of trying to put automation into small branch 22 offices in the middle of the countryside: a major issue 23 and a potential problem, yes? 24 A. Yes. 25 Q. Later on, it cost something like 40 million to fix the 75 1 problem, yes? 2 A. Yes. 3 Q. Why didn't your company point this out and say, "Look, 4 it's going to be obvious that this is going to be 5 difficult"? 6 A. We may have done but it wouldn't have been for me to 7 have done that. I can't recall, to be honest. I'm 8 being -- I'm sorry, I can't recall. 9 Q. You mentioned to Mr Blake the question of the 10 difficulties sometimes with the telephone line and then 11 the other types of lines that might be required to 12 assist with automation, yes? 13 A. Yes. 14 Q. Now, as we understand it, if the branch computer goes 15 offline within the Horizon System, there then needs to 16 be a reconciliation between the central servers and the 17 branch computer so that, essentially, they match; is 18 that right? 19 A. That's correct. 20 Q. So that this becomes vaguely comprehensible for 21 everyone, that means that one part of the system must 22 overwrite the other, so that there's consistency of 23 information between the two; is that correct? 24 A. The -- when the system comes back, the counter comes 25 back, it can be replicated from the centre. So it can 76 1 be brought back up to where it was when it failed. 2 Q. Right. So what happens is that, if the system goes 3 offline, and it could go offline because its been turned 4 off, correct? 5 A. Yes. 6 Q. It could go offline because environmental factors have 7 caused an interruption in power supply? 8 A. Yes. 9 Q. It could go offline because of problems with the cabling 10 or something similar, yes? 11 A. Yes. 12 Q. So a variety of reasons could cause it to go offline, 13 yes. Now, that doesn't mean that, necessarily, when it 14 does get back in contact, that the two parts of the 15 system are going to have the same information by that 16 point, does it? 17 A. Not necessarily but it was designed to do so. 18 Q. In fact, it was designed, as we understand it, so that 19 the computer in the branch could, even if it was 20 offline, continue to provide service to customers; is 21 that correct? 22 A. If it was up and running, yes. 23 Q. If it was up and running. So, despite the fact that it 24 may not still be connected to the Horizon main system, 25 it can still provide customer support; is that correct? 77 1 A. Correct. 2 Q. So when back online the plan was that the systems would 3 then reconcile one to the other? 4 A. Yes. The counter would reconcile back to the 5 correspondence servers, yes. 6 Q. Right, for obvious reasons, so that actually, in the 7 end, both sides of the system would end up with the same 8 information about transactions? 9 A. Yes, yes. 10 Q. Correct, right. How much of the code or the technical 11 specification for the software for the Horizon System 12 was available to you and your team? 13 A. Because the -- okay, you are referring to the Riposte 14 software that was responsible for keeping those message 15 stores in -- 16 Q. Yes. 17 A. We were not allowed to see the code. 18 Q. Were your team capable of rewriting the code if it was 19 required? 20 A. Yes, if -- our team wouldn't have been, but we would 21 have had to commission someone to do so. 22 Q. So in order to deal with any Riposte system difficulties 23 with communication and communication errors, you had to 24 go to the Riposte system people; is that right? 25 A. Yes, my chief architect would do that. 78 1 Q. What was the cost to the Post Office of doing so, if you 2 went back to Riposte? 3 A. None, as far as I'm aware. 4 Q. What about the cost to ICL: was there a cost there? 5 A. Sorry, I'm not sure about the question. 6 Q. Financial cost, was there a financial cost? 7 A. In going back to? 8 Q. To Riposte? 9 A. To change something? 10 Q. Yes, to get support or get the engineers to come through 11 or to get the code people to come and fix something? 12 A. Yes. 13 Q. Yes. Is that code still available now, so if we wanted 14 to check the system code right now, go back in time to 15 what was being used for what we call Horizon Legacy, is 16 that available? 17 A. I can't tell you. It was lodged in escrow. 18 Q. I'm going to ask you some questions about the original 19 agreement and that's going to require putting up on the 20 screen FUJ00000071. You should have on your screen, 21 Mr Austin, the first page of that. 22 I'm very grateful, Frankie. 23 As you can see, Mr Austin, this is just a reminder, 24 this is Post Office Counters Limited and ICL Pathway 25 Limited, the "Information Technology Services Agreement 79 1 for Bringing Technology to Post Offices", and this is 2 the codified agreement. This is dated, as we understand 3 it, from the system in mid-1999, okay? So this is the 4 basic agreement between the parties, all right? 5 Now, I'm going to take you to a particular part of 6 this, if I may, please. Can we go to page 97 of 914 on 7 the Relativity system, sir, for your reference. 8 I'm going to read out this particular section, 9 Mr Austin, so that you have a moment to think about what 10 it says here and, therefore, from your point of view, 11 what it means. So this is under the heading 12 "Prosecution support": 13 "The Contractor shall ensure that all relevant 14 information produced by the POCL Service Infrastructure 15 at the request of POCL shall be evidentially admissible 16 and capable of certification in accordance with the 17 Police and Criminal Evidence Act (PACE) 1984 and the 18 Police and Criminal Evidence (Northern Ireland) Order 19 1989 and equivalent legislation covering Scotland." 20 At the next paragraph, 4.1.9: 21 "At the direction of POCL, audit trail and other 22 information necessary to support live investigations and 23 prosecutions shall be retained for the duration of the 24 investigation and prosecution irrespective of the normal 25 retention period of that information." 80 1 So we can see that the heading for all of this is 2 "Prosecution support" and then between paragraphs 4.1.8 3 and 4.1.9 is basically saying that information's got to 4 be provided to the Post Office at the request of the 5 Post Office and it has to have certain evidentially 6 admissible requirements, and then the other part is 7 basically saying that, well, you need to keep the 8 information so it's available. 9 Just help us a little bit, please, with your 10 understanding of that. What was your understanding of 11 what would be evidentially admissible and capable of 12 certification in accordance with the Police and Criminal 13 Evidence Act 1984 at the time? 14 A. I'm not able to say that because that wouldn't have been 15 my responsibility. I'm not knowledgeable enough to know 16 what that -- would be required. That would be one of my 17 colleagues in the audit department there that would 18 have -- Martyn Bennett and Jan would have liaised with 19 the audit people within POCL to determine exactly the 20 nature of what they required. 21 Q. Who within ICL was responsible for making sure that ICL 22 was capable of fulfilling these prosecution support 23 requirements? 24 A. Martyn Bennett, in my opinion. 25 Q. Martyn Bennett. Help us, please, whether you are aware 81 1 that ICL sought or did not seek an opinion from 2 an experienced criminal lawyer as to what this all 3 means? 4 A. I can't -- I can't -- 5 Q. Again, are you referring that to Martyn Bennett? 6 A. I am, yes, because I wasn't responsible for that aspect. 7 Martyn would have come to me and said "This is what we 8 need to do", as he did when we produced the audit trail 9 capability. 10 Q. Right, well, let's just deal with this a little bit 11 more. What systems were you aware of that were put in 12 place, so that ICL could fulfil this requirement? 13 A policy, guidance, protocol -- 14 A. I wasn't requested to do that. 15 Q. Well -- 16 A. I was requested to put in an audit system, which we did, 17 which was to enable POCL to audit the system at various 18 points throughout, from the correspondence server right 19 through to the MIS, the back office system. 20 Q. Audit is not the same, is it? It's not the same as 21 providing information, which is evidentially admissible 22 and capable of (unclear) -- 23 A. I was never -- that never came across my desk. 24 Q. Then why are you referring to audits? 25 A. Because the next sentence, which you referred to, which 82 1 is the direction of POCL's audit trial. 2 Q. All right but what protocols or systems were put in 3 place for the first part, 4.1.8; do you know? 4 A. No, I don't. 5 Q. No. Your job, as we understand it, if I recall from 6 your statement, is that this is meant to be dealing with 7 the implementation of the system, getting it up and 8 running, getting it going, getting it going to the 9 acceptance requirements, yes? 10 A. Yes. 11 Q. Right. 12 A. And to deliver to a business requirement specification 13 as produced and signed off by POCL. That was never 14 anything that I ever saw. 15 Q. Well, this is part of their business requirements, 16 Mr Austin. 17 A. Not as I saw. 18 Q. Why not? 19 A. I don't know. 20 Q. Do you understand the importance -- 21 A. Absolutely, I understand the importance -- I absolutely 22 understand the importance but nobody said or asked me, 23 or laid down in writing what was required in order to 24 meet that requirement. That's in a contract, with 25 respect, that's not in a business requirement. 83 1 I delivered things to a business requirement. 2 That's a document that says "These are our business 3 functional requirements". 4 Q. Right. So you delivered things to, let me see, the 5 acceptance criteria, that was what you were aiming at? 6 A. No, the set of business requirements. The acceptance 7 criteria was based on the business requirements. 8 Q. Right, okay. So the business requirements, as far as 9 you can say and recall, did not include a reference to 10 the prosecution support section? 11 A. No. 12 Q. I see. If we can just scroll down the page a little bit 13 more, it says at 4.3 "Training and Training Material". 14 Frankie, if we just go down to 4.3, thank you very 15 much. 16 Now, there's a reference to: 17 "The Contractor shall prepare all training events 18 and training material ..." 19 Can you help, from your recollection, with what 20 training events and training material was even 21 considered to try and provide the prosecution support 22 that we have just been looking at? 23 A. No, that was the responsibility of someone else. 24 Q. Mr Bennett? 25 A. No. 84 1 Q. Who else? 2 A. I can't recall, but it was within Mike Coombs' area. He 3 would be able to say who was responsible for doing it. 4 Q. I see. 5 A. It was a -- training and implementation were not part of 6 my responsibility. 7 Q. If we can go one step down, please, Frankie, just 8 a little bit further down that page. Yes, at 4.4.2 you 9 see there that: 10 "The Contractor shall ensure that all materials 11 which are used for producing documentation relating to 12 POCL services, conform to relevant parts of POCL's 13 Environmental Policy Summary." 14 Well, obviously we would all agree that that should 15 be done. Do you recall whether there was any summaries 16 or policies provided to you from POCL as regards the 17 operation of the prosecution support section? 18 A. No. 19 MR STEIN: One moment please. 20 (Pause) 21 Thank you, sir. 22 SIR WYN WILLIAMS: Thank you. Could that document be taken 23 down, please. Thank you. 24 MR BLAKE: Sir, I believe Mr Henry is now going to be asking 25 some questions. There is one document that -- 85 1 SIR WYN WILLIAMS: Sorry, Mr Blake, can you stop. I can 2 hardly hear you. 3 (Pause) 4 MR BLAKE: Sir, there is one document that Mr Henry will be 5 taking the witness to that the witness hasn't had 6 an opportunity to look at before, it is going to be 7 brought up on screen. 8 Mr Austin, if you need any more time on any of the 9 documents, please do say so and we can take some more 10 time on that. Thank you very much. 11 Questioned by MR HENRY 12 MR HENRY: Hello, Mr Austin. Yes, no desire at all to 13 ambush you so please do follow Mr Blake's suggestion, 14 but could we go to FUJ00036863, please. 15 Now, sir, is this a document that you have seen 16 before? 17 SIR WYN WILLIAMS: Sorry, is that addressed to me, Mr Henry, 18 or Mr Austin? 19 MR HENRY: Well, may I, like the duck/rabbit, ask you both, 20 sir. 21 SIR WYN WILLIAMS: Well, I can't put my hand on my heart and 22 say one way or the other. I have seen so many 23 documents, Mr Henry, but I'm very happy for you and 24 Mr Austin to discuss this document and I will do my best 25 to follow. 86 1 MR HENRY: Thank you very much, sir. 2 Mr Austin -- 3 A. I have not seen this document. 4 Q. Well, look, Mr Austin, I don't want to take you by 5 surprise and I'm going to make it absolutely clear to 6 you that if you need time to think, please do so, but 7 this is produced, obviously from Fujitsu. It looks like 8 it is a sort of helpdesk file because John Simpkins 9 opens up the entry and it is talking about calls. So 10 would you be prepared to agree with that? 11 A. Yes. 12 Q. Good. It looks like it was opened, if we go to the very 13 top of the page, just underneath the grey headline, 14 "Opened", 9 December 1999 at 11.00 am. 15 A. Yes. 16 Q. And last update, 21 February 2000, and so there's a call 17 which we can see at 11.00 am. 18 Then could we go down a few lines and we've got at 19 11.00 am again, John Simpkins -- it's about six or seven 20 lines down: 21 "EPOSS transactions. The EPOSS problems look to be 22 related to Existing Reversals (often of the settlement 23 line)." 24 So that looks like it's a balancing the books 25 problem, does it not? 87 1 A. It looks like that. 2 Q. Thank you. I'm going to ask you now to go to page 2, 3 please, the lady who is assisting us, and we can see 4 an entry at line 4 from Mr Steve Warwick. It's 9.28.14 5 on 10 December, and it says: 6 "Passing to EPOSS-FP for urgent analysis. This call 7 is related to AI376 and will require resolution before 8 the commencement of Rollout in January." 9 You see that? 10 A. Yes. 11 Q. We know what 376 was and obviously that would have been 12 a major -- 13 A. Absolutely. 14 Q. Quite so. Could we now go to page 5 please. At page 5 15 there is a Francesco Chiarini. Do you remember 16 Mr Chiarini? 17 A. No. This is when my development team and a lot of my 18 technical experts had moved over into the support 19 environment. 20 Q. Yes. 21 A. So they would have been fronting and going through that 22 process. 23 Q. So they would have been customer faced? 24 A. Yes. 25 Q. Right, okay. 88 1 A. And the support teams, the actual development support 2 teams, the expertise, for example Steve Warwick -- and 3 I don't know who Francesco Chiarini is but he would be 4 someone, I would assume, would be in that environment. 5 Q. I see, thank you. So we've got here, haven't we, at 6 11.37.17, and it says, forgive me: 7 "FAD322704: Problem in the Scales transaction. 8 A fix was delivered in WP5447 on 20/8/99, but the 9 counter had a previous version ..." 10 Now is this -- and I don't ask you to speculate, but 11 is this an instance where perhaps, unknown to people 12 trying to fix things, a fix had already been applied, or 13 an update hadn't yet taken place and so therefore the 14 fix was not effective? "A fix was delivered in WP5447 15 on [20 August '99] but the counter had a previous 16 version", and then it gives WP4775. The "but" seems 17 significant, doesn't it? 18 A. What it is suggesting is that the fix was -- didn't 19 arrive at the counter, the right version of the fix 20 didn't arrive at the counter, that's what it's 21 suggesting. 22 Q. Right, thank you. Then we've got, below there -- I'm 23 going to just say FADS, about five of them: 24 "The problem has been identified and can be 25 reproduced." 89 1 Then I omit words: 2 "The other reversals are being investigated." 3 Is this again where the books are not balancing 4 reversals? 5 A. That's what it would appear to be, yes. 6 Q. Then if we just move down the page to Mr Chiarini at 7 14.30.27: 8 "Further investigation has not been able to get to 9 the root of the problem." 10 You see that? 11 A. Yes. 12 Q. So this was obviously, as you have said several times, 13 very, very difficult, complex -- complex, quite unique 14 problems arising. 15 A. A combination of circumstances, yes. 16 Q. Yes. But, obviously, this is suboptimal because you've 17 got your brightest and the best, subject to the 18 qualification you gave in evidence, trying to sort this 19 out? 20 A. Yes. 21 Q. Can we go, please, to page 7 and this is at 17.45, it 22 says: 23 "Deleted User (Mark McGrath left [July]/00). I have 24 released a fix for this in WP7029 -- to be applied AFTER 25 7012." 90 1 Can you work that out? Do you know what that means? 2 A. No, I'm afraid I can't. 3 Q. Because you see the name "Austin", and maybe this is 4 a coincidence, appears just after that, doesn't it? 5 A. It does and I can't explain that because I had nothing 6 to do -- directly to do with this exchange or in the 7 fixing, because this was a team not reporting into me. 8 Q. I see. So -- 9 A. So I don't know why "Austin" is there. 10 Q. I see. You can't -- I mean, would this team, even 11 though they weren't reporting into you, given your 12 seminal role in development, might they not have been 13 consulting with you? 14 A. No. 15 Q. No? Not at all? 16 A. Not at all. 17 Q. "The response was delivered on the system", that's at 18 17.46, and then could we go, please, to page -- forgive 19 me while I just scroll it up. 20 Forgive me, sir, I'm just having difficulty with my 21 one. 22 Yes, could we go to page 8, please. Do you see at 23 2 February 2000 at 18.17.25: 24 "Please inform me when this is released to live." 25 Can you remember what that was about? 91 1 A. I'm sorry, I'm just trying to find it. 2 Q. That's John Simpkins, it's about six up from the bottom? 3 A. Yes, yes. 4 Q. "John Simpkins notified as requested. The fix has been 5 released to live." 6 Then finally, over the page, I'm going to ask you 7 now about Colin Baker. Can you help me, please, about 8 Colin Baker, who is he? 9 A. I don't know. 10 Q. Sir, you will see it is about the last eight/nine lines. 11 21 February 2000: 12 "We are unable to test this in PI Test but we are 13 NOT aware of the problems described as occurring in our 14 environments." 15 Then this: 16 "I can only suggest that we be aware of potential 17 problems and raise them if and when they occur. 18 "The Call record has been transferred to the Team: 19 QFP." 20 Do you know what that team was? 21 A. No, sorry. 22 Q. I mean, this does not necessarily appear to be as 23 coherent as what you suggested in evidence, that this is 24 a somewhat as and when, ad hoc response to problems 25 rather than a coherent policy of investigation and 92 1 weeding out and rectification? 2 A. It does appear like that, I agree, but this was part 3 of -- as I said, the development and the testing and 4 error fixing was within the customer services arena. 5 Q. Absolutely, and it would be obvious, wouldn't it, that 6 the Post Office would be aware of this as well? 7 A. Yes, I would assume so, yes. 8 Q. Absolutely. Could we just consider now, in the time we 9 have -- and I will finish before lunch, sir -- Horizon, 10 for POCL's specific needs, was a frantic afterthought, 11 wasn't it? 12 A. No, no. The reason it may appear like that is because 13 we didn't have a documented functional requirement. 14 Q. You mention that because you thought that it was 15 apparent from both the DSS and POCL that they thought 16 that they could just give you the functional 17 requirements before it was supposed to go initially live 18 in 1997. You say that in your statement? 19 A. Absolutely. It became apparent to me that they didn't 20 understand the basic philosophy of a PFI, in that you 21 write the specification, you then get a quote and then 22 you deliver to what you said you would deliver and the 23 functional requirement. And that's the reason why, in 24 the contract that we gave to BA and POCL, in the 25 contract that was signed, it said that within 30 days of 93 1 the contract being signed they would sign off the 2 functional requirement. 3 Q. Right. 4 A. Because that's what we had. 5 Q. Let's hold that because what you said there is very, 6 very important but I want to address a more fundamental 7 problem as to why I suggest it was a bit of a frantic 8 afterthought. You, of course, worked for ICL from 1995 9 to October 2000? 10 A. Correct. 11 Q. But, obviously, Pathway and Horizon had been in 12 development before you joined the company? 13 A. Pathway bid team was in before I joined the company, 14 yes. 15 Q. Fine. But I want to ask you about this, that the 16 Benefit Payment Card was scrapped in May 1999, wasn't 17 it? 18 A. It was. 19 Q. Yes, and POCL, of course, wanted to continue with the 20 Benefit Payment Card but the DSS wouldn't support them? 21 A. Is what I understand to be the case. 22 Q. Yes. ICL were initially supportive of POCL's position 23 but, eventually, you know, had to accept the inevitable, 24 as did POCL, that the BPC was dead in the water, 25 correct? 94 1 A. Correct. 2 Q. Absolutely. So that's what I suggest is the reason why 3 it was a frantic afterthought because what effect did 4 that have on the time you had available to address what 5 POCL specifically required? 6 A. The remaining time that we had to get a release is that 7 we were agreeing with POCL what we would try and deliver 8 by what dates and in order to do that we needed POCL to 9 provide a functional requirement for us -- 10 Q. May we go -- 11 A. -- and to sign off the designs. 12 Q. I'm so sorry, sir, I didn't mean to cut across you. I'm 13 just anxious about time because I don't want you to be 14 here after lunch. 15 Could we quickly go to your witness statement, 16 please, at paragraph 8, your witness statement -- 17 forgive me, I had the number to hand but I don't think 18 it should be ... thank you so much. If we go to 19 paragraph 8, please. Thank you. Do you see the last 20 three lines? 21 A. Yes. 22 Q. You're dealing, first of all, with the DSS, Benefits 23 Agency and the rejection of the BPC. Then this: 24 "There was very little functionality introduced for 25 POCL consequently it has limited relevance to [the 95 1 Inquiry]. In fact, when the DSS made the decision to 2 cancel the project in May 1999, all the relevant card 3 software was removed from the ICL Pathway solution. The 4 POCL Horizon project only came into existence in the 5 spring 1998 and a new agreement with Post Office 6 Counters Limited signed in July 1999." 7 In fact the Benefit Payment Card was scrapped 8 in May 1999, wasn't it? 9 A. That -- yes, I have put May 1999: 10 "In fact, when the DSS made the decision to cancel 11 the project in May 1999~..." 12 Q. Exactly. So very little functionality for POCL, and you 13 give the explanation in paragraph 9, because it really 14 was because DSS was the dominant partner and the Benefit 15 Payment Card was the priority and so therefore the BPC 16 had taken precedence. Do you see your exact words: 17 "There was no doubt that the DSS were the dominant 18 partner, and the benefit payment functionality took 19 precedence over the ... EPOSS functionality which would 20 be developed in parallel over a longer period using 21 [the] iterative development approach." 22 A. I am referring to the Initial Go Live. 23 Q. Yes. 24 A. I'm not referring to the project as a whole. I'm trying 25 to say that the Initial Go Live had little relevance 96 1 because it was mainly BA functionality that we 2 delivered. 3 Q. But when you say that the DSS material was stripped out 4 of the project, would that have included the work on the 5 fraud management system because were you aware that 6 a major component, the fraud management system, was 7 removed from the contract that POCL eventually signed? 8 A. I can't be sure. I have not heard of the term "Fraud 9 management system". If you're saying the fraud 10 procedures -- we were instructed to remove all the code 11 that was to do with the benefit card system. 12 Q. So it may have been -- and I'm not asking you to 13 speculate, but it may have been that all of this 14 anti-fraud architecture was connected to the benefit 15 card and when the benefit card was scrapped, then all of 16 that architecture went as well? 17 A. I understand the point you're making and I'm trying to 18 cast my memory back, is that we were instructed that we 19 must remove all the benefit card functionality from the 20 system, which was quite a large task and really did 21 demotivate my team because -- 22 Q. It demotivated? 23 A. Very much so because they had spent quite some time 24 developing that. 25 Q. Absolutely. They had invested, as had the company, time 97 1 and effort and so therefore, at the very time that 2 you're trying to actually produce something for POCL, 3 your team are obsessed and under the cosh at actually 4 doing completely dead work, removing all of that 5 architecture? 6 A. There were two separate teams. The people that were 7 removing the BPS, all the functionality for that, were 8 not the same people that were developing on the POCL 9 side. It was, as I said, being run in parallel. So I'm 10 referring to the fact that we literally had to kill all 11 the software and all the work we had done on the 12 benefits payment system and then run in parallel and put 13 all our resources into EPOSS. 14 Q. So this brings me to my final point and this was the 15 point that you raised earlier, in fact, about "They 16 didn't understand the PFI", et cetera, et cetera, 17 because this actually turned from being a PFI to 18 a turnkey contract, didn't it, effectively? 19 A. My colleagues who were responsible for that negotiation 20 and the agreement that came to pass and, yes, it became 21 more traditional. 22 Q. Yes, absolutely. So this was, as you anticipated in 23 fact at the beginning of your statement at 24 paragraph 5 -- you know, you thought it was an 25 incredibly ambitious timescale and you thought, really, 98 1 from the beginning it ought to have been a turnkey or 2 off-the-shelf? 3 A. I did, that was my personal view. 4 Q. But of course turnkey or off-the-shelf doesn't mean 5 bespoke, does it? 6 A. No. 7 Q. It doesn't even mean made to measure, does it? 8 A. No. Well, it means we developed -- we had got 9 a product, the Riposte product, and then we developed 10 the BPS aspect of it and that was live and ready to go. 11 Q. But it's very much working out compromises and make do 12 and mend, isn't it? 13 A. We still had to deliver to the business functionality. 14 We still -- we were guided -- we had to develop a system 15 that met the requirement. 16 Q. I know, but -- 17 A. The business requirement. 18 Q. Unfortunately however, as is plain, it didn't. 19 A. Which bit? 20 Q. Well, I mean you even said to Mr Blake, you didn't 21 understand how you could achieve a product which would 22 deliver to the business what they wanted. I mean you 23 were doing your best, but it had, as it were, morphed so 24 many times -- 25 A. That's true. I mean, from a development team, we were 99 1 getting hit by functionality we had never seen before. 2 Q. Exactly. 3 A. And it was growing at a massive rate and my team was 4 growing at a massive rate. 5 Q. Right, so it comes to this and this is, I promise you, 6 my final question: you have stated, and Mr Blake has 7 taken you through it, that paragraph 22 of your 8 statement -- no need to bring it up, paragraph 23 -- you 9 were aware of bugs, errors and defects. July 1997, it 10 wasn't sufficiently robust, you bring in Escher and 11 Riposte, correct? 12 A. Correct. 13 Q. But paragraph 26, September 1999/October 1999 -- no need 14 to bring it up on the screen, FUJ00079782 -- I mean, 15 post-Escher and Riposte, it still, quite frankly, is not 16 fit for purpose, is it? 17 A. It didn't go to Escher for that reason. It went to 18 Escher for a specific reason, in that the product, 19 EPOSS, was not interfacing with Riposte correctly. 20 Q. Right. 21 A. So the rest of the EPOSS product was as is. Escher 22 didn't touch it. 23 Q. Leaving that aside, September/October 1999 -- 24 FUJ00079782 -- it was still not fit for purpose? 25 A. No, and we had to put substantive corrective actions in 100 1 place. 2 Q. But the substantive corrective action, as you say in 3 your statement, was only to concentrate on 80 per cent 4 of the bugs, errors and defects? 5 A. No, no. 6 Q. Well, let's go to your statement. 7 A. No, I did say that. I just want to clarify. What we 8 said was that we could stabilise the product. We 9 couldn't fix everything, we could stabilise the product 10 by targeting the 80 per cent because 80 per cent of the 11 bugs and defects were due to a particular area of the 12 system and we attacked that. 13 We also attacked the others, but -- and fixed a lot 14 of problems in those areas, but the instability issue 15 and the not fit for purpose was mainly due that people 16 were seeing so many large amount of errors that they 17 were using that as a yardstick to determine whether the 18 project was fit for purpose or not. 19 So if you fixed the 80 per cent, you were then left 20 with the 20 per cent, of which a lot of them you did fix 21 but, overall, the product was becoming more and more 22 stable. 23 Q. But, nevertheless, given the imminent -- well, given the 24 rollout and given what we now know, it was 25 a catastrophically dangerous thing to have happened 101 1 because of the number of unknown unknowns and the 2 factors you have already said, that when you were fixing 3 things, you were sometimes introducing further problems, 4 unintentionally and unwittingly? 5 A. Well, when you do that, that's what regression testing 6 was due to find and it did, so if that happened -- you 7 couldn't be 100 per cent, but regression testing was 8 there to try and avoid that happening. 9 Q. But all the time the Post Office was aware of this as 10 well: the level of errors -- the level of bugs, errors 11 and defects? 12 A. I believe so. 13 Q. Thank you. 14 A. Because I -- sorry, just to -- 15 Q. Yes, please. 16 A. I didn't specifically hide that from POCL. I wasn't -- 17 as far as I was concerned, acceptance and the errors and 18 defects we were having, whether it was under the 19 instability issue which I have discussed prior to that, 20 which was very difficult to crack, 298, and then 376 21 which was the balancing issues, we took those extremely 22 seriously and if we would have not solved those, we 23 wouldn't have achieved acceptance. 24 Q. Who was your counterpart at POCL then, if you didn't 25 hide it from POCL? 102 1 A. I didn't have a counterpart. 2 Q. You didn't? But you -- 3 A. The nearest person that would -- if you like, that 4 I will be talking with would be John Marr. 5 MR HENRY: John Marr. Thank you very much, sir. 6 MR BLAKE: Sir, can you hear me? 7 SIR WYN WILLIAMS: Yes. 8 MR BLAKE: Thank you. I just have one very brief follow-up 9 question and also Mr Maloney is going to ask a question 10 or two as well. 11 MR MALONEY: Sir, no. We have no questions. We have 12 provided very detailed rule 10 requests and, if we may 13 say, with respect, Mr Blake has very ably covered all 14 the questions we identified in respect of this witness. 15 So we have no questions. 16 SIR WYN WILLIAMS: That's fine. 17 Further questioned by MR BLAKE 18 MR BLAKE: Thank you very much. So it's just one question 19 from me then. 20 We have seen the PEAK and PinICLs and I think your 21 evidence was that you believed the Post Office had 22 access to those. Would it surprise you if, in fact, 23 they didn't have access to those? 24 A. I believed that POCL were -- had people that were part 25 of the helpdesk community so, from that perspective, 103 1 I can't believe that they weren't given access to that 2 system. I don't know whether they were able to 3 contribute to it, but I can't believe they weren't given 4 access to it. They were part of the same team of 5 experts. 6 Q. Would it have been best practice for them to have had 7 access? 8 A. I would -- yes. 9 MR BLAKE: Thank you very much. Sir, I don't have any 10 questions unless you do, sir? 11 Questioned by SIR WYN WILLIAMS 12 SIR WYN WILLIAMS: I just wanted to ask Mr Austin, there was 13 reference in one of the emails which Mr Blake showed you 14 to a man called Gareth Jenkins. 15 A. Yes. 16 SIR WYN WILLIAMS: Was Mr Jenkins part of your team? 17 A. Not that I can remember. 18 SIR WYN WILLIAMS: Did you know him at all? 19 A. Not that I can recall. 20 SIR WYN WILLIAMS: Fine. Thanks very much. 21 Thank you very much for making your statement, 22 Mr Austin and also for coming to give oral evidence. 23 A. Thank you. 24 MR BLAKE: Thank you. 2.00, please, sir. 25 SIR WYN WILLIAMS: Certainly, yes. 104 1 (1.02 pm) 2 (The luncheon adjournment) 3 (2.00 pm) 4 MS KENNEDY: Good afternoon, sir. Can you see me and hear 5 me? 6 SIR WYN WILLIAMS: I can indeed. 7 NEW SPEAKER: Our next witness is John Bennett. 8 JOHN BENNETT (sworn) 9 Questioned by MS KENNEDY 10 MS KENNEDY: Mr Bennett, I am Ruth Kennedy and I ask 11 questions on behalf of the Inquiry, as you know. 12 Could you confirm your full name, please? 13 A. John Hamish Bennett. 14 Q. In front of you, you should have a copy of the witness 15 statement that you prepared; do you have that? 16 A. Yes, I do. 17 Q. Have you read through this statement recently? 18 A. Yes, I have. 19 Q. If you turn to page 10 of that statement -- 20 A. Yes. 21 Q. -- is that your signature there? 22 A. Well, the one I've got here hasn't been signed -- yes, 23 it is, I can see it here. Yes, it has been signed; that 24 is my signature. 25 Q. Is it true to the best of your knowledge and belief? 105 1 A. Yes. 2 Q. I'm going to start by asking a few questions about your 3 background. Please could you explain your background 4 prior to joining ICL? 5 A. I -- when I graduated I joined a company called English 6 Electric-Leo-Marconi, which eventually became part of 7 English Electric and therefore picked up what's in 8 paragraph 1 here. 9 Q. What did you initially do when you arrived at ICL? What 10 roles did you have? 11 A. Sorry, what? 12 Q. Roles. 13 A. Rules? 14 Q. Roles. 15 A. Roles, I apologise. 16 Q. Excuse my Northern Irish accent. 17 A. Yes, I started my career as a computer programmer and 18 progressed from that into systems engineering and 19 eventually into sales. Then through sales management 20 and then followed the track which you have here in 21 paragraph 1. 22 Q. What role did you have between 1994 and 2000? 23 A. 1994 and 2000. In 1994, I took over the responsibility 24 for managing the ICL team, which was bidding for the 25 BA/POCL PFI contract, so it started as a pre-sales 106 1 exercise. It ran -- took about two years for the 2 contract to be eventually let. The contract was 3 eventually placed with ICL. I continued to work on that 4 contract. 5 The company was reformed as ICL Pathway in -- 6 Q. I think that happened -- sorry, I was just going to say 7 I think that happens in 1995 -- 8 A. I'm sorry, yes. 9 Q. -- that ICL Pathway is set up. You then take over as MD 10 of that company in 1995, in the newly set up company? 11 A. Yes, that's correct. 12 Q. Can you briefly explain, in your view, what the purpose 13 of what would become the Horizon project -- what the 14 purpose of the project was? 15 A. Well, it started, of course, with two major customers, 16 two major clients, and they both had somewhat separate 17 but joint agendas -- well, they joined up, but the 18 essential feature for Post Office Counters was to fully 19 automate their extensive network of post offices, which, 20 until that time, was still running on mainly manual, 21 paper process. So they needed to automate their entire 22 network, for one thing to deliver their existing 23 services but, more importantly, to build a platform for 24 future business within the Post Office network. 25 The Benefits Agency, which were, of course, a -- and 107 1 had been for a long time -- a customer of the Post 2 Office Counters needed to develop a new and more 3 automated system for the payment of benefits and the 4 contract called for a Benefit Payment Card to achieve 5 that, and the key driver there was a more efficient 6 system but perhaps, even more importantly, a system 7 which had a substantially lower element of fraud. 8 So the two customers were joined together to both 9 deliver their separate business objectives. 10 Q. At the time of the procurement exercise, did you feel 11 that ICL Pathway could deliver and meet those two aims? 12 A. When the business -- when the sales campaign started and 13 from all the invitations to tender and everything, yes, 14 I was confident that we could deliver a system to meet 15 this requirement. 16 Q. How important was training in respect of that vision? 17 Did you see that as an important element of the project? 18 A. A key feature of the programme, particularly for Post 19 Office Counters -- not so much the Benefits Agency, but 20 for Post Office Counters was the training of, I think, 21 in excess of 30,000 staff to use the new system. And 22 many people referred to the programme as one of the 23 biggest management of change programmes in the UK and 24 perhaps in Europe and, as a management of change, it did 25 affect the working practices of an extremely large 108 1 number of people, so training was a key feature of that 2 programme of change and we were responsible within the 3 contract for delivering training facilities to all those 4 numbers of staff I have just referred to. 5 Q. I'm now going to take you to some board minutes of 6 a meeting on 3 October 1995. If we could turn up 7 FUJ00077832, please, and if we could turn to page 3 8 please. You can see that this is a note of the board 9 meeting on 11 October 1995. If we turn to page 3 and at 10 (d), the top (d), it says: 11 "As noted at the last meeting, we were still 12 experiencing considerable difficulty with the way the 13 procurement was developing. POCL/BA were attempting to 14 rewrite the SSR [which is the statement of service 15 requirements] via detailed contract schedules, then 16 would implement change control so that 'level playing 17 field' would be achieved with only one substantial 18 variable left upon which to make a final decision 19 between the three shortlisted suppliers, namely price." 20 At this stage, the three shortlisted suppliers were 21 Cardlink, IBM and Pathway; is that right? 22 A. Correct. 23 Q. In relation to price, what did you mean when you said 24 that the final decision would be made on "namely price"? 25 A. Well, it was a competitive international tender and once 109 1 all three shortlisted suppliers were compliant then 2 price would be, in my reckoning, the predominant 3 criteria for selection. 4 Q. Did you feel it was important then to be the cheapest 5 option in order to win the contract? 6 A. There was no way of judging whether we were the cheapest 7 bidder. We had no access to the costs and prices of our 8 competitors' bids. We could only price our bid at the 9 best level we could, compatible with delivering the 10 contract and making a return for the company, but we had 11 no access to other people's prices. 12 Q. No, but you have just said that you had an instinct or 13 a desire to pitch it at the cheapest workable amount; is 14 that right? 15 A. We would always bid at a level we thought we could win 16 the contract. 17 Q. Looking down the page to (c), if we could just scroll 18 down, it says: 19 "It was acknowledged that Riposte, produced by 20 Escher, was vital to our proposed solution yet there had 21 as yet been no effective technology transfer path agreed 22 to Pathway. Mr Bennett would prepare a paper on 23 proposals to deal with this issue, and try to ascertain 24 any technical or commercial/legal concerns that Escher 25 had." 110 1 Pausing there, Riposte, that was the software that 2 was required to make the whole project work, isn't that 3 right? 4 A. It was one of the essential products, it wasn't the sole 5 product, but it was a critical product to be installed 6 in every post office, so, yes, it was a critical 7 component of our solution. 8 Q. At this stage, did you have a good grasp of the Riposte 9 product and how it worked and how it would function with 10 the project? 11 A. Can you remind me what the date of this document is? 12 Q. Yes, it's 3 October 1995. 13 A. 1995? Well, that's very early in the process. I think 14 we had, at that time, deduced that Escher, with their 15 product, was ideally suited for this solution. It had, 16 as you probably know, already been implemented in the 17 Irish Post Office. An Post had also used this software 18 from Escher and, having seen the An Post use of this 19 software, we were satisfied that this would be a good 20 fit for the requirement of meeting the needs with this 21 contract. 22 Q. But at this stage, you still had quite a lot of work to 23 do to understand how it would fit in this specific 24 project; would you agree with that? 25 A. We had a lot of work to do to understand how it would 111 1 fit and, also, which I think this leads on to, how to 2 manage our relationship with this small company. Both 3 of those factors were well understood in 1995. 4 Q. You also wrote a report that's appended to the end of 5 this document and if we could turn to page 8, please, 6 and if we look at the first paragraph, it says: 7 "The last four weeks have seen very slow progress on 8 the Stage 3 plan. BA/POCL have found it increasingly 9 difficult to meet their timescales for schedule 10 production and release. Progress has slipped 2/3 weeks 11 in 4 weeks with little confidence in future dates and 12 their achievement. 13 "Pathway has not been without its own concerns. 14 Product descriptions have progressed only after intense 15 and tiring effort; a number of key components for WINDEM 16 are still outstanding; in short there is a sense of the 17 procurement becoming bogged down." 18 What did you mean by that last sentence, that there 19 is a sense of the procurement becoming bogged down? 20 A. I think there was quite a big issue here which reflects 21 later through the documents -- and perhaps you will 22 review -- which was that we were dealing with what was 23 essentially a PFI contract and a PFI contract being let 24 for a complex IT bespoke system was something I hadn't 25 experienced before. And there was a conflict beginning, 112 1 I think, to build up between how you could achieve a PFI 2 contract with all the risk transfer requirements, 3 alongside how and who had responsibility for design. 4 And, I think, somewhere in here that is an element of 5 where the procurement became difficult. 6 It took a long time for the procurement to complete. 7 It was -- I think I said over two years, which is 8 an awful long time for a IT procurement and I think 9 a lot of it reflected the difficulty of letting a PFI 10 contract. This was not a standard IT supply or even 11 service contract, and the whole basis of it required 12 a lot of hard work and difficulty in the procurement 13 phase, and I think it's -- I just say here, and even 14 when that procurement eventually became a contract, the 15 contract documentation was still in many ways incomplete 16 and had to be dealt with by all the parties 17 post-contract. 18 Q. Yes, and if I can ask you to turn to a board minute from 19 1996. If we pull it up, it's FUJ00077839. This is 20 a board meeting on 21 February 1996 and if we turn to 21 page 2, please, and if we can pull up this second 22 paragraph, thank you. In that second paragraph, picking 23 up on this point that you mentioned, you say: 24 "If this work is left until after the Service 25 Provider's selection then it is difficult to see how 113 1 a binding contract could be entered into and accepted by 2 either party." 3 So that was an issue you were very aware of in 1996? 4 A. I'm just trying to look in this paragraph to where 5 that -- what you just read out. Is it in the -- can you 6 just remind me where it is? 7 Q. Yes, let me find it for you. My apologies, it's 8 actually the paragraph before. 9 That's why you couldn't find it, and it's the last 10 bit of the second paragraph, if you see it says: 11 "If this work ..." 12 A. I see, yes. Yes, I understand that now, yes. And the 13 question was? 14 Q. So at this stage, you are raising this as an issue, the 15 difficulty of securing a valid and binding contract 16 between all the parties. 17 A. Yes, and it's accentuated by the fact that, being a PFI 18 contract as the provider, then we would only begin to 19 earn revenue from the contract during the steady state 20 or operating phase of such a contract, and the longer 21 the procurement -- or the longer it took post-contract 22 to sort out these outstanding issues would quite 23 naturally delay the time by which we could enter steady 24 state and, by implication, how we could get our returns 25 from this substantial investment. 114 1 So one thing is delaying the contract but equally 2 important is the issues left unresolved in the contract, 3 which still have to be sorted out and you -- later on, 4 but perhaps it is worth mentioning now, we entered into 5 a situation where there was a -- things called 6 agreements to agree, you will find this percolating 7 through a lot of this documentation. 8 An agreement to agree was a sort of legal camouflage 9 for hiding difficult issues. It said that the parties 10 would use reasonable endeavours to sort out things they 11 couldn't sort out during the contract negotiation but, 12 by their very nature, it led to delay and difficulty for 13 everyone and I would say that we didn't get to the 14 bottom of agreements to agree for quite a few years. 15 Q. If we could scroll further down this document so that it 16 is showing what it was showing before and at the end of 17 the second paragraph that's shown there, again picking 18 up on something you have already mentioned, the last 19 sentence says: 20 "There is a growing awareness that the structural 21 weakness of this procurement is having two customers who 22 see the world from quite different perspectives." 23 A. Yes. 24 Q. At this stage, though, you still felt like the project 25 was doable and that it was something that ICL should 115 1 continue to bid for; is that right? 2 A. I think I'm going to say yes, but presumably the date of 3 this -- is this -- this is before the contract was let, 4 isn't it? 5 Q. It's 21 February 1996? 6 A. So it's a few months before the contract was let. 7 Q. Yes. 8 A. Yes, I think we did. I think we were alive to the fact 9 that this was an unusual contract in having two 10 customers in one contract. That did give rise to, 11 downstream, quite a few difficulties which we might 12 refer to later but at this stage in the process we 13 accepted this as being what the contract was and what we 14 were bidding for. 15 Q. The next paragraph goes on to mention the risk register. 16 A. Yes. 17 Q. Could you explain a bit about that? 18 A. Well, we started constructing a risk register both in 19 a pre-sales environment, in other words what are the 20 risks of bidding or winning this contract, and in some 21 ways were the risks worth taking, should we in fact 22 continue with the bid, that's what the risk register was 23 about in the pre-sales environment and it is true 24 because I can read here that we refer to Escher and 25 I guess we refer to Escher more later. 116 1 Q. Yes. 2 A. But I think there was a recognition, both in the 3 pre-sales risk register, that dealing with a very small 4 company would be a substantial risk to mitigate. We 5 recognised that in bidding and that risk remained with 6 us, I think, throughout the whole time I was on Pathway. 7 Q. I was going to ask you about Escher because we have 8 moved forward in time since the last board minutes. Was 9 there any work that you recall that was done with Escher 10 between these two board meetings to minimise or work 11 together with Escher to minimise that risk? 12 A. I can't really pick out the chronology of that, but what 13 I can say is that in mitigating this risk we took, 14 either at this stage or perhaps later, I'm not sure, 15 a number of actions to mitigate this risk, for example 16 one I remember doing, we took their entire source code 17 of their software and put it into escrow account, such 18 that if anything happened to the company we would have 19 access, at least at source code level, to everything 20 they had provided us with. 21 We also -- and I suspect it was probably a bit later 22 than this -- did draw up a comprehensive teaming 23 agreement which spelled out what their responsibilities 24 and ours were and how we should handle it. That was 25 a help. 117 1 We also, and I guess it's probably later than this 2 as well, did allocate key staff to go to Boston, which 3 was where Escher was based, to work directly within 4 their team to learn first-hand elements of their 5 software. 6 Q. As you have mentioned that can we pull up the next board 7 meeting minute please, which is FUJ00077842, and if we 8 go to page 2 and if we can go to (d) please. This is 9 what you were just mentioning, Mr Bennett, about going 10 and working with Escher in Boston. This is in 11 April 1996, so very close indeed to being awarded the 12 contract. 13 A. Yes. 14 Q. I'm going to suggest to you that this is quite late to 15 be doing this type of work with a company where it was 16 so important to the delivery of the project. 17 A. We had -- our knowledge of Escher wasn't just going to 18 Boston. We had knowledge of Escher through the work and 19 our association with the Irish Post Office, An Post, and 20 one of the key directors of An Post was actually at 21 an early stage, I think, on or attended our board 22 meetings, which were used to supervise the project. 23 So yes, this visit to Escher was, as you say, in 24 April, but we had access to and knowledge of Escher, 25 particularly through An Post, probably -- this is 118 1 April 1996, probably -- maybe up to -- certainly well 2 into 1995, so we had access and knowledge of Escher 3 starting in 1995, perhaps even earlier, I can't recall. 4 Q. But is that knowledge principally from An Post, as you 5 say? 6 A. I'm not sure, but from reading this it sounds very 7 likely that's the case. 8 Q. At this stage it is anticipated that the rollout will 9 take place in July 1997; is that right? 10 A. Yes. Yes. 11 Q. What were the reasons for the delay? 12 A. I think in my statement I highlight three areas, all of 13 which contributed to the delay in rollout. The first 14 one, which I referred to, is the difficulties we were 15 having within Pathway in designing and completing our 16 software development. Those problems caused delay and 17 we had to take action to deal with them, so that was one 18 cause of delay. 19 Another cause of delay, which I have already hinted 20 at and referred to, was the fact that this was a PFI 21 contract with a very unclear baseline when the contract 22 was set and those requirements -- there was a phase 23 which I seem to think was called "drop down", which is 24 a bit of an ugly phrase, but it reflected the fact that 25 everyone knew that the contract, although it was let and 119 1 was in about four lever arch files thick, was incomplete 2 and drop down was a process defined whereby those 3 missing pieces could be sorted out. 4 In actual fact, they weren't sorted out in the 5 three months allowed for drop down and they -- many of 6 them weren't sorted out even after six months, which was 7 an extension and, as I said earlier, it still left 8 a huge number, or a large number of agreements to agree 9 which continued to be addressed well after the drop 10 down. 11 So the PFI contract -- I think I might say it now, 12 if I don't say it later, I think was totally ill suited 13 to this contract but none of us at the time spotted 14 that. Anyway, PFI itself caused delay as we struggled 15 to understand the design processes and the clear 16 statement of the requirements. We were developing 17 a system and writing software against a moving baseline. 18 And the third issue -- I have developed two -- the 19 third issue was, I think, a recognition that everything 20 Pathway did was equally dependent on large development 21 work taking place both within the two clients. The 22 Benefits Agency had a huge amount of work to do in 23 preparing themselves for their new Benefit Payments 24 System, as indeed did the -- and the Post Office 25 themselves, Post Office Counters had massive development 120 1 work as well, and there were dependencies and 2 interdependencies between all three parties. 3 Each of us had to interface with each other, take 4 information from and deliver information to, in order to 5 eventually deliver an end-to-end service. So the delays 6 were of those three components. 7 Q. I'm now going to ask you some questions about the 8 Initial Go Live. Could you tell us about the Initial Go 9 Live that happened shortly after the contract was 10 awarded? 11 A. Perhaps -- only in, perhaps, outline really. A small 12 number -- and it was a small number -- of post offices 13 were selected to be the first post office to use the new 14 system for benefit payments. As I remember, the first 15 benefit which was available through the new Benefit 16 Payment Card was child benefit. 17 The other benefits were not -- were yet to be 18 developed by BA but, for Initial Go Live, the Child 19 Benefit payment was available and was able to be used in 20 those selected post -- and I think, at some point, 21 something like 40,000 beneficiaries on Child Benefit 22 were able to collect their Child Benefit payments using 23 the Benefit Payment Card through that selected number of 24 post offices. 25 I might not have it absolutely right, but that's the 121 1 sort of -- my understanding of what happened in 1990 -- 2 I don't know what -- '98, was it? 3 Q. 1996. If we could turn up the report that was written 4 by one of your colleagues, FUJ00058278. You can see 5 there in the abstract that, as you have described, it 6 was introducing the benefit payment system into 7 "10 offices in Stroud early". 8 If we could turn to page 8 of that document. Just 9 looking at the first paragraph there: 10 "IGL was implemented as a result of a political 11 imperative from Peter Lilley which called for the 12 introduction of payments of benefit by card, with the 13 deadline of the 23rd September." 14 Do you remember a political imperative? 15 A. I remember Peter Lilley coming to Stroud to look at the 16 system. I remember meeting him there and, obviously -- 17 I didn't -- whether there is a political imperative, I'm 18 not sure, but I do remember him being there and being 19 very keen on the system. 20 Q. Did you feel like you were under pressure to roll out 21 this Initial Go Live quickly? 22 A. No. 23 Q. What did you think of the Initial Go Live? Was it 24 a useful process? 25 A. I think two responses. Yes, it was a useful process but 122 1 it was extremely short and extremely small. It was not 2 really real time representative system. It was too 3 limited in scope. 4 Q. Was it quite high level? 5 A. Well, it paid benefits, which is what it was there to 6 do. So for what it did, it was a proper operational 7 system. It's just that there were very few offices and 8 only one benefit payment. That's the limitation. But 9 within that limitation, it was a true operational 10 system. 11 Q. You presented the results of the Initial Go Live at 12 a board meeting on 25 November 1996. Can we get up the 13 document FUJ00077844. If we go on to -- if you scroll 14 to the bottom of that page we can see that you presented 15 a report and if we turn over the page we can see that it 16 says: 17 "Initial Go Live 2 had been successful on 18 21 October, although it took more effort than planned. 19 Mr Bennett confirmed he was likely to sign off the drop 20 down process at the end of the present week." 21 Can you explain how it took more effort than planned 22 or why you would have said that? 23 A. No, I can't remember. 24 Q. At this stage, did you feel like things were operating 25 well and that you were making good progress with the 123 1 Horizon project? 2 A. I think this is -- the date of this is -- can you remind 3 me of the date? 4 Q. Yes, the date of this is 25 November 1996. 5 A. Okay. This is very early in the life of the contract 6 and it was too difficult to judge at that time how 7 things would unfold, so it was an early start but only 8 an early start, and there was an enormous amount still 9 to do. 10 Q. I'm now going to pick up again the topic of delays. If 11 we could turn up FUJ00078972. This is a memo that you 12 sent on 7 January to all Pathway staff, and we can see 13 there in the first paragraph: 14 "At the PDA board meeting [that's the Programme 15 Delivery Authority] in December, it was proposed that 16 a joint review should be undertaken between ICL Pathway 17 and the PDA to assess all the remaining risks which must 18 be handled prior to a successful implementation starting 19 with the Live Trial and leading in to National Rollout." 20 If we can scroll down a bit, please. Yes. Up 21 again -- no, there it is: 22 "It is very likely that you will hear views 23 expressed that the programme will be delayed by anything 24 between one to as much as six months. At this stage 25 I must ask you to treat any such views as speculative. 124 1 If additional time is considered appropriate, it will be 2 handled through our formal change control process." 3 So, at this stage, you seem to be suggesting that 4 rollout can happen reasonably quickly? 5 A. I think there was -- and this letter rather implies it. 6 There was a recognition that, although that was the 7 date, delays were not ruled out. I think at this stage 8 we could begin to anticipate that there would be delays 9 yet to come and be clarified. 10 Q. It says "as much as six months". That appears to be the 11 outer limit of what you thought the delay would be at 12 this time? 13 A. Well, I think it says here that other people were 14 expressing the view about one to six months. I don't 15 think I was expressing the view of one to six months. 16 Q. Quite right. At this stage do you recall how long you 17 might have thought that it was going to take? 18 A. No, I didn't have a considered view. 19 Q. In February 1997 there was a no fault replan of the 20 programme between the three contracting parties and then 21 in June there was a meeting of the Programme Delivery 22 Authority. If we could pull up POL00028304. Could you 23 explain to us what the Programme Delivery Authority was 24 and how it worked? 25 A. Well, I can try. I mean, it goes back to something 125 1 I was saying earlier, which is this contract was unusual 2 in having two customers and one supplier. Most of my 3 experience of IT contracts with government were you had 4 one customer and one supplier and you dealt directly 5 between the supplier and the customer in all aspects of 6 the contract and contract delivery. 7 Now, in the case of this PFI contract, that didn't 8 function, so instead -- and this is something which the 9 contract documentation, early on, didn't really specify 10 in any details -- there was a need to create a new 11 organisation called the PDA, which sat above and 12 in-between the supplier and the two customers. 13 So the PDA was a new organisation, staffed from both 14 BA and POCL, and it was put in place to really discharge 15 the contract on behalf of both customers, but in many 16 ways it did actually sit between us and the customer and 17 it got quite large and I think you will -- maybe you 18 will come to this later on -- there were -- but I will 19 mention it now -- there was a major review -- this is 20 1997, it wasn't soon after this, it was a major review, 21 external audit, carried out by the PA Consulting group 22 which I -- which looked at the entire range of project 23 Horizon. 24 And they addressed this issue of the PDA and its 25 size and, I think, echoed some of the comments I have 126 1 just made, that it was a large organisation, it was 2 interceding between us as a supplier and the two 3 customers and it didn't have -- and this I think is 4 quite important -- any executive authority, by which 5 I mean it could press ICL Pathway to do things and it 6 could disagree with what we're doing and it could point 7 out the things we hadn't done. 8 But when it came to making decisions about change or 9 things which would affect the so-called sponsors, this 10 case BA -- it always had to revert back to the two 11 customers, so that process of interceding did actually 12 take a lot of time. 13 So if there were critical issues which needed rapid 14 resolution, we might be able to agree to them within 15 ICL Pathway, but PDA could not agree to them itself, it 16 had to revert back to the sponsors. So you will often 17 find, in a lot of the documentations, the "This will be 18 reverted to the sponsors for agreement" and indeed it 19 was but it was a process which took time. 20 Q. If we turn you to page 3 of this document, we can see at 21 these PDA meetings you had the opportunity to present 22 reports, or report ICL Pathway issues; is that right? 23 A. Sorry, which report is this? This is? 24 Q. So you can see there it says "John Bennett reported on 25 ICL Pathway issues". 127 1 A. This is a PDA board meeting, is it? 2 Q. Yes. 3 A. Okay, I'm with you. 4 Q. Yes, the same document. 5 A. Okay, thank you. 6 Q. Would that typically happen? 7 A. At the PDA board, yes, I think on -- on the PDA board we 8 would always report on progress and issues. 9 Q. If we look at the fourth -- 2.4.4 it states: 10 "John Bennett, Paul Rich and Colin Baker NFSP had 11 visited sites in the North East. It was noted that that 12 had been a useful exercise in terms of achieving early 13 exposure of some of the issues needing NFSP decisions." 14 Can you tell us a bit about what that site visit 15 involved? 16 A. I don't think that these site visits were ever designed 17 to resolve any issues. I don't recall -- I do recall 18 the meetings. I mean one of the things I was very 19 interested in -- and this is one example of it -- I was 20 very interested in visiting Post Offices around the 21 country. I was keen to meet subpostmasters and their 22 staff and I was very keen to see how they were finding, 23 adopting and using the new system or their enthusiasm 24 for anticipating the new system, so I liked -- I enjoyed 25 and found it very useful to meet and have these visits 128 1 and I did request a number of them which were always -- 2 and Paul Rich, I recall, often would sponsor and 3 organise these meetings and, as you can see, would 4 attend. So we had a number of these meetings over time, 5 which I found very useful, but I don't believe they were 6 ever designed to deal with any particular specific 7 issue. There might have been feedback and an 8 understanding of how subpostmasters were thinking, but 9 I don't recall they ever led to specific resolution of 10 defined issues. 11 Q. As managing director of Pathway, you wrote a number of 12 monthly progress reports on the development of the 13 Horizon System; is that right? 14 A. Correct. 15 Q. What can you tell us about how those reports were 16 written and drawn together? 17 A. This was something we started quite early in the 18 process, probably after contract I guess, but quite 19 early in the process and I determined that it would be 20 very useful to have a comprehensive monthly report of 21 everything we were doing in Pathway and to construct -- 22 and I -- and to construct that report I got every one of 23 my direct reports, obviously, to send me their input. 24 I would consolidate their input and write a management 25 summary. 129 1 And that report -- as I think you've got quite a few 2 of them in this bundle -- would often run to 20 or 3 30 pages of information and that report was sent both to 4 my boss, which was the chief executive of ICL, it was 5 shared with all my direct reports and it was also shared 6 with other people who needed it, for example we had 7 a dedicated team from Fujitsu from Tokyo with us, for 8 quite a time, who came to work with us later on in the 9 process and they had a copy of this report. 10 Since it was sent to all my team, it meant that all 11 my team, not only knew what their contribution this last 12 month had been, but they could see quite clearly what 13 the contribution of all their colleagues had been. So 14 I looked upon it as a very important piece of 15 communication and I continued it every month for, 16 I think, quite a few years. 17 Q. Shall we look at one of those reports. If we could pull 18 up FUJ00058162. Thank you. If we could turn to page 4. 19 This is what you were just describing, isn't it, one of 20 these reports? 21 A. Yes, it is indeed. 22 Q. When you were talking about a managing director's 23 summary, this is the bit that you wrote? 24 A. Correct. 25 Q. This report is dated 11 July 1997 and we can see here on 130 1 this page, looking at the first paragraph: 2 "This month saw the drawing together of the findings 3 of three separate activities. The first concerned the 4 difficulties in testing of Release 1c, the second review 5 concerned the plans for future Pathway releases in 1998 6 and the third was the output and findings from the 7 programme audit conducted by Mike Coombs and Andrew 8 Boswell across ICL Pathway. This has led a programme 9 review being undertaken and a comprehensive reassessment 10 of the achievable outputs from now through to national 11 rollout." 12 It goes on to say: 13 "The programme review has been presented to PDA, 14 POCL, BA, ITSA and SSA ... The consistent response has 15 been one of disappointment and shock that yet another 16 slippage has come to the surface so soon after two 17 previous replans." 18 So, at this stage in your report, you're reflecting 19 the frustration of the various interested parties; is 20 that right? 21 A. Yes. Everyone was deeply disappointed and affected by 22 these delays. 23 Q. If we could scroll down to the bottom of that page, the 24 final bullet point, it says: 25 "The anticipated delays have caused a serious impact 131 1 on the business cases of BA, POCL as well as 2 ICL Pathway. As a result, quite a few people have now 3 voiced their concerns that the programme may not be 4 viable and this has given rise to suggestions of 5 significant contract readjustment as well as exploring 6 the position on termination." 7 Was this a very difficult time to be in your job? 8 A. I think this was a very difficult time for everyone, not 9 just me, but the sponsors and all our staff. 10 Q. If we could turn to page 19 of that document please. If 11 we could scroll down, please, to the "Current Critical 12 Problem". Is it a fair summary that often, if not 13 always, in your report you would summarise each theme 14 and set out a current critical problem? 15 A. I think in the report we -- I had settled on a format 16 which was always followed every month. It was followed 17 both by my summary and I think the format was reflected 18 in all the reports from my management team, so we all 19 had a strict format and this is -- I think that's 20 right -- and this is a part of that format. 21 Q. If we look at the bottom bullet point, we can see that 22 there is a reengineering programme being devised to: 23 "resolve the issues surrounding the EPOS product." 24 Do you see that there? 25 A. Yes, I do. 132 1 Q. That involves redevelopment work being carried out by 2 Escher. 3 A. Mm-hm. 4 Q. So, already at this stage in 1997, there are issues with 5 the EPOS product, aren't there? 6 A. Yes. 7 Q. Was that something that you were acutely aware of and 8 concerned about? 9 A. I think the problems with EPOSS started quite early and 10 they continued for a long time. There was a lot of 11 action to resolve and deal with the issues, but they 12 persisted through -- what -- I don't -- what month -- 13 this is July -- this is June 1997, isn't it? 14 Q. Yes, written in July but dealing with June 1997. 15 A. I mean, EPOSS took the attention of my technical team, 16 not just in 1997 but all the way through to 1999 and 17 probably into year 2000. EPOSS was always high on the 18 agenda of things to deal with. 19 Q. Is that partly because it was such an important part of 20 the programme itself? 21 A. It was a vital part of the programme for Post Office 22 Counters and it was -- and I think I might have 23 mentioned this -- the service, the EPOS Service was 24 an end-to-end service, so it reflected not only what we 25 were doing but it reflected what was happening across 133 1 the interfaces with the work done elsewhere, 2 particularly in this case, or later on with Post Office 3 Counters. 4 So it was a complex product. It had extensive 5 end-to-end implications. It involved an awful lot of 6 people and we, in our responsibility, had our own 7 problems in delivering our side of that picture as well. 8 Q. If we could turn up a PDA minute. If we could pull up 9 POL00028311. This is another PDA board minute. If we 10 could scroll down, this is, as I say, from August 1997. 11 The first item is: 12 "Mr Bennett to identify how long it will take to 13 obtain all the information required for all post offices 14 and provide a progress update with timescales for 15 collation of full details and resolution of the issues 16 to the next PDA board meeting." 17 So, at the same time that the EPOSS issue is being 18 reengineered, you're also being put under quite a lot of 19 pressure to identify how long it is going to take to get 20 the Horizon project completed, aren't you? 21 A. Yes, I mean, this is a different but a major strand of 22 the programme, which is the more practical -- well, 23 I say practical, it's the more physical end of the 24 programme, which is the fact that the software and the 25 equipment associated with it needed to be installed in 134 1 either 17,000 or 19,000 post offices. The number 2 I think came down during the period, but even so it was 3 getting on to 20,000 outlets. 4 I have to say on this that the information available 5 to Pathway at the contract stage about the state, 6 physical state, of the Post Office network was very 7 slender, very thin, one might say very thread bare, and 8 it was only -- and we weren't, I think -- and I might 9 have this slightly wrong, but we were not -- not only 10 us, but the other shortlisted suppliers as well -- we 11 were not encouraged or allowed during the sales 12 campaign, during the procurement phase, to see for 13 ourselves what a typical or a range of post offices 14 might actually look like, in reality. That was 15 something which was not to be available until 16 post-contract. 17 Q. When you say it wasn't allowed, was it something that 18 you asked to be able to do? 19 A. Do you know, I can't remember. 20 Q. Are you conscious of anyone asking the Post Office 21 whether they could go and inspect -- 22 A. Oh, I think so, although I dare say this wasn't the 23 highest thing on our list. We rather took it that the 24 post offices would be suitable for automation and the 25 Post Office Counters themselves would have told us in 135 1 advance if they themselves had serious doubts that their 2 estate was not or had defects or difficulties. 3 So we relied, I guess, on any information from the 4 Post Office and took on -- perhaps on trust or however 5 you might describe it -- the view that the network would 6 be in good condition. 7 Q. You made assumptions about the network? 8 A. I guess we did. 9 Q. Shortly after this -- if we could pull up POL00028442 -- 10 Peter Crahan on the 24th -- if we could look at page 3 11 please -- wrote to you to give notice of breach of 12 contract. If we could maybe scroll down so that we can 13 see the whole letter, thank you. 14 Were you surprised when you received this? 15 A. Sorry, can you just remind me of the date of this 16 letter? 17 Q. Yes, it's 24 November 1997. 18 A. 1997. Was I surprised at this letter? Probably not. 19 Q. Did you see the way that things were going? 20 A. Well, I think that everyone was having difficulty and 21 major IT contracts nearly always have difficulties in 22 their early days. Normally, and in most cases, those 23 problems are dealt with but, if they can't be dealt 24 with, then contract cancellation is always an option. 25 So I think, in this case, I certainly couldn't put to 136 1 one side that the government would take the option to 2 cancel. It was always an option. 3 Q. Mr Todd's evidence was that you were involved in 4 producing a response to this letter and, if we could 5 pull it up, it's POL00031117. If we could turn to the 6 third page, please. Were you involved in drawing up 7 this position paper? 8 A. I wasn't the author, but yes, I was aware and I inputted 9 to this position paper. I suspect, reading it, it was 10 produced by our legal counsel, the way it's written, but 11 yes, I was involved and I was aware of the content and 12 the thrust of this position paper. 13 Q. What did you understand "without prejudice" to mean? 14 Was that the legal person's drafting? 15 A. One of the -- let me just stand back from that a bit. 16 One of the constant themes in this contract -- it didn't 17 just start with this position paper, it started very 18 early on, which was somewhat disappointing that so much 19 correspondence between the sponsors and Pathway was 20 always headed "Without Prejudice". This is, for me, 21 quite unusual because, in most contracts, or most IT 22 contracts I have been involved with, the contract, once 23 it had been signed, was normally put to one side and 24 everyone worked on delivery. 25 I don't recall in my previous roles ever constantly 137 1 having correspondence which is always without prejudice. 2 It was an unusual feature of this contract and this is 3 perhaps another example of it. 4 Q. So the fact that it is headed "Without Prejudice" didn't 5 impact or affect the information that you fed into this 6 document? 7 A. No, it was a constant feature of the contract. 8 Q. If we could turn to page 5, please, and if we could 9 scroll down. This is a long document but I'm only going 10 to take you to some key parts. Further down and, in 11 fact, it's across the page, but this is a section 12 dealing with "The Authorities/PDA". 13 If we could go over -- yes, great, and scroll to the 14 bottom of that page, please. 15 At the bottom of that page, it says: 16 "At the pre-contract stage, had the true role of the 17 PDA been accurately described, Pathway would have 18 reconsidered the commercial terms upon which it entered 19 into the Contract." 20 Do you agree with that? 21 A. Yes, I do. 22 Q. Have you read through this document and is there 23 anything in its contents that you take issue with? 24 A. I have read this document several times and it was the 25 result of a considerable amount of thought and work in 138 1 constructing it, and I think, I would say, that this 2 document, within the bundle of documents you have sent 3 me, is one of the more significant documents I would 4 subscribe to. 5 Q. If we could look at page 10 of that document and if we 6 could scroll down. In that first line, it says: 7 "It became apparent during installation work for the 8 first 200 Post Offices that many post offices are not 9 fit for the purpose of installing automation equipment." 10 Was that something you also fed into? 11 A. Well, I think I was talking about that a few minutes 12 ago, so, yes. I mean, what -- just what we did, as soon 13 as we were beginning to draw up our plans for rollout, 14 we made a point of visiting every single one of the 15 19,000 post offices to do our own survey, on the grounds 16 that there were no other survey reports to work from. 17 So within this contract -- I don't know whether we 18 anticipated it at pre-sales time, I suspect we didn't, 19 but, as soon as we started, it was clear that, since we 20 had to install this equipment, we had to do our own 21 surveys. And we did survey every single office to 22 determine did it have the physical attributes to take 23 even the limited computer equipment we provided, did it 24 have sufficient power supplies to drive those devices, 25 did it have the capability to be connected to a network 139 1 and how much work, by whom, needed to be done to bring 2 it up to a satisfactory state. 3 That was a job which fell on ICL Pathway to do and 4 we did it for every office. 5 MS KENNEDY: Sir, that might be a convenient moment to take 6 a short break. 7 SIR WYN WILLIAMS: Thank you and what time should we start 8 again? 9 MS KENNEDY: 3.10? 10 SIR WYN WILLIAMS: Certainly, okay. 11 MS KENNEDY: Thank you. 12 (3.00 pm) 13 (Short Break) 14 (3.10 pm) 15 MS KENNEDY: Sir, can you hear me? 16 SIR WYN WILLIAMS: Yes, I can, thank you. 17 MS KENNEDY: Thank you. 18 Picking up where we left off then, if I could bring 19 up a further report that you wrote on 10 July 1998. 20 It's FUJ00058174. Thank you. If we could turn to 21 page 5, please, and if we look at the fourth bullet 22 point down it states: 23 "The main stress point on NR2 testing continues to 24 be with the EPOSS system, testing together with POCL 25 reference data. Resources have been reallocated within 140 1 Pathway to bring more effort onto this area and extra 2 skills are being brought in from elsewhere in ICL." 3 If we turn over to page 11, and we're looking 4 halfway down the page, it states -- so towards the 5 bottom of the screen: 6 "The volume of system incidents (PinICLs) is 7 perilously high. Clearance plans are being devised to 8 are all products." 9 If we look to the bottom of that page, it says: 10 "The current architecture for the TIP interface 11 module may not be capable of supporting the entire 12 network. The simulated tests suggest that it cannot 13 harvest the daily transactions generated by EPOSS and 14 send them to POCL in the overnight time slot available." 15 So, as at July 1998, the EPOSS issue is reaching 16 a point where you are needing to deploy extra resources 17 into it; is that right? 18 A. Yes, yes, we did need to put more effort in. 19 Q. Were you being put under pressure to clear PinICLs at 20 that stage? 21 A. The pressure was self generated. There was a clear 22 recognition that there were many more errors than we 23 expected and that these were growing rather than 24 retreating, if I put it that way, and so the effort was 25 doubled down on this system and I think there were -- 141 1 you may talk about it later. There were a number of 2 technical audits carried out on the development process 3 to try and find a way of bringing our development work 4 of EPOSS more under control. 5 Q. Shortly after this report, you set up a task force, 6 an EPOSS task force; do you remember that? 7 A. I remember it but let me put it this way, personally 8 I don't remember that task force, it was 24 years ago, 9 but if I look at the bundle of documents you sent me 10 I have had a read of that task force report so I'm 11 familiar with it in terms of it within your bundle, 12 although I don't particularly remember it directly 13 myself. 14 Q. The task force at the time would have been reporting to 15 you as the MD; is that right? 16 A. That's not quite right. Let me step back a bit. In 17 Pathway, we conducted every year a range of audits, 18 in-house audits, technical audits and in -- I'm not sure 19 what year this is, is this 1998? 20 Q. 1998. 21 A. Well, in 1998 -- particularly in 1999, which I remember 22 better, we embarked upon probably several technical 23 audits and one technical audit was indeed to do with 24 this task force, looking at the development of EPOSS. 25 This technical audit was produced for my development 142 1 director. It was he and his staff who were writing the 2 software and this audit report was produced for him and 3 his team. 4 And if you look at my report, which you have done, 5 you will see it was sent to a lot of people and I indeed 6 was sent a courtesy copy so, in a sense, I was sent 7 a copy of this task force report. But, being an audit 8 report, it was sent to my development director and his 9 boss, which is the overall programme director. 10 Now, the principle in Pathway was that if we had 11 an in-house audit then the process was that when the 12 audit report was completed, all those recommendations 13 would be reviewed by the responsible directors and the 14 corrective actions would be registered, recorded and 15 were placed with owners. And I think you will find 16 that, as far as that task force report was concerned, 17 that's exactly what happened and the corrective actions 18 from that task force report were recorded, owners were 19 placed and there is a record of that in this bundle. 20 Q. We will come back to the eventual report from the EPOSS 21 task force in a moment, but if we could move forward to 22 August 1998 and we could pull up POL00031127. This is 23 a Bird & Bird expert report produced on the BA/POCL 24 payment card programme and if we look at page 4, we can 25 see the purpose of the report and the terms of 143 1 reference. So this is at the stage of assessing the 2 strength and weaknesses of the sponsoring authority's 3 case. 4 If we could turn to page 12, please, and scroll 5 down, we can see at paragraph 304 that you are mentioned 6 in the middle of that paragraph. The paragraph reads: 7 "We have not yet conducted a review of the project 8 management of the programme by ICL Pathway, the PDA and 9 the Sponsoring Authorities. Our initial impression is 10 that no formal methods, such as PRINCE, were used. 11 There are well prepared, high level, plans documented in 12 the various versions of the Master Plan and we 13 understand that lower level plans may have been prepared 14 in some areas. The comment by Mr Bennett of ICL Pathway 15 at the PDA Board meeting on 13th August 1996, when 16 Master Plan version 1 was approved, is revealing. The 17 minutes state ... 'Mr Bennett confirmed that ICL Pathway 18 were anxious to progress from drawing up plans to 19 actually using them and were content therefore to sign 20 off this version for that purpose. He did however want 21 it placed on record that the plan quoted dates which 22 were a mixture of agreed baseline, planned targets and 23 hoped for dates and that this weakened the effectiveness 24 of the plan and the process for review'." 25 If we turn over the page and scroll down it says: 144 1 "Our impression from the work we have carried out to 2 date is that Mr Bennett's comments of 13th August 1996 3 reflect the weakness in the project management approach 4 taken by ICL Pathway at that time." 5 Do you have any comments on those words in the 6 report? 7 A. I've never seen this report before so I don't know its 8 pedigree at all. I do notice that it is August 1996, 9 which is early in the process and I would have -- I was 10 of the view that, certainly around that time, the 11 programme management processes ICL Pathway was adopting 12 were quite extensive, but I think that's probably 13 a better question for my programme director or my 14 development director to answer, so I can't say much more 15 about this report. 16 Q. If we then move forward in time to May 1999, you 17 produced an update for the board at that stage, which is 18 at FUJ00117463. Could we turn that up, please. If we 19 look at the top, you have stated: 20 "The ICL proposal submitted in December 1998 has not 21 been progressed or discussed since that date and we have 22 now formally withdrawn it." 23 It then sets out a number of options and, if we 24 scroll down we can see option B1.2 and you set out there 25 the proposal for that suggestion should the BA Benefit 145 1 Payment Card be cancelled. So, at this stage, you are 2 already sketching out a plan for if the BA benefit card 3 is cancelled, aren't you? 4 A. It looks like it, yes. 5 Q. Do you remember this from the time? 6 A. Not specifically, but it's -- it sounds quite familiar. 7 Q. If we turn to page 7 of that report and we look in the 8 "Programme Status" section, in the middle of that 9 section, it says: 10 "All the 200 post offices running release 1c have 11 been successfully converted to NR2 [New Release 2] which 12 is running well in terms of the day-to-day transaction 13 operations. There are however, major problems with the 14 cash account balancing which takes place on a Wednesday 15 evening and this has been analysed as much more to do 16 with the business processes, business support and the 17 skill knowledge of subpostmasters, rather than 18 structural issue with the Pathway system. Nevertheless, 19 it will require joint effort to straighten out." 20 That's quite a crucial issue that you have 21 identified in that report, isn't it? 22 A. Yes, I think it is. I think it begins to draw us to the 23 understanding that the complete delivery of this aspect 24 of EPOSS, which is the critical one to do with the 25 accounts and the balancing of these accounts, required 146 1 quite a lot of activity across a wide spectrum. It 2 clearly involved Pathway and what we were doing, but it 3 required -- and it hasn't been -- well, it's referenced 4 a little bit in the document you pulled up a little 5 while ago. 6 It relied, to start with, by very good input to us 7 from the reference data system being developed within 8 POCL and I think, if you look at the correspondence, in 9 order to address these end-to-end issues, we do advise 10 a thorough review of the end-to-end process. I think 11 people probably understand that, to produce accounts in 12 the Post Office, you need to know how much you are 13 charging for individual products being sold. Now, it's 14 a very similar -- I'm not an expert on reference data 15 but, in its simplistic form, we relied on reference data 16 to tell us the status and pricing of every product to be 17 transacted across the counter. 18 And, if you think about it, taking a very simple 19 example, if the Post Office changes the price of, let's 20 say, a First Class stamp -- it sounds very easy and it's 21 probably the simplest product to think about -- what we 22 have to do is to make sure that every single one of the 23 maybe 40,000 counter positions in perhaps 19,000 post 24 offices, uses the new price at exactly the same time and 25 that's just one simple product and, of course, there are 147 1 hundreds -- I don't know how many, but there were 2 several hundred products which are marketed through the 3 Post Office network. 4 And to get the accounts right at the end of the day 5 or the end of the week, every single product has to have 6 its exactly right price at the right time and I think it 7 was recognised that -- and this isn't from a Pathway 8 audit, but I think it came out from other audits, that 9 just as much as we had work to do in Pathway, there was 10 a lot to do, to do with the integrity of the reference 11 data information which would impact on the ability to 12 produce good accounts at the end of the week. 13 So that was a factor, but I think overall this 14 paragraph does draw you into the complexity of getting 15 end-to-end EPOSS working at the level everyone wanted. 16 Q. You also suggest, don't you though, that the skill 17 knowledge of the subpostmasters was an issue, rather 18 than the structural issue -- there being a structural 19 issue with the Pathway system? 20 A. Well, I'm not trying to dodge the issue that we had 21 problems ourselves. We had a lot of problems and you 22 have referred to those earlier but another end of the 23 EPOSS service is, of course, what happens in the post 24 offices themselves and that, I think, takes you into 25 a broader issue, which is to do with the adequacy of the 148 1 training for the staff who had to use the system. 2 We were contracted to provide substantial training 3 to staff, both the managers and, indeed, the other staff 4 operating the system, but this was -- has often been 5 referred to as a major management of change programme 6 and whether it was sufficient to rely just on technical 7 training on how to use the system to really cope with 8 the issues of helping 30,000 people with a variety of 9 skills, fears and motivation to use the system well is 10 a big challenge. 11 And I think -- my understanding, my knowledge of the 12 EPOSS system end to end is that parts of it were well 13 used and people liked, which was the daily transactions, 14 the ability to construct -- to register a transaction 15 during the day when dealing with people across the 16 counter, in other words the interaction across the 17 screen. I think that system, from my recollection, was 18 generally well used and well liked. 19 The real weaknesses, which is what you have pointed 20 to, were really to do with the end-of-week accounting 21 process, which proved to be, for many people, very 22 difficult. 23 It was perhaps my understanding -- I admit I might 24 have this wrong -- that the question of balancing had 25 been a big issue well before Pathway introduced EPOSS. 149 1 In other words, the paper-based system which pre-dated 2 the Pathway or the POCL system was never an easy system 3 to use at that time and required -- I might have this 4 wrong and I'm perhaps talking outside my knowledge, but 5 I think balancing at the end of the week had never been 6 an easy thing to do. 7 Q. On 14 June 1999 you gave evidence to the House of 8 Commons Select Committee on Trade and Industry. Do you 9 remember that? 10 A. Yes, I do. 11 Q. Your position, is this right, was that you didn't think 12 that there was a technological or technical issue with 13 the system and that this system was extremely robust and 14 applicable? 15 A. If I said that, I must have said it. 16 Q. Shortly after that, the Post Office agreed to accept 17 Horizon on a conditional basis, provided criteria were 18 met. If I could ask you to look at a document from 19 27 September 1999. It is FUJ00079189 and it is page 2, 20 and it says: 21 "I am delighted to confirm the completion of 22 acceptance: the second stage has been successfully 23 finalised by the signing of the second supplemental 24 agreement on 24 September 1999." 25 Did you feel at this time like you had a complete 150 1 endorsement of the system and that no further work was 2 required? 3 A. No, but I would pick issue with you about the acceptance 4 being conditional. There was no question of it being 5 conditional. The system was either accepted or it was 6 not accepted. It is true that, once the system was 7 accepted, there were a number of Acceptance Incidents 8 which had to be addressed later on, but acceptance was 9 acceptance, it was not conditional. 10 Q. If we could look at an audit -- you mentioned audits 11 earlier. This audit was carried out on 28 October 1999 12 and it is at FUJ00079782. If you look at that there, 13 you can see this was one of the audit documents you were 14 referring to earlier; is that right? 15 A. Well, this is an audit document. Whether this is -- 16 yes, it's an audit document, yes. 17 Q. We can see that you are on the distribution list for 18 this one. 19 A. Am I? 20 Q. John Bennett, distribution. 21 A. Oh, yes, correct, yes. 22 Q. If we could turn to page 19 of this document. Mr Austin 23 was taken to this document earlier, so I don't intend to 24 read it out again, but if you could read it to yourself, 25 Mr Bennett, it's the third paragraph -- well, from the 151 1 first paragraph onwards and let me know once you have 2 finished. 3 A. Sorry, from the first paragraph? 4 Q. Yes, from under "Commentary". 5 (Pause) 6 A. Right, I have read what's on this screen. 7 Q. Thank you, and if we could turn over the page, thank 8 you. Pausing there: 9 "The figures indicate that the problems facing EPOSS 10 during the Task Force period have not diminished. Of 11 greater concern are the non-EPOSS PinICLs within the 12 group suggesting that there are still serious quality 13 problems in this vital, customer facing element of the 14 system." 15 Then underneath: 16 "The EPOSS Solutions Report made specific 17 recommendations to consider the redesign and rewrite of 18 EPOSS, in part or in whole, to address the then known 19 shortcomings. In light of the continued evidence of 20 poor product quality these recommendations should be 21 reconsidered." 22 So, at this stage, there's the suggestion that EPOSS 23 should be rewritten; isn't that right? 24 A. Correct. 25 Q. This is 28 October 1999. 152 1 A. Yes. 2 Q. So this is very late in the design of the programme, 3 would you accept? 4 A. Yes, it is. 5 Q. If we could turn to the next document in the series, 6 which is FUJ00079783. This is another development audit 7 and, again, we can see you are on the distribution list 8 and, if we could turn to page 6, again, Mr Austin was 9 taken to this, this morning. Your initials, are they 10 under MTM, "JHB"? 11 A. Yes. 12 Q. So you were the -- I believe this means the managing -- 13 even though Mr Austin was the owner, you were ultimately 14 managing him? 15 A. Well, his direct manager was Mike Coombs, MJBC, and Mike 16 Coombs was the technical -- was the programme director 17 working for me, who had technical authority for the 18 programme. So, yes, the development director Terry 19 Austin worked for the programme director Mike Coombs who 20 was a member of my management team. But he was the 21 overall programme director. 22 Q. Were you aware of this suggestion or recommendation that 23 EPOSS should have been rewritten or could have been 24 rewritten at this time? 25 A. Well, I'm aware of this -- only by reading these 153 1 reports. I have read this report and I think under 2 "Actions" you can see that the view of my -- I don't 3 know, is this the one? Yes, the view of my development 4 director, which it says here at the end: 5 "We will ... continue to monitor the PinICL stack 6 ... and ... re-evaluate this decision. Would [you] 7 please close this issue formally using the rationale 8 described." 9 So that was the assessment of my development 10 director and that was accepted by his boss the programme 11 director. 12 Q. So you wouldn't or didn't have input into that decision? 13 A. No. It never came to me to make a managing director's 14 decision on this point. 15 Q. If we could move forward in time to 20 December 1999. 16 You wrote a further report. If we could pull it up, 17 it's FUJ00058188. If we could turn to page 6, please, 18 and scroll down. It states: 19 "The most serious issue on acceptance resolution 20 concerns AI376 and the integrity of the accounting data 21 being managed from the end to end basis with Horizon. 22 This in turn requires more disciplined and strict 23 accounting integrity controls, some of which can be 24 achieved through the EPOSS reconciliation software and 25 others through process and independent tools and the 154 1 balance through stronger end to end control of the 2 reference data processes." 3 So, at this stage, this issue is causing you still 4 a great deal of concern? 5 A. Yes. 6 Q. What do you understand in relation to the integrity of 7 accounting data being managed from the end to end basis 8 with Horizon? 9 A. Well, I think it's -- well, I'm only interpreting what 10 it says here, that the -- whether there were errors in 11 the accounting data which were being promulgated through 12 the system and could you therefore trust the way the 13 accounting data was being managed. And it does make the 14 point, I think, which I think we talked about a bit 15 earlier, that there's a lot of areas where the integrity 16 can be compromised and there were steps put in place to 17 try and address these. 18 Q. Did you feel that you were across or had adequate 19 knowledge of the steps that were going to be taken to 20 get across this and resolve this issue? 21 A. Well, I think, as we said earlier, we thought that there 22 was work we had to do with our software. We had a lot 23 of discussion about how the training facilities with 24 Post Office based staff should be improved or augmented 25 and I think it called for a new set of training courses 155 1 to help people with -- I think there's new 2 documentation, as well, for guidance and guides how to 3 use the system. 4 There was plenty of opportunities for user errors in 5 running the end-to-end system, which we thought had to 6 be dealt with. So you had a component in the Post 7 Office themselves with how the staff can be helped to 8 use the system, you have problems in the use of 9 software -- and it's not recorded here, but I know there 10 was an issue concerning how long it took to print-out 11 reports in the post office itself, which was a source of 12 concern -- and then there was, as it says here, stronger 13 end-to-end control of reference data. So there were 14 plenty of steps in this process where improvements could 15 be made, a number of them obviously were with 16 ICL Pathway. 17 Q. Did you feel that the programme was fit for purpose when 18 it was rolled out? 19 A. At some point, and I'm not quite sure in this 20 documentation where it is recorded, AI376 was deemed to 21 have been handled to such a level that the software was 22 judged suitable to go into live use. So, at some point, 23 yes, this was considered, notwithstanding these 24 weaknesses, the system was judged fit for purpose for 25 live use. 156 1 Q. When in 2000 did you leave your role in ICL Pathway? 2 A. Well, I was beginning to withdraw towards the end of 3 1999. I mean, if you look at some of the ICL Pathway 4 board meetings you will see that my successor was in 5 fact attending board meetings in 1999 and I think my 6 last board meeting was either January or February 2000, 7 I'm not absolutely sure. 8 Q. If we could turn up WITN04600104 please. This is 9 a further version of a document that we looked at 10 earlier, dated 10 May 2000, so this may well postdate 11 your departure. But if we could look at page 7 -- 12 Next page, I think. Sorry, one moment. Page 9. 13 Yes, over the page. 14 So we have seen all of this before but if we could 15 turn over the page, and you commented on the agreed 16 action before, but this is new. Could you take a moment 17 to familiarise yourself with the "Agreed action" column. 18 (Pause) 19 A. Yes, I have read that. 20 Q. Are you able to help us with what happened between 21 November and May that results in this particular issue 22 being closed? 23 A. No, I can't fill in the gap, but I can notice that you 24 will see that this report was actually sent to my 25 successor, Mike Stares, so he is on the distribution 157 1 list for this and you can further see that Mike Coombs, 2 who was the programme director to whom the development 3 director reported, really supported the previous view of 4 my development director that this code should not be 5 redesigned or rewritten at this time. In fact, 6 Mr Coombs makes it more clear in that final paragraph of 7 actions -- let me just have a look: 8 "Effectively as a management team we have accepted 9 the ongoing cost of maintenance rather than the cost of 10 a rewrite." 11 And I think that was the judgement of my programme 12 director and I -- although I wasn't, as you say, on -- 13 in post in May 2000, I would have supported his 14 judgement. 15 Q. To your knowledge, were Post Office employees able to 16 review PinICLs? 17 A. Do you know, I'm not sure, but I would have thought they 18 would. We were working very closely with Post Office 19 and AI376, which lies behind an awful lot of this, was 20 actually raised by the Post Office Counters' people so 21 they would have been very much aware of everything going 22 on with end to end EPOSS service, including the closure 23 of AI376. It was their decision to close it. 24 Q. You may have heard the evidence of Mr Oppenheim 25 yesterday. Did you consider that remote access by ICL 158 1 and Fujitsu a facility that was so obvious that it 2 didn't need to be set out or explained? 3 A. Sorry, could you repeat the question. 4 Q. So remote access, the ability to access branch data 5 remotely, would you consider that so obvious that it 6 need not be minuted or explained, the ability? 7 A. Well, I was -- I did see that statement elsewhere. 8 I didn't hear the evidence from Mr Oppenheim, but I have 9 seen in one of the reports the suggestion that ICL staff 10 or ICL Pathway staff would somehow have access to 11 transaction data. I found that an amazing allegation or 12 suggestion and, in my view, that was never contemplated 13 or happened. It was never suggested it should happen 14 and I really don't know why and where anyone should 15 think that could have taken place. 16 MS KENNEDY: Mr Bennett, I don't have any further questions 17 for you. 18 Chair, do you have any questions for Mr Bennett? 19 SIR WYN WILLIAMS: No, thank you. 20 MS KENNEDY: If I could pass to Mr Jacobs. 21 MR JACOBS: Thank you. Sir, can you see and hear me? 22 SIR WYN WILLIAMS: As usual, Mr Jacobs, I first hear you and 23 now I see you. 24 MR JACOBS: Thank you, thank you. 25 SIR WYN WILLIAMS: There's always a delay between the two. 159 1 Questioned by MR JACOBS 2 MR JACOBS: Yes. 3 Mr Bennett, good afternoon. I have some questions 4 for you on behalf of 153 subpostmasters who are 5 represented by Howe & Co and I want to ask you some 6 questions concerning paragraph 30 of your witness 7 statement, and can I ask that we could just have that on 8 screen. It is WITN04 -- 9 Ah, it is here already. Thank you very much, 10 Frankie. 11 So you say: 12 "Throughout the rollout the Post Office had full 13 visibility of ICL Pathway's key activities including all 14 the incident recording and resolution processes." 15 Now, we know, Mr Bennett, from the findings that 16 were made in the High Court, that the Post Office was 17 unable to access audit data or ARQ data, which was held 18 by Pathway and are you aware of what that is, are you 19 familiar with that? 20 A. I'm afraid I'm not. I assume that's information 21 captured at the counter when transactions were 22 processed. 23 Q. Absolutely correct, yes. It's the keypad activities, 24 what keys are pressed -- 25 A. Yes. 160 1 Q. -- when they are pressed by the subpostmasters. It 2 emerged that Post Office had to request that from 3 Pathway and, after a certain amount of requests, then 4 Fujitsu were contractually entitled to charge them for 5 providing that information. 6 I want to ask you about what you say about 7 visibility of Pathway's activities. Are you aware of 8 any other data or key material, which was held by 9 Pathway, which Post Office was unable to access without 10 requesting it of Pathway? 11 A. Well, first of all, I don't -- I mean, your original 12 comment about audit data, I didn't realise that this 13 wasn't available to Post Office. It was captured by 14 Pathway because we captured everything which went on, so 15 I could see why people would want access to it, I can 16 understand that. I can't understand why it wasn't 17 available. 18 Taking your second point, can I think of any other 19 areas, well, I dare say someone will find something 20 which we didn't pass on but I would make the general 21 point that, under my direction, most of my staff 22 operated a very transparent process of the work we were 23 doing and the problems we were having and we would share 24 them as appropriate with people who needed them. This 25 was a technical programme to get right and there was no 161 1 value in not sharing critical information. So 2 I can't -- I can't think of anything, no. 3 Q. That's helpful, and perhaps you might not be able to 4 answer my next question, but I will ask it anyway. 5 You said in your evidence today that you were 6 working very closely with Post Office. 7 A. Yes. 8 Q. What mechanisms were available for Post Office to access 9 data held by Fujitsu? If they wanted to do so, how 10 would that be done? 11 A. Well, I suspect one of the key providers or key conduits 12 for that would be the PDA. 13 Q. Right. 14 A. Because, I think as we discussed a lot earlier, most of 15 our contact with the Post Office had to be through the 16 PDA. 17 We had very little direct, technical programme 18 contact, directly with the Post Office themselves. 19 I think the PDA was our source of interface. 20 Q. Thank you. 21 A. I mean, I don't -- I'm quite sure, from time to time, we 22 met with Dave Miller and his people and all sorts of 23 things but, in terms of programme management and 24 programme development, the sponsors were very clear they 25 saw us routing our information, both ways, through the 162 1 Programme Development Authority. 2 Q. But you're not aware of a specific mechanism? 3 A. Not as a mechanism, no. 4 Q. Okay, and just for completeness perhaps, you have heard 5 that some information Fujitsu was contractually entitled 6 to charge Post Office for. Were you aware of charges 7 for any other information or data? 8 A. In my experience, we charged for virtually nothing for 9 five years on this programme. I mean, one of the big 10 issues for us was that we worked for five years without 11 any payment and our payments didn't start until pretty 12 well the end of 1999. 13 Q. Yes. 14 A. Even though we started work in 1994 and, in-between 1994 15 and 1999, I would consider the income we would have 16 earned through any charging mechanism to be 17 infinitesimally small. 18 MR JACOBS: Well, thank you. I might have some further 19 questions for you but I'm just going to ask those who 20 instruct me if there's anything else that I need to ask 21 you. 22 I don't have anything else for you. Thank you very 23 much. 24 MS KENNEDY: Chair, apologies, I understand that Mr Henry 25 may be putting three documents to Mr Bennett, I believe 163 1 two of which Mr Bennett has seen but one of which has 2 not been seen and I believe someone should be handing 3 a copy of that now. 4 Unfortunately, one of the documents, which is 5 an organogram, is not on the system and can't be shown 6 on the screen. 7 MR HENRY: The organogram is actually POL00089867. 8 MS KENNEDY: I'm told it's not with RTS and can't be shown 9 on the screen. 10 MR HENRY: Oh, I'm so sorry. That's a shame, because I had 11 quite a number of questions to put on it but not to 12 worry. Let me see. 13 Questioned by MR HENRY 14 MR HENRY: Mr Bennett, I'm very sorry, what have you got in 15 front of you now? Have you got -- 16 A. I've got two documents I have seen before -- 17 Q. Yes, and you've got one page. 18 A. -- and I've got a one-page organisation chart. 19 Q. What a shame because, although I gave that as 20 a reference, the organogram, on reflection, I would 21 quite like to put about ten pages out of that 22 presentation, which of course you and ICL Pathway gave 23 in a sort of all-day session with POCL on 24 19 November 1996 and you couldn't possibly remember that 25 all-day session, but let's just go to the page you have 164 1 got and that shows the accountability hierarchy at 2 ICL Pathway in 1996, doesn't it? 3 A. Yes. 4 Q. So we have you as managing director, correct? 5 A. Correct. 6 Q. Then if you're the admiral, your rear admirals or vice 7 admirals are director quality and risk management, 8 Martyn Bennett, correct? 9 A. Correct. 10 Q. Director of commercial and finance, Tony Oppenheim, 11 okay? So they sit, as it were, to your left and right. 12 A. I think that's drawing too much from this chart. 13 Q. Really? 14 A. I think you should consider all these people as my 15 direct reports. 16 Q. So they are your direct reports? 17 A. Yes, rather than some form of different -- it doesn't -- 18 Q. But you are the managing director? 19 A. Oh, yes, but I don't know why you refer to these two -- 20 I consider them all to be my direct reports, simple as 21 that. 22 Q. I see, all right. But you are the managing director? 23 A. Correct, yes. 24 Q. Oh, yes, quite clearly. Then we have a line, a direct 25 and unbroken line, like a plumb line, down to programme 165 1 director Terry Austin, so he was your direct report as 2 well? 3 A. Correct. 4 Q. Right. Then we have a business development director, 5 Liam Foley, we have customer services director -- is it 6 Stephen Muchow? 7 A. Stephen Muchow. 8 Q. Muchow, right. I won't -- customer requirements 9 director John Dicks, et cetera, et cetera. 10 Could I just ask you to consider various things and 11 I'm going to make it absolutely plain to you that, if 12 you cannot remember, then you must obviously say so, 13 because the last thing I would want to do is to put you 14 in an invidious position. But let me just ask you about 15 what POCL wanted because you had, up until a very late 16 stage, as it were, two customers, didn't you? 17 A. Yes. 18 Q. Yes, and the dominant party was the DSS? 19 A. As far as I'm concerned, they were joint partners, they 20 were joint customers. 21 Q. Really? 22 A. They each had their own requirement but they were both 23 equal standing customers. 24 Q. But 30 per cent of POCL's revenue came from DSS BA, 25 didn't it? You knew that? 166 1 A. I -- well, no, I didn't know that. 2 Q. I mean, would you not have become familiar with what 3 your client, or one of your clients depended upon? 4 A. I was very well aware that Benefits Agency were one of 5 the most important customers POCL had. I didn't know it 6 was 30 per cent. 7 Q. You, of course, were present, weren't you, at a meeting 8 which took place on 3 July 1998, correct, with Mr Keith 9 Todd and also Frank Field? 10 A. Yes. 11 Q. You have that, it's FUJ00075721. Do you recall that 12 Mr Field, the Minister, talked about the Post Office 13 dependency culture and the concern in maintaining the 14 Benefit Payment Card? 15 A. Can you refer me to which paragraph? 16 Q. Yes, of course. We have a reference to the Benefit 17 Payment Card at 4.5, do you see? 18 A. Yes. 19 Q. "We spent a lot of time on the benefits of the payment 20 card and not surprisingly picked up this question of 21 part payments. The minister was particularly keen to 22 know that our system was quite capable of handling part 23 payments and that the constraint was essentially a BA 24 rule." 25 Then we considered -- 4.7: 167 1 "... some time discussing the wider use of cards 2 across government to connect various government services 3 together." 4 A. Yes. 5 Q. 4.8: 6 "We discussed the need for the Post Office to move 7 to a more commercial outlook and Keith explained that 8 their difficulty with payment systems was more to do 9 with concern about their funding arrangements and that 10 if the Post Office was block funded then they would 11 probably take a more relaxed and open minded view of how 12 to meet modern payment systems. Frank Field called this 13 the Post Office dependency culture." 14 But then 4.9: 15 "There was no strong reaction to our key comment 16 that the progress to ACT [Automated Cash Transactions] 17 was inevitable but would take time and had to be managed 18 alongside re-engineering of the Post Office network." 19 Now, you were effectively drafting this diary note 20 from that meeting, weren't you? 21 A. Yes, I wrote this note. 22 Q. Absolutely. So what I'm trying to suggest to you is 23 that that was primarily the concern at that time and, so 24 therefore, the DSS was the dominant partner? 25 A. I'm prepared to accept your opinion. 168 1 Q. Could I just ask you, please, to just help me a little 2 bit about Mr Todd's managerial style. 3 So far as you were concerned, you were reporting 4 back to him, were you, candidly and frankly everything 5 that was going on from those who were directly reporting 6 to you? 7 A. Yes, I was his direct report on this project. It was 8 a bit unusual because most of the business in ICL went 9 through established business divisions. We had a retail 10 division, we had a government division, we had a finance 11 division, et cetera. But this programme was sufficient 12 size and importance that this organisation broke the 13 normal organisation rules. So, rather than reporting 14 through an established line business, which is what we 15 would normally do, I was a direct report to my 16 chief executive. 17 Q. What about Sir Graham Corbett? 18 A. Graham Corbett? He was nothing to do with -- 19 Q. The chairman of ICL? 20 A. Graham Corbett was not the chairman of ICL. You have 21 the wrong man. 22 Q. Forgive me, I will withdraw that. So you were, 23 therefore, tasked with being the messenger to Mr Todd? 24 A. I reported to Mr Todd and he was my boss. 25 Q. Yes. 169 1 A. That's not a -- 2 Q. I didn't mean that pejoratively -- 3 A. Well -- 4 Q. -- but if you had to bring him bad news -- 5 A. I reported everything to Mr Todd. 6 Q. You reported everything to Mr Todd. Clearly, at that 7 time, sir -- and I note what you say about gradually, as 8 it were, distancing yourself from direct responsibility 9 from about, what do you say, 1999? 10 A. Well, the end of 1999/early 2000. 11 Q. Yes, but clearly, at that time -- if we go over the page 12 to page 3 and paragraph 7 -- you were clearly the -- 13 A. Sorry, can we just -- which document -- oh, this is the 14 same document. 15 Q. Same document, sir, yes. 16 A. Yes, I've got it? 17 Q. Paragraph 7 which is on page 3? 18 A. Page 3, which paragraph? 19 Q. Paragraph 7. 20 A. Right, thank you. 21 Q. This was a list of the action points that arose from 22 that meeting and, clearly, with one exception, you were 23 deputed or, of your own initiative, had to execute the 24 vast majority of those actions, didn't you? 25 A. Yes. 170 1 Q. So, for example, the draft letter confirming the 2 parallel approach, the sort of compromise approach, 3 correct? 4 A. Well, all these actions I can see are directed to me to 5 discharge. 6 Q. Yes, yes, the visit to Feltham, that's not particularly 7 important. You then talking to Derek Sayers for details 8 of the employment services business, so trying to, as it 9 were, capture more business, correct? 10 A. Yes, that's another aspect of ICL's business. 11 Q. You to write to Mr Field reconfirming Horizon's 12 capability of handling part payments under the BPC, 13 correct? 14 A. That's what it says. 15 Q. Yes: 16 "... to discuss with Terry Reynolds the impending 17 visit of [Mr] Field to ... An Post ..." 18 And family budgeting, in other words his campaign 19 against poverty, you know, to sort of teach people how 20 to manage their own money under the system; would that 21 be right? 22 A. This is the action. 23 Q. Yes. Then: 24 "[Mr Todd] to speak urgently to John Roberts to get 25 much more vigour and energy behind the Post Office's 171 1 move to financial services." 2 A. Yes, I think on that I would say that Keith Todd had 3 a very strong view, and he expressed it on many 4 occasions, he thought the Post Office had a huge 5 opportunity in being more dynamic and active in the 6 banking field and he thought that, just as the Post 7 Office network had a huge coverage in the UK, against 8 a climate of high street banks retreating from towns and 9 villages, that this opened up a new business opportunity 10 for Post Office Counters and he was thinking well beyond 11 the BA/POCL contract as to how some of these more 12 visionary services could be used to enhance the network, 13 Post Office network, and provide a better service to the 14 community. 15 Q. You obviously bought into that vision? 16 A. We did, indeed, although our focus was delivering the 17 contract, not -- 18 Q. Yes, of course? 19 A. -- not thinking too far ahead. 20 Q. But if I could bring you back, therefore, to paragraph 2 21 of that document on page 1, because of course the 22 Minister, his opening remarks were that he was keen to 23 see Mr Todd to talk about social banking, whereas 24 Mr Todd's opening remarks were that he was here to talk 25 about the programme in the round and the key points were 172 1 that Horizon is: 2 "deliverable, that it is critical to ICL as well as 3 DSS, POCL and Government and that the infrastructure 4 being built is essential for all aspects of fraud, 5 welfare reform, the future of the Post Office and all 6 aspects of better government." 7 He was going at it not less than 100 per cent, 8 wasn't he, from the beginning of the meeting? 9 A. Yes. 10 Q. So when you say that you didn't feel under any pressure 11 to deliver, if I ask you -- 12 A. Sorry, can you just repeat what you say I have said? 13 Q. Well, you didn't feel under any great pressure to -- 14 A. Do what? 15 Q. -- deliver the project. You said that earlier. 16 A. No, I didn't say that. 17 Q. Really? 18 A. I said that we were always under -- I'm not sure I would 19 use the word "pressure", but if you run a programme like 20 this you are always energised to get on with the job. 21 Q. Yes. 22 A. I wouldn't call that somehow unrealistic pressure. If 23 you run a programme like this then you, every day, are 24 working hard to make progress. 25 Q. Of course. 173 1 A. I wouldn't call that pressure, I would call that normal 2 operating. 3 Q. Could I ask you -- and I have nearly finished because, 4 obviously, I can't put things to you from a document 5 which you haven't seen and which isn't on the system, so 6 I'm going to abandon that whole tranche. But you appear 7 to say that it was Mr Coombs' decision -- and do correct 8 me if I've got this wrong -- that it wasn't necessary to 9 rewrite the IT; is that right? 10 A. No, I think we were talking about rewriting the code for 11 the EPOSS. Redesign and rewrite a part or all of the 12 EPOSS software. 13 Q. Exactly, that's what I meant. Are you saying that that 14 was his exclusive responsibility? 15 A. He was the technical authority on the programme and 16 I trusted his judgement and his authority was what 17 I followed. His judgement was what carried and 18 I implemented. 19 Q. But, ultimately, were you not -- I'm not saying you were 20 ultimately accountable, I will rephrase that. I mean, 21 obviously, ultimately, those above you are ultimately 22 accountable, but you were the managing director, were 23 you properly and fully sighted on that momentous 24 decision? 25 A. There were many momentous decisions during the six years 174 1 I was managing director of Pathway and I learned, over 2 that time, whose judgement I would follow and who had 3 the authority. Now, Mike Coombs -- if you're talking 4 about Mike Coombs in this case -- was probably the most 5 professional, skilled and experienced man in ICL in 6 programme management and I would not overrule his 7 judgement. 8 Q. But surely this is not simply a technical issue? 9 A. Isn't it? 10 Q. This is a governance issue about: (a) suitability of the 11 programme for your customer; and also (b) as you have 12 alluded to in the answers you have given a short while 13 ago, cost must have played some consideration in this? 14 A. I think it was a technical decision, based upon 15 economies and cost, and I think Mr Coombs in the paper 16 we looked at a moment ago explained the two options 17 which were in front of us and the two options were 18 either to continue to manage the system we had 19 developed, with its benefits and drawbacks, or to 20 involve a major rewrite, which, again, had benefits and 21 drawbacks. His judgement, which is well recorded in the 22 papers, was that we were better off managing the system 23 we had and, at that time, that was what was done. 24 Q. If you had not chosen to go down that path there would 25 have been further delay, would there not? 175 1 A. Well, we didn't go down that path, so I couldn't 2 speculate on that. 3 Q. Well, obviously, if you were going to -- it's not 4 a question of speculation really, it's a question of 5 common sense. If you weren't going to rewrite that 6 EPOSS issue and EPOSS would have been part of the 7 backbone of the system, wouldn't it? The Electronic 8 Point of Sale Service, it would have been part of the 9 backbone of the system for a subpostmaster in his 10 office? 11 A. It was very important to get that decision right. 12 Q. It would have been the backbone, wouldn't it? 13 A. It was a critical -- it was one of the most critical 14 systems in the Post Office. 15 Q. Absolutely. Transaction information processing: again, 16 part of the backbone, correct? 17 A. It's a key system. 18 Q. Inventory management: part of the backbone? 19 A. A key system. 20 Q. Yes, so if you were going to rewrite that there would 21 have been delay, wouldn't there? 22 A. Well, I think there is clearly pros and cons of either 23 two options, and both -- 24 Q. I won't ask the question again. I won't ask the 25 question again. I suggest to you it would have been 176 1 obvious that there would have been delay. 2 Cost. Obviously, you would have had to have 3 invested more in rewriting, wouldn't you? 4 A. The judgement was that that was not the best option 5 available. 6 Q. Again, I won't ask the question again. 7 Opportunity cost. The cost involved in rewriting 8 would prevent you from utilising staff and resources to 9 work on other projects? 10 A. That's absolutely true. Every decision you take has 11 an opportunity cost. 12 Q. Then, of course, Her Majesty's government might have 13 something to say about this, wouldn't it? 14 A. I can't answer that one. 15 Q. Well, you can because, of course, they had already -- 16 and you would have been aware of this -- put you in 17 breach of contract in November 1997. 18 A. We rejected that breach. 19 Q. I'm sorry? 20 A. We rejected that breach. 21 Q. Well, that -- 22 A. And that breach was never carried out. 23 Q. Yes, but the position was, wasn't it, as well, that 24 there were -- I mean, I won't go through them in detail 25 but, eventually, you were proposing, and I suggest that 177 1 you should have been aware of this if you were dealing 2 with this with Mr Todd, that there was a proposal not to 3 do live testing or to drastically reduce testing of the 4 system. Do you recall that? 5 A. I don't recall a debate about reducing testing, no. 6 Q. Were you aware of a problem, as well, about evidence of 7 ownership of assets involved, such as perpetual licences 8 for intellectual property? 9 A. No, I have not -- that's not a subject I have any 10 recollection of. 11 Q. I mean, if I put something to you (unclear) accept it 12 but, isn't it right that An Post, which was supposed to 13 be, forgive me, the poster boy for a wonderful system, 14 was, in fact, in April 1998, in a massive court case and 15 dispute with no less than Escher over who actually owned 16 the IP? 17 A. I don't remember that at all but I take your word for 18 it. 19 Q. Well, because I want -- final this: when you say that 20 you, in fact, put the source code into escrow in case 21 anything might have happened, and you said that earlier 22 this afternoon -- 23 A. Yes. 24 Q. -- you must have had a memory of why that was? 25 A. We put it there because we saw the risk of a small 178 1 company not being in business during the life of this 2 contract. 3 Q. Forgive me, sir, were Escher aware that you put their 4 source code into escrow? 5 A. They certainly were. 6 Q. Were you aware -- 7 A. We had -- I mean -- sorry, I don't mean to interrupt, 8 but we couldn't put their source code into escrow unless 9 they gave it to us. 10 Q. Thank you, because obviously then -- I mean, weren't 11 they supposed to be bought out by Andersen Consulting? 12 A. I don't remember that. 13 MR HENRY: You don't remember that. Well, thank you very 14 much for answering my questions. 15 SIR WYN WILLIAMS: Is there anyone else who is intending to 16 ask any questions? 17 MS PATRICK: Yes, sir, Ms Patrick. 18 Questioned by MS PATRICK 19 MS PATRICK: Mr Bennett, my name is Ms Patrick. I ask 20 questions for 64 subpostmasters and I'm instructed with 21 Mr Maloney KC, who sits beside me, by Hudgells 22 Solicitors. 23 I only have a few questions for you, following up on 24 two points that Mr Henry was raising with you, a visit 25 to Feltham, but instead of looking backwards, I may be 179 1 looking forwards, as it were. 2 The potential for the future commercial exploitation 3 of Horizon was important, both for ICL and for POCL? 4 A. Yes. 5 Q. But first Horizon had to work and be seen to work? 6 A. Yes. 7 Q. Can we look, just to perhaps refresh your memory, at how 8 you closed out 1999 in your end of year communication to 9 the ICL staff. I think you have seen this document but 10 we can bring it up at FUJ00075736, and it is the first 11 page. 12 On the left-hand column -- you see there are two 13 columns to the document and I'm looking at the second 14 paragraph on the first column to start with and I'm 15 going to start a little of the way down, maybe about 16 a third of the way down the paragraph, you will see 17 there's "There": 18 "There are still some important issues which we are 19 working on with the Post Office but we plan to restart 20 rollout on 24th January 2000 at which time we will 21 rapidly reach the rate of 300 post offices per week, 22 completing as planned in spring 2001. Meeting these 23 dates is just as important to the Post Office as to us, 24 since they are keen to get started on the exploitation 25 opportunities, which Horizon provides. The next major 180 1 new activity for both parties will be around banking 2 services. Work here will accelerate rapidly through 3 2000." 4 Now, I'm going to skip ahead a little and, if we 5 look over to the second column at the top, it is visible 6 on screen: 7 "We have recently hosted, here in Feltham, Alan 8 Johnson MP, Minister of State at the DTI, responsible 9 for the Post Office. Like many of his colleagues, he 10 found the session an excellent introduction to the 11 project. We have also been very fortunate to have 12 visits from the Parliamentary IT Committee and the 13 Performance & Innovations Unit from HM Treasury during 14 November. Those attending were particularly excited by 15 the impact that this project can have on the way that 16 government services are delivered in the future, 17 exploiting the natural strengths of the Post Office 18 network." 19 It goes on: 20 "2000 will be a challenging and exciting year for us 21 all and I thank you all in advance for all your efforts 22 and wish you and your family a Happy Christmas and 23 New Year." 24 You see it is signed off there at the bottom by you; 25 is that correct? 181 1 Now, POCL and ICL, and the government, were 2 particularly excited by the future commercial 3 exploitation of Horizon, weren't they? 4 A. I'm sorry, I didn't -- I have seen this. Could you just 5 repeat the question? 6 Q. So we have seen what you said at the end of 7 December 1999 to your staff. POCL, ICL and the 8 government were particularly excited by the future 9 commercial exploitation of Horizon, weren't they? 10 A. I believe so, yes. 11 Q. Now, looking at that, we have seen the meetings that 12 were going on with ministers and Parliamentarians to 13 explain the system. Did those meetings include 14 an update on just how many receipts and payments 15 imbalances had occurred in 1999? 16 A. Are you quoting from a document or is that -- 17 Q. No, I'm just asking you a question. 18 A. I see, yes. 19 Q. You were involved in those meetings, you knew they had 20 happened. In those meetings with ministers and with 21 Parliamentarians to explain the system, did those 22 meetings include an update on how many receipts and 23 payments imbalances had occurred in 1999. 24 A. I'm sorry, maybe my hearing is bad but I don't quite -- 25 could you just. 182 1 Q. When the MPs -- sorry, it has been a long day, 2 Mr Bennett. I will go really slowly. 3 When you had those meetings -- 4 A. Yes. 5 Q. -- with MPs -- 6 A. Yes. 7 Q. -- and with ministers to explain the system -- 8 A. Yes. 9 Q. -- did they include an update for them on how many 10 receipts and payment imbalances had occurred in 1999? 11 A. I -- I do recall that element, no. I don't -- I doubt 12 whether that was discussed. I mean I think the main 13 purpose was to show them an example of how the system 14 worked and raise their sight as to how it could be 15 utilised in the future. I don't think we discussed any 16 particular issues or programme activities which were of 17 the subject you have just raised, so I doubt whether we 18 did discuss that. 19 Q. Thank you, Mr Bennett. 20 Our last point: did the meetings involve 21 an explanation of the continuing recommendation from 22 some within ICL that the EPOSS required a redesign or 23 a rewrite? 24 A. It wouldn't have been discussed in those meetings, no. 25 MS PATRICK: Thank you, Mr Bennett. That's all the 183 1 questions we have. 2 MS KENNEDY: Chair, I -- 3 SIR WYN WILLIAMS: Mr Bennett -- Ms Kennedy, there's one 4 aspect of Mr Bennett's evidence that I think may require 5 a little further clarification and that relates to the 6 questions you asked him about the potential for remote 7 access -- I use that phrase as a very shorthand 8 phrase -- and what's concerning me is that I haven't got 9 before me the transcript of Mr Oppenheim's evidence 10 yesterday, which was similar to Mr Bennett's evidence in 11 the sense that he didn't think that remote access in the 12 strict sense could be achieved, but he did give quite 13 a detailed account of how use of the central servers 14 might have an effect on branch accounts, which ought, if 15 the system was operating properly, to have been flagged 16 up in the data produced. 17 Now, I'm loath to ask Mr Bennett questions about 18 that without having precisely what Mr Oppenheim said 19 before me so that I can quote him accurately to see 20 whether the two of them agree or disagree about this 21 because, as you will know, in the civil litigation there 22 appeared to be an acceptance that remote access -- and 23 I use that again somewhat loosely -- was not just 24 a possibility, it occurred. 25 So, as it seems to me, there is this area of 184 1 evidence which may need to be clarified and I'm saying 2 all this publicly because I'm contemplating asking 3 Mr Bennett to make a short supplementary statement, once 4 we have a transcript of Mr Oppenheim's evidence, so that 5 he can say whether or not he agrees with it. 6 That's a very long-winded statement by me, but 7 I wanted to make it public to everyone because I don't 8 want this aspect of the case to be left up in the air, 9 so to speak, with different nuances and different 10 accounts from different witnesses, unless ultimately 11 that is the state of the evidence, so to speak. 12 There we are. It's not a question to you, 13 Mr Bennett, but it is, now that I have said what I have 14 said, raising the possibility that at some stage the 15 Inquiry might write you a letter with an extract from 16 Mr Oppenheim's evidence and ask you to comment on it. 17 You understand that? 18 A. Yes. 19 SIR WYN WILLIAMS: Yes, thank you. So that's it for this 20 afternoon, Mr Bennett, and I would like to thank you for 21 making your written statement and coming into the 22 Inquiry to give oral evidence. 23 A. Thank you. 24 SIR WYN WILLIAMS: So one witness tomorrow, Ms Kennedy? 25 MS KENNEDY: Yes, Mr Miller. 185 1 SIR WYN WILLIAMS: Can we use that well-known legal phrase 2 "not before 10 o'clock", just in case my train is late 3 tomorrow? 4 MS KENNEDY: Yes, thank you. 5 SIR WYN WILLIAMS: All right, fine. 6 (4.27 pm) 7 (The Inquiry adjourned until 10.00 am on Friday, 8 28 October 2022) 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 186 1 2 INDEX 3 4 TERENCE AUSTIN (affirmed) ............................1 5 Questioned by MR BLAKE ...........................1 6 Questioned by MR STEIN ..........................73 7 Questioned by MR HENRY ..........................86 8 Further questioned by MR BLAKE .................103 9 Questioned by SIR WYN WILLIAMS .................104 10 JOHN BENNETT (sworn) ...............................105 11 Questioned by MS KENNEDY .......................105 12 Questioned by MR JACOBS ........................160 13 Questioned by MR HENRY .........................164 14 Questioned by MS PATRICK .......................179 15 16 17 18 19 20 21 22 23 24 25 187