1 Tuesday, 26 September 2023 2 (10.00 am) 3 Statement by MR BEER re Disclosure 4 MR BEER: Sir, good morning, can you see and hear 5 me? 6 SIR WYN WILLIAMS: Yes, I can, thank you. 7 MR BEER: Sir, before we start Mrs Chambers' 8 evidence this morning, there's just a short 9 statement I would like to read, if I may, about 10 some recent disclosure. 11 SIR WYN WILLIAMS: Yes, of course. 12 MR BEER: Sir, you will recall that on Tuesday of 13 last week I provided an update to the Inquiry on 14 the Post Office's late disclosure to the 15 Inquiry. I noted that, since 3 July 2023, so 16 just before the first disclosure hearing, the 17 Post Office had disclosed some 23,000 18 potentially relevant documents to the Inquiry of 19 which some 15,200 are said by the Post Office to 20 relate to Phase 4 of the Inquiry. I noted that 21 this meant that a high number of potentially 22 relevant documents had been disclosed in the 23 recent past by the Post Office and many of them 24 were presently being processed by the Inquiry. 25 I also said that nobody should be surprised 1 1 if witnesses had to be recalled as a consequence 2 of this and other disclosure issues. 3 On Friday afternoon and yesterday at midday 4 the Inquiry received further updates from the 5 Post Office regarding its disclosure. The 6 update related to an issue that had been 7 highlighted by the Post Office to the Inquiry 8 previously, known as the back-up tapes. It 9 relates to 37 back-up tapes that the Post Office 10 located at one of its sites. The Post Office 11 had provided an update on the back-up tapes on 12 22 August 2023 in a letter to the Inquiry, 13 disclosed to the Core Participants prior to the 14 5 September disclosure hearing. 15 Just for your reference, no need for it to 16 be displayed, the document reference for that 17 letter is POL00124517. 18 That update had noted that the Post Office 19 had identified some 42 back-up tapes with 20 unknown contents at a site and that the Post 21 Office had, for some time, notified the Inquiry 22 of the steps that it was taking to understand 23 the contents of the back-up tapes, which the 24 Post Office said were substantial in volume, 25 including in particular whether the back-up 2 1 tapes might relate to matters being investigated 2 by the Inquiry. The number of the back-up tapes 3 was later changed in correspondence to B37. 4 However, the Post Office noted at that time 5 that, as the back-up tapes: 6 "... ordinarily reflect a copy of the data 7 that exists elsewhere, they are not ordinarily 8 considered as key repositories to process and 9 search for the purposes of disclosure." 10 The Post Office also noted that: 11 "The files recovered from the dataset 12 sessions on the tapes may be partially or 13 significantly duplicative of files already 14 harvested, searched and disclosed to the 15 Inquiry." 16 The Post Office provided a detailed 17 description of the steps that it was taking 18 before it begin to investigate the back-up tapes 19 in that 22 August 2023 letter. This update was 20 detailed and largely of a technical nature. 21 On 31 August 2023 the Post Office provided 22 further back-up tapes, this was also provided to 23 the Core Participants before the 5 September 24 disclosure hearing. That's POL00126338, no need 25 to display. 3 1 The Post Office noted that it had started 2 its investigation into the tapes and was taking: 3 "... a two-pronged approach, firstly, using 4 technical policy-based and other means to 5 identify whether the back-up tapes contained 6 meaningful new information and, secondly, 7 concurrently taking steps to allow the Post 8 Office to review the actual data on the tapes to 9 identify whether the tapes contain new 10 information." 11 At that time, the Post Office noted that it 12 was not yet able to confirm how long certain 13 steps would take but that it would continue to 14 keep the Inquiry informed and would welcome our 15 thoughts on an appropriate approach to adopt. 16 Last Friday afternoon at 3.08 pm, the Post 17 Office sent a further update to the Inquiry, 18 regarding the 37 back-up tapes. That back-up 19 tapes update highlighted three key points. 20 Firstly, an initial analysis of the tapes as 21 performed by KPMG suggested that the content of 22 the tapes are largely non-duplicative. In other 23 words, while the tapes are called back-up tapes, 24 they contain new material. They're made up of 25 documents that are not already in the Post 4 1 Office's Relativity platform from which it 2 reviews documents for potential disclosure to 3 the Inquiry. 4 Secondly, after performing initial searches, 5 it seemed that a significant number of the 6 back-up tape documents may be relevant to the 7 Inquiry's terms of reference, including Phase 4. 8 Third, the date range of the parent emails 9 on the back-up tapes was primarily 2001 to 2008. 10 The Post Office noted that this was a period 11 where their electronic disclosure was more 12 limited and the Post Office had therefore relied 13 more heavily on their hard copy materials. In 14 this sense, the Post Office has, it said, 15 "a present working assumption", that this 16 material may be unique at least in the main. 17 The Post Office noted that it was intending 18 to prioritise the review of this material on 19 a witness-by-witness basis and, with this in 20 mind, the Post Office was prioritising the 21 documents responsive to search terms for Anne 22 Chambers and Mandy Talbot. It noted that it had 23 run some searches over the documents for "Anne 24 Chambers" and "Mandy Talbot", both in speech 25 marks, and this had generated some 208 and 1,060 5 1 hits respectively, excluding families. 2 The Post Office noted that they would 3 provide relevant documents relating to 4 Mrs Chambers as soon as possible and envisage 5 doing so informally by end of day on Friday. 6 Sir, the Inquiry legal team considered this 7 information urgently on Friday afternoon. In 8 order to get a better understanding of the 9 issue, the legal team asked the Post Office to 10 indeed provide the documents responsive to the 11 search terms for Anne Chambers as soon as 12 possible and also asked the Post Office to 13 provide search hit results for the Phase 4 14 witnesses who are to be called between today and 15 Friday, 20 October 2023: namely John Jones, 16 Simon Recaldin, Catherine Oglesby Andrew 17 Hayward, John Scott, Rob Wilson, Paul Inwood, 18 Thomas Pegler, John Breeden, Alan Lusher, Alison 19 Bolsover and Marie Cockett. 20 The Inquiry asked for two sets of results, 21 one with family members and one without, no 22 later than 11.00 am yesterday. At 7.03 pm on 23 Friday evening the Post Office provided 24 documents that had been responsive to Rule 9 25 requests or otherwise said to be of interest 6 1 from the "Anne Chambers" search hits of 208 2 documents. They provided 84 documents in total. 3 In the meantime and over the weekend, the 4 Inquiry Team urgently considered the 84 new Anne 5 Chambers documents. The Inquiry is of the view 6 that none of the material is sufficiently 7 relevant to Phase 4 issues, however that 8 disclosure did contain new material that may be 9 relevant to Mrs Chambers' evidence in relation 10 to Phase 3, in particular in relation to 11 knowledge of and action taken in response to 12 bugs, errors or defects. 13 At 12.14 yesterday, the Post Office provided 14 the search hit results for the remaining four 15 Phase 4 witnesses to be called between 16 29 September and 20 October. This included 17 expanded search terms beyond just "first name" 18 and "last name" in inverted commas. 19 Sir, as you would expect, the search results 20 from the back-up tapes vary from witness to 21 witness. One had no hits, some have hits in the 22 hundreds and some have hits in the tens of 23 thousands. Indeed, one witness has some 93,699 24 hits when you include family members in 25 a broader search. 7 1 Last week, I noted that, as had been made 2 clear on previous occasions, when the Inquiry 3 has addressed the issue of late disclosure, all 4 of those interested in the work of the Inquiry, 5 including but not limited to witnesses and Core 6 Participants, should understand that the fact 7 that the Inquiry has decided to continue to hear 8 evidence does not mean that witnesses from whom 9 evidence is about to be taken will be giving 10 evidence just once. 11 I also noted that the Inquiry will not 12 hesitate to request further witness statement or 13 witness statements from witnesses and call 14 witnesses back to give evidence, in the event 15 that sufficiently relevant material is either 16 disclosed before the witness gives evidence but 17 the Inquiry hasn't had the opportunity to 18 process it or such evidence is disclosed after 19 a witness has given evidence. 20 This not only to ensure that all 21 sufficiently relevant material is put to 22 witnesses but also in fairness to witnesses, so 23 they have the opportunity to address all 24 sufficiently relevant material. 25 On behalf of the Inquiry, I repeat those 8 1 comments now. Like other late disclosures, once 2 the Inquiry has had a reasonable opportunity to 3 analyse the material that the Post Office has 4 provided from the back-up tapes, the Inquiry 5 will disclose that material to Core Participants 6 and witnesses as required. 7 If this necessitates the need for 8 supplemental requests, then those supplemental 9 requests will be sent. 10 Having spoken with you, sir, I make the 11 following additional comments: on 15 September 12 you published a statement following the 13 5 September directions hearing. You stated that 14 there was a need for close monitoring of the 15 disclosure process during the remainder of the 16 Inquiry, especially as it relates to disclosure 17 from the Post Office. You also directed that 18 the Inquiry would hold a further hearing to 19 consider disclosure issues, on a date to be 20 fixed in the period commencing 8 January and 21 ending 19 January 2024. 22 Sir, your legal team consider that the 23 material that is disclosed as a result of the 24 back-up tapes ought to be closely monitored, 25 alongside other issues in the lead-up to the 9 1 further disclosure hearing and, if witness 2 evidence is required on the back-up tapes, both 3 in relation to specific documents that come to 4 light results of that disclosure or on Post 5 Office's updates in relation to the back-up 6 tapes more generally, then it will be sought in 7 accordance with your directions. 8 Sir, that's all I would propose to say on 9 the issue of disclosure now. 10 Sir, I think you're still muted. 11 SIR WYN WILLIAMS: I said thank you, Mr Beer, and 12 I was would just like to add that, as I hope has 13 been clear from Mr Beer's statement, he has been 14 consulting me throughout the period since Friday 15 afternoon about how we should approach what he 16 has described in relation to the Post Office 17 disclosure. 18 I wish to make it clear that I did consider 19 whether it was appropriate to stop the process 20 of evidence gathering. Having reflected upon 21 that, I decided it wasn't, because I am 22 satisfied, at least at the moment, that we can 23 deal with the evidence of Mrs Chambers and those 24 who follow her immediately and also hopefully 25 deal with the appropriate disclosure to Core 10 1 Participants on necessary documents which have 2 been recently disclosed to us. But, as Mr Beer 3 has stressed, I am keeping all of this under 4 very close review and, if it is necessary to 5 change the approach that I have determined is 6 appropriate, at least for Mrs Chambers' 7 evidence, I won't hesitate to change my 8 approach. 9 So that's all I wanted to say about that. 10 Thank you, Mr Beer. 11 MR BEER: Thank you, sir. May Mrs Chambers be 12 sworn. 13 SIR WYN WILLIAMS: Well, does she need to be? That 14 was the question I was going to -- she was sworn 15 on the last occasion. I'm very happy for her to 16 be resworn, but is it strictly necessary? 17 MR BEER: Sir, I reflected on that and I read it the 18 ending of the transcript of the last session and 19 you released Mrs Chambers because there would 20 have been a need for her to speak to her legal 21 representatives and I took that to be a release 22 from the oath, essentially. 23 SIR WYN WILLIAMS: All right, well, I think 24 Mrs Chambers, unless you have any objection to 25 the course of action proposed by Mr Beer, it's 11 1 probably sensible that you be resworn. I'm not 2 going to repeat it but the statement I made to 3 you at the beginning of your evidence back in 4 May, about self-incrimination, applies equally 5 to the evidence you're about to give, all right? 6 ANNE OLIVIA CHAMBERS (re-affirmed) 7 Questioned by MR BEER 8 MR BEER: Thank you, Mrs Chambers. As you know, 9 I ask questions on behalf of the Inquiry. You 10 gave evidence on 2 and 3 May 2023, on all of the 11 first day and for some of the second day, about 12 the operation of the SSC, about other service 13 support offered by Fujitsu to the Post Office, 14 about the operation of PinICLs, PEAKs and KELs, 15 about problem management between Fujitsu and the 16 Post Office, and then, for the balance of the 17 second day, about some bugs, errors or defects 18 in the Horizon System. 19 I said that I would continue to examine some 20 other bugs, errors or defects when you came back 21 to give your Phase 4 evidence. I shall do that 22 probably tomorrow and I should say I'm not going 23 to look at bugs 19 and 23 when we get to them. 24 They, as it seems to me, are adequately 25 addressed in Mr Justice Fraser's judgment and in 12 1 other documents that the Inquiry now has. So 2 you're principally here to give evidence about 3 the work that you undertook in Mr Castleton's 4 case. 5 A. Yes. 6 Q. Before I get to either the additional bugs or 7 Mr Castleton's case, I'd just like to pick up on 8 four topics that we addressed on the last 9 occasion where we've now got some additional 10 disclosure that I would like to ask you about. 11 Firstly, ARQs and the SSC. 12 A. Okay. 13 Q. At the end of the last evidence session, this is 14 on 3 May 2023, Mr Moloney asked you about a work 15 instruction that you had been asked to write by 16 the SSC management team in August 2011. 17 A. Yes. 18 Q. There's no need to display the document at the 19 moment but if you want to refresh your memory we 20 can go back to it. The document is FUJ00138385. 21 It was suggested to you that a reason that 22 the management team may have asked for all 23 issues concerning litigation to be forwarded to 24 the SSC management team for sign-off might be 25 a financial reason. 13 1 A. Yes. 2 Q. Yes, that motive or purpose was suggested to you 3 and you said, in summary -- and I do 4 summarise -- well, that would really be an issue 5 for your managers to answer what their 6 motivation or purpose was? 7 A. Yes. 8 Q. Can we look at a document to see whether this 9 sheds light on the issue, please. It's 10 FUJ00154665. Can you see that this is an email 11 thread dated 8 August 2007? 12 A. Yes. 13 Q. It's copied to you, you'll see you're last of 14 the copy-ees? 15 A. Yeah. 16 Q. But it's a message to Penny Thomas and Peter 17 Sewell -- 18 A. Yes. 19 Q. -- from Mik Peach, your then line manager, yes? 20 A. Yes, that's right. 21 Q. If we just read the message, Mr Peach says: 22 "Penny, 23 "I am not saying that you are confused about 24 the difference between an ad hoc request, an ARQ 25 and a support call -- I am saying that the 14 1 customer is either confused about the 2 difference, or else is making a deliberate 3 attempt to avoid the cost of raising ARQs or 4 ad hoc data requests by raising these as support 5 calls. 6 "Bottom lines for SSC on these problems is 7 as follows: 8 "(a) If it is believed there is a system 9 problem which has caused discrepancies, then we 10 will investigate as normal ... this includes the 11 calls passed over yesterday although none of 12 these calls says that they believe there is 13 an FS problem, all of them actually indicate 14 there is a mismatch in the figures in POLFS, 15 cause unknown. 16 "(b) If it is believed that POL are using 17 the support process as a means of avoiding ARQ 18 or ad hoc data request costs then the calls 19 should be referred back to POL (by Liz?) 20 requesting payment. 21 "(c) If there is any hint of litigation, 22 then we won't deal with the calls as support 23 calls, but will assist the security team in 24 their analysis. 25 "There is a significant difference in the 15 1 system now which is leading to this sort of 2 call, and why there needs to be a more robust 3 application of the process -- in the past, 4 reconciliation on the system was done in two 5 different streams within the FS domain -- if 6 there was a reconciliation issue, or mismatch in 7 the figures then it had to be in our domain 8 somewhere (even if it was chord by [postmaster] 9 user error). 10 "The new system means that much of the 11 reconciliation and auditing figures are produced 12 by POLFS, which is not in the FS domain, is 13 a POL system and is managed for them by PRISM. 14 "Regards 15 "Mik." 16 You'll see that, in the course of that 17 message, Mr Peach says expressly -- he expressly 18 states that a concern is that the Post Office is 19 using referrals to the SSC for investigations to 20 bypass requests for ARQ data that are 21 chargeable; can you see that? 22 A. Yes. 23 Q. Now, although this was before the preparation of 24 the work instruction that you were taken to, 25 does it assist you with your memory as to what 16 1 the reasons for or the motives were for asking 2 for the work instruction? 3 A. That was the management's line, if you like, and 4 had obviously been so for some time. SSC were 5 primarily there to investigate problems as they 6 happened. If you get to the point that it's 7 months down the line and the evidence is needed, 8 then it was certainly harder for us to 9 investigate because we would need to effectively 10 get the data out of audit anyway and, possibly, 11 this was then seen as, if it's at that point in 12 the process, then should Post Office have been 13 paying for it? 14 But I still think this is a question for my 15 management, who were the ones saying this, 16 rather than for me. 17 Q. I understand. You see in the first paragraph 18 Mr Peach raises the suggestion that the client, 19 POL, may be making a deliberate attempt to avoid 20 the cost of raising ARQs, essentially by getting 21 the SSC to do the work, and then in (b) says: 22 "If it is believed that POL are using the 23 support process as a means of avoiding ARQ or 24 ad hoc data request costs ..." 25 Can you recall whether this was a theme in 17 1 your time that was spoken about within the SSC? 2 A. I think it was seen that this was an additional 3 service that Post Office were expected to pay 4 for, if they needed this data, and that was part 5 of -- I assume, was part of the contract between 6 the two companies but I don't know that. 7 Q. Although you may be right that the motivation 8 for the request for the work instruction is 9 a matter for your managers, either Mr Peach or 10 Mr Parker, I'm asking you: can you recall 11 discussion within the SSC about the facility 12 being abused, essentially, by the client, in 13 order to avoid the costs of ARQ data? 14 A. It wasn't something we sat there talking about 15 every day. It's more likely that there were 16 times, perhaps, when I was being helpful and 17 perhaps doing more things than I should have 18 been, that should have been charged for. 19 Q. The paragraph, two from the bottom, beginning 20 "There is a significant difference"; can you see 21 that? 22 A. Yes. 23 Q. Can you read that to yourself or reread it to 24 yourself and assist us, if you're able, with 25 what it means. 18 1 A. I think this is talking about the back-end 2 systems, which I cannot remember all the details 3 of, but there was a big change at some point, 4 and Post Office's back-end database was now 5 something called POLFS, which was not, as it 6 says, part of the Fujitsu domain but was managed 7 by a third party. 8 And so before, when -- even though it wasn't 9 necessarily a front-end of Horizon problem, 10 because investigation would have to be done by 11 Fujitsu, then we would have done that, and Mik 12 is now pointing out that now, you know, you've 13 got two parties resolved so Fujitsu, perhaps, 14 should not be taking on more than their fair 15 share of it. 16 Q. So if an enquiry is raised or an issue is raised 17 by a subpostmaster with the NBSC and work is 18 done, for example, by an investigator to either 19 rule in or rule out a problem on the counter, is 20 this paragraph suggesting that the SSC should 21 take the position that there ought to be further 22 work done behind the scenes at the Post Office 23 to consider any discrepancy or shortfall before 24 the SSC would do any investigatory work? 25 A. No. This is talking about the reconciliation 19 1 between the back-end systems, not the counter 2 reconciliation. But there were cases then when 3 Post Office would be looking at the data that 4 was held in POLFS for a branch and there might 5 be some inconsistency with branch figures, 6 whereas they -- that had been fed through from 7 Horizon. 8 This would not necessarily -- well, it -- 9 I can't remember specific examples but this is 10 not the postmaster at the end of the week or the 11 month saying "I've made a discrepancy"; it's 12 more possibly the two different systems having 13 different numbers and Post Office trying to work 14 out why that might be the case. 15 Q. So Mr Peach isn't saying that the SSC won't do 16 the work: he's just saying that SSC won't do the 17 work for free; is that right? 18 A. He's saying that, in these situations where 19 there's some inconsistency between -- it may be 20 if there's an inconsistency between POLFS 21 something at an individual branch -- I'm not 22 saying -- not sure that he's saying that we 23 didn't -- wouldn't look at that but this is less 24 about individual branch things. I don't think 25 I'm going to be able to get much further with 20 1 this, sorry. 2 Q. Well, looking at the email as a whole, is 3 Mr Peach saying that the SSC will only 4 investigate once data has been requested by the 5 Post Office through the Security team and the 6 Security team has asked for assistance from the 7 SSC? 8 A. It would depend entirely on what the problem was 9 that we were being asked to investigate. 10 Obviously, you know, day-to-day 'happening now' 11 problems, they would come into SSC and that 12 would be our normal investigation. 13 This is things where it is Post Office, 14 probably not NBSC, but the people looking after 15 POLFS and the Post Office financial systems at 16 the back end, or potentially investigating 17 individual branches for litigation, or whatever, 18 those requests would come in through the 19 Security team and I believe it's just sort of 20 pointing out to Penny that she mustn't then just 21 pass them on to SSC as support calls but they 22 needed to be considered as to whether they 23 fitted better into the ad hoc request or the ARQ 24 process. 25 Q. The paragraph at (c), "If there is any hint of 21 1 litigation, then we won't deal with the calls as 2 support calls but will assist the Security team 3 in their analysis"; can you remember whether 4 that was fleshed out at all, what "any hint of 5 litigation" might be? Whether it was the 6 postmaster alleging faults with Horizon, if 7 there were unexplained shortfalls, if there were 8 investigators or auditors involved? What was 9 the trigger for saying, "No, we won't deal with 10 this, this needs to be routed through a Security 11 team"? 12 A. That was a decision for Penny to make, what she 13 was hearing from the people raising the calls 14 within Post Office. 15 Q. So she was the gatekeeper, was she? 16 A. For the type of queries and calls that are being 17 talked about here, it is talking about calls 18 that are coming in through the Security team. 19 Q. What I'm trying to work out is how it worked on 20 the ground, in the light of this email after 21 2007. If a PEAK is raised with the SSC, how it 22 was established whether this was a genuine 23 support call or whether it was something which 24 disclosed a hint of litigation and had to go 25 down a different route? 22 1 A. Yes, I think -- this is talking about the calls 2 that were coming in through the Security team, 3 not the calls that were coming in through the 4 Helpdesk? 5 Q. So the whole email you read as being only about 6 calls coming in through the Security team? 7 A. Yes, yes. That's what it's addressing. 8 Q. Where do you see that? 9 A. Because it's sent to Penny Thomas and is 10 obviously part of a discussion with her and 11 it's -- the top two people on the distribution 12 list are the Security team, and that's who it's 13 sent to, and then it's copied to Service 14 Managers and just a couple of us within SSC who 15 possibly had been dealing with calls that had 16 come in through Penny, and so Mik wanted to make 17 us aware of what he was saying. I really can't 18 remember. 19 Q. Can you recall any instruction or advice that, 20 where there was believed to be a risk of 21 litigation, which, for the most part, meant 22 criminal prosecutions, the Post Office and 23 Fujitsu should work especially closely together, 24 rather than bouncing the issue back to the 25 Security team and saying "Either you pay for 23 1 an ARQ set of data before we do any further 2 investigation work"? 3 A. I think we were expecting, as I say, Penny to 4 talk to her Post Office contacts and find that 5 before taking it over. I mean -- yeah, sorry, 6 I don't think I can answer your ... 7 Q. Yes. Would you agree that, in cases that may be 8 heading towards a criminal prosecution, it was 9 necessary for the Post Office and the SSC to 10 work particularly closely together to ensure 11 that the right data was analysed and, if 12 necessary, harvested and retained? 13 A. I think this was -- that was the responsibility 14 of the Security team, not the responsibility of 15 SSC at that point. If there was examination of 16 that data then needed in the run-up to a case, 17 then I believe the Security team would normally 18 talk to Gareth Jenkins, or somebody like that, 19 to do the analysis and it's possible that 20 somebody like me might then have been roped in 21 to assist. I think, certainly from 2007, you 22 could say, yes, that's what would happen. 23 But if a call had come in through the NBSC 24 from a postmaster saying "I'm making losses and 25 I think it's the system at fault", then, if this 24 1 was a current ongoing situation, then, yes, that 2 would certainly be looked at by SSC in the first 3 instance. The fact that it might at some point 4 later end up as litigation wouldn't stop us from 5 looking at it at that very early stage. 6 Q. Lastly on this document, at the end of (c) "but 7 will assist the Security team in their 8 analysis". Was there any developed, written 9 protocol that set out the steps that would be 10 taken by the SSC in assisting the Security team 11 in their analysis of data where there was a hint 12 of litigation? 13 A. I have no recollection of anything written about 14 that. 15 Q. Was there, short of anything written, any 16 clearly articulated set of requirements on what 17 needed to be analysed, what data needed to be 18 harvested -- 19 A. No, because -- 20 Q. -- and what needed to be retained? It was done 21 on a case-by-case basis; is that right? 22 A. The data would have been the ARQ data and SSC 23 had no access to the servers on which that was 24 stored. 25 Q. Thank you. 25 1 That was the first topic. The second topic, 2 if I may turn to it, Horizon data integrity. 3 Could we look, please, at FUJ00155493. If we 4 just look here, we can see your email at the top 5 of the page to Mr Jenkins. 6 A. Yes. 7 Q. Then if we scroll down, please, we can see 8 Mr Jenkins' email to you and others at the foot 9 of the page. If we just go over the page, we 10 can see he signs it off. So, if we start at the 11 foot of page 1, please, and see what Mr Jenkins 12 said. "All", and you can see this is addressed 13 to Allan Hodgkinson, Jeremy Worrell, you, Jim 14 Sweeting, Chris Bailey, copied to Latoya Smith 15 and John Burton. 16 A. Yes. 17 Q. He says "Jeremy has asked"; is that Jeremy 18 Worrell in the distribution list? 19 A. I presume so but I don't really have any memory 20 of him. 21 Q. Were there other, relevant Jeremys in the SSC 22 or -- 23 A. Certainly nobody -- that wasn't somebody in SSC. 24 Q. No. 25 "Jeremy has asked me to produce a paper on 26 1 Horizon Data Integrity to support [the Post 2 Office] in refuting claims by postmasters that 3 Horizon is causing money to be lost. 4 "I've put together an initial draft. I've 5 ignored the 'front bits' for now and am 6 currently looking for comments on the technical 7 aspects and in particular the comments in 8 yellow." 9 Then there's a character string indicating 10 that there was an attachment, I think. 11 A. Yes. 12 Q. "Also if anyone is aware of other material to 13 feed in on this, I would be grateful. I had 14 a quick look through PVCS ... " 15 Can you recall what PVCS was? 16 A. That was the document storage system. 17 Q. "... and a search through the TED ..." 18 Can you recall what the TED was? 19 A. No. 20 Q. "... and found nothing useful there. 21 "Note that this is NOT yet in a state to go 22 to [the Post Office] and once I've had feedback 23 from all of you (and anyone else you think is 24 relevant), Jeremy needs to pass this through 25 Commercial." 27 1 What was Commercial? 2 A. I presume the -- a department in Fujitsu. 3 Q. Reading on: 4 "We have a [conference] call with [the Post 5 Office] at 4 pm on Friday, so I'd appreciate any 6 feedback by lunchtime Friday ..." 7 Then a comment addressed to Jeremy, just 8 scrolling down: 9 "Do you have any thoughts as to where this 10 should be lodged in Dimensions. I assume we 11 need to make this a formal document, but if not, 12 I'm happy to remove the 'front bits'." 13 Then it's signed off "Regards, Gareth". 14 Just going back to the distribution list, 15 peace. Can you help us with the other people 16 there: Allan Hodgkinson, do you recall him and 17 what position he had? 18 A. He was one of the senior designers. I can't 19 remember. 20 Q. In the development team? 21 A. Development or architects, yes. 22 Q. Okay. You've said that you don't recall Jeremy 23 Worrell? 24 A. No. 25 Q. Jim Sweeting? 28 1 A. I don't remember that name. 2 Q. Chris Bailey? 3 A. He was another of the very senior 4 architects/designers. 5 Q. Latoya Smith? 6 A. I don't remember. 7 Q. John Burton? 8 A. I think he was the manager of a group of 9 development teams. 10 Q. So he's giving you a paper, Mr Jenkins, and is 11 asking you, on his initial draft, for views. 12 You'll see in the first paragraph he says: 13 "Jeremy has asked me to produce a paper on 14 Horizon Data Integrity to support [the Post 15 Office] in refuting claims by postmasters that 16 Horizon is causing money to be lost." 17 A. Yes. 18 Q. Would you have understood this literally: the 19 purpose of the paper was to be to support the 20 Post Office in refuting claims, rather than to 21 explore whether there may be anything in the 22 claims being made by subpostmasters? 23 A. Gareth wrote that sentence and not me. 24 Q. Yes, but you're a recipient of it. Would you 25 understand that the direction was to produce 29 1 a paper that would support the Post Office in 2 refuting claims rather than the more open 3 question of an exploration of whether there was 4 anything in the claims made by the 5 subpostmasters? 6 A. Yes, well, that's what he says he is doing. 7 Q. Yes, and how would you -- say for example you 8 thought "Well, I have got some knowledge about 9 the way that Horizon is structured and, by this 10 time, 2009, by the way in which bugs, errors or 11 defects have manifested themselves and how we in 12 SSC and the company more broadly has treated 13 them, I've got some evidence that may assist 14 claims by subpostmasters" -- 15 A. I -- 16 Q. -- would you have -- 17 A. I didn't -- 18 Q. -- included that in reply? 19 A. I was not aware of any bugs, errors and defects 20 that were causing money to be lost without them 21 leaving any sign that a problem had occurred. 22 In general, although, yes, of course there were 23 bugs, errors and defects, they were not causing 24 continual ongoing losses. 25 Q. You have introduced a qualifier there 30 1 Mrs Chambers, "I was not aware of bugs, errors 2 and defects that did not leave a sign that they 3 were occurring", essentially? 4 A. Mm, yeah. There obviously were bugs, errors and 5 defects that, in some cases, were causing money 6 to be lost but my view at that time was that 7 Horizon was robust in general. There would have 8 been specific cases when it was not. 9 Q. Would you -- 10 A. The -- 11 Q. I'm sorry? 12 A. No, it's all right. 13 Q. Do you agree that this is suggestive of 14 a request to put forward the best case in 15 refuting the claims made by subpostmasters, 16 making the best case for Horizon's integrity? 17 A. That would appear to be how Gareth has put it 18 there. I mean, my view was that, you know, 19 I would investigate each case individually, 20 which was my job, you know, when the support 21 calls came in, and that -- yeah. I mean, that's 22 how Gareth has put it there and I was not aware 23 of problems with data integrity that were 24 causing losses left right and centre, leaving no 25 indication behind them. 31 1 Q. I don't believe that we have the family 2 documents for this email, so I can't presently 3 show you the draft of the Horizon Data Integrity 4 report that's referred to there in that 5 character string, ie the document that you were 6 being asked to comment on. 7 A. Yeah. 8 Q. We do have a version of the document dated 9 2 October 2009, the day after your email and, 10 when we look at it, we'll see that it appears to 11 include or reflect the comments that you made in 12 part. 13 A. Yeah. 14 Q. Can we scroll up to your reply, please: 15 "Section 2 16 "You mention incremental consequence numbers 17 in the audit section at the end, but could 18 mention it earlier too, to make clear that each 19 message has a unique identifier which stays with 20 it when it is replicated. 21 "Each individual message has a checksum. 22 "What quite sure what you should be saying 23 about CRC read failures. We aren't currently 24 checking old event logs for these when doing 25 audit retrievals, and I don't want us landed 32 1 with even more checks to make. If there are CRC 2 errors, SMC normally raise a call and we trash 3 the message store and let it rebuild -- but 4 probably don't want to say that! But if we 5 don't say, [I think that's meant to read 'will'] 6 they ask?" 7 Is that right? 8 A. Yes. 9 Q. "[will] they ask? 10 "Event logs -- more than 18 months but not 11 sure if it is 7 years." 12 Then: 13 "[Paragraph] 3.1.1, 3.1.2 The user will get 14 an AP message warning them that the last session 15 ended in error, but it only tells them to check 16 AP transactions, not others. 17 "3.1.3.1 If banking recover is not completed 18 immediately after the counter is replaced, this 19 is reported on the banking reconciliation 20 reports and followed up." 21 So the part of the reply where you say, "We 22 aren't currently checking old event logs for 23 these when doing audit retrievals, and I don't 24 want us landed with even more checks", can you 25 just explain what you're referring to when you 33 1 say, in this context, "CRC read failures"? 2 A. This is when a message has been -- a Riposte 3 message, when read by a process, either on the 4 counter that it was originally written on or one 5 of the other counters, every time it was read, 6 the checksum on the message was recalculated and 7 if it didn't match it implied there'd been some 8 sort of a corruption, and that would raise 9 a critical red event. 10 Q. You say: 11 "We aren't currently checking old event logs 12 for these when doing audit retrievals ..." 13 A. As part of the audit retrieval process after 14 2008 sometime, the Security team would also 15 extract the Tivoli events for the branch over 16 the relevant period, and SSC staff would look at 17 those events to see if there was anything of 18 concern. In particular, we were looking for the 19 Riposte lock events which might indicate some 20 silent failure that might not have been noticed 21 at the time. 22 This was a not exactly a time consuming 23 process but it was part of the process. Now 24 we -- I can't remember now but this implies that 25 we wouldn't have seen those CRC read failure 34 1 events in that process. They should, however, 2 have been noticed at the time because they were 3 one of the events that the SMC monitoring team 4 were monitoring for, so there should have been 5 a PowerHelp call raised at the time, if these 6 events were occurring on a particular counter. 7 That PowerHelp call would then have been in 8 the audit trail, so this isn't something that 9 would have been happening and then wouldn't 10 notice as part of the -- and would not be seen 11 as part of the audit retrieval. 12 Q. You continue that: 13 "[The] SMC normally raise a call and we 14 trash the message store and yet it rebuild ..." 15 What do you mean by that, trashing the 16 message store? 17 A. I think that's the bit that I probably didn't 18 want to say, "trash the message store", because 19 it sounds horrible. What it meant was delete -- 20 on the affected counter, you would delete the 21 message store at a time when the counter -- 22 I mean often -- one thing I can't remember is 23 whether these sort of errors actually shut down 24 Riposte on the counter, which meant that you 25 then couldn't use it or, if it could still be 35 1 used. Obviously, if it was unusable, then it's 2 not going to -- this underlying error couldn't 3 possibly cause any financial impact because you 4 can't use the counter. But I cannot remember if 5 that was the case or not. 6 But then trashing the message store is 7 actually deleting the message store file on the 8 affected counter, and then it would rebuild 9 itself, and I can't remember if it used the -- 10 and it would then be copied from one of the 11 other counters to get the complete set of 12 messages back in again, and that would then 13 hopefully not have any corruptions in it. 14 Obviously, if this kept happening because 15 you've got a dodgy disk on a counter you might 16 still get more errors being reported and then 17 the action would be to replace the counter, 18 which would, as part of that process, also 19 rebuild the message store. 20 Q. Why wouldn't you want to say this? 21 A. I think because "trash" is not a particularly 22 good-sounding word. 23 Q. But there was an acceptable way to describe what 24 you were doing, wasn't there -- 25 A. That was -- 36 1 Q. -- without using the word "trash"? 2 A. Yes, you would say, "delete the message store" 3 that -- 4 Q. Yes. Why wouldn't you want to say that? That 5 you were deleting the message store and then 6 letting it rebuild itself? 7 A. I really cannot remember exactly why I put it 8 into those words and, obviously, I wish I hadn't 9 put it into those words but I cannot remember my 10 thinking at the time. 11 Q. I'm not concerned at the moment with the word 12 you used, "trashing the message store", I'm 13 concerned with -- I'm asking you why wouldn't 14 you want to reveal this in a report about 15 Horizon data integrity? 16 A. Yes, again, that's why I can't remember, really, 17 why I said that. It wasn't because I thought it 18 was an absolutely awful of thing to be doing. 19 Q. What other reasons could there be for not 20 revealing it? 21 A. I don't know. 22 Q. Were you saying "What they don't know won't harm 23 them"? 24 A. It certainly looks as if that was what I was 25 saying but, if that is the case, I don't know, 37 1 really, why I was saying that, because this was 2 not doing -- this was, you know, to get the 3 counter up and running again with a non- 4 corrupted set of data. 5 Q. Or were you saying, because it's not going to 6 help the Post Office prove Horizon's integrity, 7 let's not tell them? 8 A. No, I don't think I was saying that. 9 Q. Again, can you, reflecting back, think what the 10 reason would be not to reveal this information 11 in a report on Horizon data integrity? 12 A. The only thing I can think of with hindsight is 13 perhaps when these errors should -- had 14 occurred, we should have specifically been 15 looking to see if it could have had any impact 16 on the -- anything, you know, if they were 17 balancing at the time, and that wasn't a check 18 that we were making at the time. 19 But did I think this was a situation that 20 was causing -- was a cause of ongoing problems, 21 you know, no, I don't -- even now, I don't think 22 that was the case. But no, I don't know why 23 I put that sentence in those terms. I think 24 I just wanted to let Gareth know of what the 25 situation was and then he could decide precisely 38 1 what he wanted to say about it. But I agree, 2 that doesn't look good. 3 Q. You say, "SMC normally raise a call". 4 A. Yes, that was the process and then that was so 5 we knew to do something about it. Obviously, if 6 we had a call direct from a branch about any 7 sort of problems and then we looked at that 8 call, we would have seen those CRC errors and 9 would have looked to see what needed to be done 10 about it. But I cannot be 100 per cent sure 11 that if they weren't -- if they were saying, you 12 know, that they'd got financial problems and we 13 saw the CRC errors, then, obviously, we would 14 look to see if they could have been a cause of 15 it. 16 But, if we're talking about looking at data 17 retrieved some years or months -- months or 18 years afterwards, again, if we were examining -- 19 if we were actually investigating, rather than 20 just doing the audit retrieval, we would look 21 for anything wrong. But, no, sorry, I'm just 22 not getting anywhere with this. 23 Q. Okay. You say the SMC was normally to raise 24 a call. Were there times, therefore, when the 25 SMC failed to pick up this issue or there were 39 1 problems that therefore slipped through the net? 2 A. I could not say that they 100 per cent picked up 3 on everything that they were always meant to, 4 but they would certainly have picked up on the 5 vast majority of things like CRC errors. 6 Because that was their job to do. 7 Q. Was there a process to ensure that they ensured 8 that that process always operated as it was 9 intended? 10 A. I don't know what their processes were and how 11 much cross-checking was done by their 12 management. 13 Q. Further down the email, when you're commenting 14 on what were, I think, paragraphs 3.1.1 and 15 3.1.2 of the draft Integrity Report, you say: 16 "The user will get an AP message warning 17 them that the last session ended in error, but 18 it only tells them to check AP transactions, not 19 others." 20 Can you explain what you mean by that, 21 please? 22 A. Not properly, without seeing the underlying 23 document but I think we're talking about what 24 has happened when a counter has failed in some 25 way and it's going into the recovery system, 40 1 when the counter becomes available again. 2 Q. You continue: 3 "If banking recovery is not completed 4 immediately after the counter is replaced, this 5 is reported on the banking reconciliation 6 reports and followed up." 7 Did you either work in or see BIMS? 8 A. The BIMS were produced by the -- what I think of 9 as MSU, Management Support Unit. SSC did not 10 see that system but the data that went into the 11 BIMS reports was often taken from PEAKs, which 12 the MSU raised asking SSC to check out various 13 reconciliation report entries and then we'd send 14 our response back and they would, if necessary, 15 send a BIMS report, often just cutting and 16 pasting our response, and that went to Post 17 Office. 18 Q. Your comment on 3.1.3.1, is that a description 19 of what the process was, ie what was supposed to 20 happen? 21 A. Yes. I mean, if anything went wrong with 22 banking transactions, there was a huge amount of 23 central reconciliation that was done matching up 24 the counter outcome -- the outcome was known at 25 the data centre -- and also data received from 41 1 the various financial institutions, the banks, 2 and so on. And everything was matched up in 3 there and any inconsistencies gave an entry on 4 the banking reconciliation reports. 5 Specifically, in a recovery situation, it's 6 possible that a banking transaction had been 7 authorised by the bank and the money removed 8 from the customer's account but, if it hadn't 9 settled on the counter before the counter 10 failed, then the money might not have -- that 11 transaction would not be included in the branch 12 accounts and so this banking reconciliation 13 process was all intended to get everything into 14 a consistent position, and recovery was part of 15 that. 16 Q. So failures in banking recovery were always 17 supposed to be reported accurately on banking 18 recovery reports? 19 A. No, if there was any consistency -- any 20 inconsistency between the outcomes and 21 a transaction needing recovery -- a banking 22 transaction needing recovery was going to be 23 incomplete until somebody logged back on to that 24 counter again. 25 Q. So that was supposed to be reported on a banking 42 1 reconciliation report? 2 A. It would all be reported somewhere if there was 3 any inconsistencies. 4 Q. It was supposed to be followed up with a BIMS 5 notice; is that right? 6 A. It depends what the outcome was. When the 7 recovery was completed, if the branch could 8 confirm whether they'd paid the money out or 9 not, then that would resolve the, you know, the 10 inconsistency, if you like, and then that will 11 get fed through and matched up in the 12 reconciliation system. 13 Q. During your evidence back in May, you were asked 14 to consider a circumstance in which a system 15 error was to result in a BIMS notice, and you 16 gave evidence that it wasn't your role to follow 17 up what happened with the BIMS notice and later 18 actions; do you remember? 19 A. Yes. 20 Q. So where it says here "If banking recovery", 21 et cetera, it is "followed up", again, are you 22 describing the process rather than your 23 knowledge of what actually happened? 24 A. This is describing -- yeah, by following it up, 25 I mean that, if necessary, SSC would investigate 43 1 what had happened, in some cases would phone the 2 branch to ask them to complete the recovery by 3 logging on to the counter. 4 Q. So, essentially, by this email you're adding in 5 some further points that may assist Mr Jenkins 6 to assist the Post Office in refuting claims 7 that Horizon is causing money to be lost; is 8 that right? 9 A. I was just trying to clarify where he had put 10 things in his draft that I felt could be perhaps 11 explained. I can't remember, without seeing the 12 draft, whether this is in more detail or saying 13 that, you know -- you know, I knew a little bit 14 more about the process at this level. 15 Q. Can we turn, please, to a draft of the report. 16 FUJ00080526. You'll see that under "Document 17 Status" this is marked as a final draft. You'll 18 see in the bottom of the document on page 1 it's 19 version 1.0 and it's dated 2 October 2009. The 20 email exchange we've just looked at was 21 1 October 2009. 22 A. Could I just ask, is it possible to have this 23 screen raised up slightly? 24 Q. Yes. Do you want to try it yourself? If you 25 grab the top of it. Maybe one of the ushers can 44 1 assist. 2 A. Yes, that's right better. Thank you. 3 Q. We were looking at the date at the bottom, 4 2 October 2009, the day after your email. 5 A. Yeah. 6 Q. Then if we can go to page 3, please. We can see 7 the version number. 0.1b is said to be dated 8 2 October 2010, I think that must be a typo for 9 2009. Then: 10 "First informal draft. Changes from version 11 0.1a were marked in red (like this) with 12 strikeout for significant deletions." 13 Then this version, 1.0 "Version for release 14 to Post Office". 15 Yes? 16 A. Yes. 17 Q. It's those, 0.1a and 0.1b, that I don't think 18 that we have got. If you just look down at the 19 reviewers, you don't seem to be listed there -- 20 A. No. 21 Q. -- and yet you did review? 22 A. Yes, I don't know how informal that first draft 23 was but, yes, I did review it. 24 Q. Would that sometimes be the case, that if you 25 were an informal reviewer, you wouldn't be 45 1 listed? 2 A. I don't know, I wasn't -- no, I don't know. 3 Q. Or is this a description of those who are to 4 review this draft, rather than those which had 5 reviewed previous drafts? 6 A. Yes, that could well be. I mean, we are now up 7 at a far higher level, if you like, of people 8 than the level at which I worked. And I note 9 that Chris Bailey and Alan Hodgkinson aren't on 10 there either. So I assume, you know -- maybe, 11 you know, Gareth had just asked around useful 12 people to try to get a picture in his first very 13 informal draft. 14 Q. Can we go to page 5, please. We can see the 15 purpose of the document set out: 16 "This document is submitted to Post Office 17 for information purposes only and without 18 prejudice." 19 You can see at the top it says in 20 "Commercial in Confidence and Without 21 Prejudice". It says: 22 "In the event that Post Office requires 23 information in support of a legal case Fujitsu 24 will issue a formal statement. 25 "This document is a technical description of 46 1 the measures that are built into Horizon to 2 ensure data integrity, including a description 3 of several failure scenarios, and descriptions 4 as to how those measures apply in each case." 5 So that's the purpose of the document. Then 6 if we go to page 6, please. There is the 7 section on Horizon data integrity. 8 If you scroll down, please, if you just see 9 the paragraph that's now at the top of the page: 10 "Every record that is written to the 11 transaction log has a unique incrementing 12 sequence number. This means it is possible to 13 detect if any [transactions] have been lost." 14 Then right at the foot of the page: 15 "Each record generated by a counter has 16 an incremental consequence number and a check is 17 made that there are no gaps in the sequencing." 18 So that is, I think, the first comment you 19 made, that, whereas the sequence number had been 20 referred to by Mr Jenkins in the audit section, 21 which is what this part of the report is, it 22 could usefully be explained earlier. Can you 23 see that at the top? 24 A. Yes, because -- yes. 25 Q. Then the second point that you made in your 47 1 email "If there are CRC errors, SMC normally 2 raise a call and we trash the message store and 3 let it rebuild but we probably don't want to say 4 that, but if we don't say, will they ask", 5 that's not referred to here or, indeed, so far 6 as I can tell, anywhere in the report. Do you 7 think, because that was a potentially tricky 8 point, it was left out? 9 A. I don't know. I mean, what we were doing was 10 fixing the problem and not in a particularly 11 tricky way. It was the clean way to fix the 12 problem. I don't know what Gareth had said 13 about it before. I think what he's put there is 14 correct but what he doesn't do is go into any 15 detail as to the action that was taken when it 16 did happen. 17 Q. Then if we go forwards, please, to 3.1.1 and 18 3.1.2 on the next page and just scroll down, 19 please. Do you remember these were the 20 paragraphs, 3.1.1 and 3.1.2 -- 21 A. Yes. 22 Q. -- where you had said "The user will get an AP 23 message warning them that the last message ended 24 in error but it only tells them to check AP 25 transactions, not others"? 48 1 A. Yes. 2 Q. I think the point you were making, is this 3 right, is that the subpostmaster was not getting 4 information telling them to check all 5 transactions, only the AP transactions, yes? 6 A. Yes. I mean, AP transactions were different to 7 other transactions in that they were 8 recoverable, whereas other transactions were 9 not. 10 Q. So why were you raising that point, to make the 11 distinction that the subpostmaster wasn't 12 getting a message to check other transactions? 13 A. It must have been something that was in the 14 previous draft where I felt there was -- 15 a little bit of clarity was needed but I can't 16 remember. 17 Q. It doesn't look as if your point has been 18 addressed here, does it, in 3.1.1 and 3.1.2? 19 A. Unless there had been a sentence in saying that 20 the user will be prompted. 21 Q. Which has now gone? 22 A. Which has now gone. That is possible. 23 Q. Thank you very much, that's all I ask on the 24 second topic. 25 Sir, I wonder whether that would be 49 1 an appropriate moment with the morning break? 2 Sir, you're still on mute. 3 SIR WYN WILLIAMS: It would but I may have 4 misunderstood something. I think it's page 5 of 5 that document, but there is certainly a page 6 that we've looked at, which describes the 7 version of Horizon to which it relates. Have 8 I misunderstood that? 9 MR BEER: Yes, that is page 5. 10 SIR WYN WILLIAMS: And -- 11 MR BEER: It's the last sentence. 12 SIR WYN WILLIAMS: This document only covers 13 Horizon. It does not cover HNG-X, Horizon 14 Online. Am I being too simplistic? Does that 15 mean that it only covers the version of Horizon 16 which existed until Horizon went online? 17 A. Yes, this covers the -- what we sometimes call 18 Legacy Horizon, which was based all around the 19 Riposte system because the data integrity side 20 of things, it was all to do with the way that 21 Riposte was working behind the scenes. Horizon 22 Online was based on a totally different set of 23 centrally-held databases. 24 SIR WYN WILLIAMS: Sure, and at the date of this 25 document, Legacy Horizon still existed? 50 1 A. Yes, but -- 2 SIR WYN WILLIAMS: Had there been -- yes, the 3 transition to Horizon Online hadn't yet 4 occurred? 5 A. No. 6 SIR WYN WILLIAMS: Fine. Sorry. I just wanted to 7 be clear in my mind about that. Thank you. 8 Yeah. Yes, let's have our morning break. 9 MR BEER: Sir, 11.30, then please. 10 SIR WYN WILLIAMS: Yes, fine. 11 MR BEER: Thank you. 12 (11.15 am) 13 (A short break) 14 (11.30 am) 15 MR BEER: Good morning, sir, can you see and hear 16 me? 17 SIR WYN WILLIAMS: Yes, I can, yes. 18 MR BEER: Thank you. 19 Mrs Chambers, can we turn to the third 20 topic, please, which is informal examination of 21 issues within the SSC and the process for the 22 examination of issues within the SSC that may 23 end up in court proceedings. 24 Can we start, please, by looking at an email 25 chain. FUJ00156153. I should make clear that 51 1 this isn't an email chain that you were included 2 on but I'm going to ask you questions about 3 whether what's in it reflected the position at 4 the time. 5 Can we go to page 3, please, and look at the 6 foot of the page, which is the first in the 7 chain. Foot of the page, please. Thank you. 8 So 2 June 2010, from Penny Thomas to Steve 9 Parker: 10 "Hi Steve 11 "You wanted to change the way we request 12 these checks to PEAK ..." 13 The subject is "Event Analysis via PEAK": 14 "... and I think we need to agree the 15 format; have you got a few minutes to agree 16 process?" 17 At this time, we can see who -- and we know 18 who -- Mrs Thomas was, a security analyst. What 19 position did Mr Parker hold in June 2010? Was 20 he the manager of the SSC by then? 21 A. I think so. He took over certainly around about 22 then, so I assume that, yes, he was the manager. 23 Q. So your line manager? 24 A. Yeah. 25 Q. Then if we scroll up, please. We can see his 52 1 reply: 2 "Penny, 3 "Yes, I'd like to change it as well so that 4 we get formal PEAKs raised for ARQs as discussed 5 last week, establish audit trail, spread the 6 work, etc." 7 What do you understand that first part to 8 mean, "We get formal PEAKs raised for ARQs"? 9 A. I presume it means, so for each -- for each ARQ 10 extract that the Security team were doing, 11 where, by this point, we were also checking any 12 events that were raised in the same period, that 13 the request for SSC to make those checks, each 14 one would have a PEAK raised for it. 15 Q. So for each request for an ARQ, there was 16 an equivalent PEAK? 17 A. Yes, effectively they'd do the ARQ extract of 18 the transaction data and Riposte events. They'd 19 also get the Tivoli events and then raise the 20 PEAK, with the Tivoli events attached, and route 21 it to SSC for us to examine. 22 Q. Would you understand why it would be necessary 23 or desirable to establish an audit trail? 24 A. I presume just so that somebody could say, if 25 asked, "Yes, those events have been checked by 53 1 somebody in the SSC". 2 Q. Would that be in the context of court 3 proceedings or in other contexts too? 4 A. I don't know. 5 Q. Okay: 6 "Before I can do this we may need to talk to 7 Tom as well." 8 I think that must be a reference to Tom 9 Lillywhite. Can you see him amongst the copy 10 list? 11 A. I can see him, yes. 12 Q. Do you remember who he was? 13 A. No. 14 Q. "I'm concerned that if we put this on a formal 15 level like this it may mean that further down 16 the line random members of the SSC get 17 a subpoena and we have to testify. If there is 18 any chance of this happening then we (SSC) will 19 not be giving guidance on the events. We need 20 guidance from Tom (or Fujitsu legal) on how we 21 protect ourselves from the possibility of court 22 appearance before we formalise the process?" 23 Can you recall a concern within the SSC at 24 about this time, that's mid-2010, that the SSC 25 should not be giving guidance or speaking to ARQ 54 1 data because it may result in the SSC giving 2 evidence in legal proceedings. 3 A. I don't remember anything specifically from 2010 4 but I think, after I gave evidence in 2006, the 5 SSC management were not keen on any SSC members 6 having to be put in that position again. 7 Q. Why were they keen that members of the SSC 8 should not give evidence? 9 A. Because they felt it was not our job to give 10 evidence. 11 Q. Do you know why they felt it was not your job to 12 give evidence? 13 A. Because we'd had no training, it was not part of 14 our job description. You know, we did not take 15 on the job thinking that we might find ourselves 16 in court. 17 Q. Was that a theme, that was the subject of 18 discussion, that there was a need for the SSC to 19 protect itself from the possibility of giving 20 evidence in court? 21 A. Yes, I think that was how our management felt. 22 Q. What about individual members of the SSC too? 23 A. Yes, I don't think -- I don't think any of us 24 had ever joined a support team thinking that 25 that is where we might end up. 55 1 Q. So, although there was a desire to formalise 2 matters, is this right: you would understand 3 what Mr Parker is saying here, that that 4 formality can't come at the expense of exposing 5 ourselves to court appearances? 6 A. That seems to be what he's saying there, yes. 7 I'm not sure that I was aware of this discussion 8 at the time, though. 9 Q. That was my next question: did you understand 10 the relationship between a formal process for 11 administering potential litigation cases and 12 an increased possibility of a court appearance 13 by SSC staff? 14 A. I'm not sure that that was anything I was 15 thinking about at the time when I was checking 16 these events. But, obviously, it was a -- 17 appears to have been of concern to Steve. 18 Q. Ie an instruction, keep in informal, do things 19 by email and discussion, otherwise you may 20 expose yourself to giving evidence in court? 21 Did that come down to the workers in the SSC? 22 A. I don't recall anybody ever saying that. 23 Q. Okay, let's read on, please. Scroll up. We can 24 see Penny Thomas's reply: 25 "OK, Steve, I'll continue requesting via 56 1 email until you are fully satisfied that [the] 2 SSC are protected." 3 Yes? 4 A. Mm. 5 Q. Scroll up still further, please. We can see 6 Mr Parker refers it on to Mr Lillywhite: 7 "Tom, 8 "Any comment on this please? 9 "It is important for the ARQ process that 10 SSC examine the events generated and then 11 comment on their potential impact on the 12 financial status of the branch." 13 Would you agree with that? 14 A. Yes. 15 Q. "This has been done in the past on an informal 16 basis (email to Anne Chambers normally!) ..." 17 Is that right, that it was normally you that 18 received these informal requests via email? 19 A. I don't actually remember. I'd forgotten that 20 this was done on an informal basis at any point. 21 But that's what it says. So ... 22 Q. "... but that informal process leads to requests 23 being lost when somebody may be on leave, etc. 24 "We need to formalise this but I'm concerned 25 about the legal implications. SSC staff are not 57 1 trained on evidential requirements or as 2 witnesses in court." 3 That's something that you've just mentioned? 4 A. Yeah. 5 Q. Just stopping there, I'm going to come back to 6 ask you some questions about that in the context 7 of the Lee Castleton case. In general terms, 8 were you ever given any guidance on when you 9 were carrying out enquiries and carrying out 10 investigations in a case that may end up in 11 litigation, that you had to do or not do certain 12 things as a potential witness, like retaining 13 your working notes? 14 A. I think you're -- when we were just looking at 15 ordinary support calls, I don't think it 16 occurred to any of us at the point at which we 17 were doing that investigation that it could, at 18 some point in the future, result in us needing 19 to be a witness. 20 Q. Just stopping there, I'm not asking about those 21 ordinary support calls, as you describe them. 22 I'm talking about these informal requests from 23 Security? 24 A. Yes, I don't think anybody, when we started 25 checking these Tivoli events, I don't think 58 1 there was any discussion, when that process 2 started, that we might then be expected to 3 appear in court about it. 4 Q. No training or guidance that "Look, if you do 5 end up in court, the court has certain 6 requirements for somebody that's performing 7 a task of specialist expertise" -- 8 A. No. 9 Q. "-- such as retention of working notes, 10 retention of copy documents"? 11 A. No. 12 Q. The duties that you might owe to a court? 13 A. No, there was no discussion or training. There 14 was, I believe, a KEL that sort of outlined the 15 sort of things that -- the events that we needed 16 to check, but it was purely technical. 17 Q. Mr Parker continues: 18 "If there is any possibility of a court 19 appearance or a witness statement being required 20 then we have to refuse to process the ARQ 21 requests. 22 "Do your own what the legal situation is 23 here?" 24 Do you recall that, that it got to the stage 25 that, such was the concern about the possibility 59 1 of giving a witness statement or making a court 2 appearance, that the SSC would refuse to look at 3 the ARQ data? 4 A. No, I think I was completely unaware of any of 5 this. 6 Q. This refers to emails being sent to you on 7 an informal basis asking you to do this work. 8 Did that continue or can't you remember? 9 A. As I said, I don't remember ever doing it by 10 email. We did switch to PEAKs being raised and 11 then they were shared out, so it wasn't always 12 me doing them. 13 Q. Scrolling up the page, please. Thank you. We 14 can see Mr Lillywhite's reply. 15 "If there is indeed legal implications, and 16 you are all agreed on that then I think we seek 17 advice from our legal department ... it is too 18 important to get wrong!" 19 Then up to page 1, please, and then scroll 20 up, please, to look at Mrs Thomas' email. We 21 can see her reply: 22 "I am not convinced there is, Tom." 23 You remember he said if there are legal 24 implications on the SSC staff undertaking 25 investigation of these events and looking at ARQ 60 1 data. She says: 2 "I am not convinced there is, Tom. 3 "While Anne has been helping us she has been 4 fully shielded from any form of [Post Office] 5 litigation. Why would we specifically identify 6 the checking of events as more vulnerable than 7 any other part of the process considering the 8 total end-to-end process employed here? The 9 names of those checking events for us are not 10 notified to the [Post Office] and we have the 11 ability to identify and select any expert 12 witness we consider appropriate to support [the 13 Post Office's] prosecutions. No one in the 14 company can be forced to sign a witness 15 statement if they do not want to; and [the Post 16 Office] cannot cherrypick our staff. 17 "Gareth has the responsibility of covering 18 transaction records for all litigation facing 19 activity until now and there has been no issue. 20 Do we need to life a suitable 'expert' to cover 21 event filtration and analysis? That's another 22 question." 23 Was that the subject of discussion within 24 the SSC, that you had been shielded from 25 involvement in Post Office litigation? 61 1 A. I don't recall it ever being discussed. 2 Obviously, I knew that, after the Marine Drive 3 case that I'd been involved in, Gareth had taken 4 on subsequent witness statements and trials but 5 I don't recall any discussion about it. 6 Q. What about the point that the names of those 7 checking events for us are not notified to the 8 Post Office? Was that an instruction that was 9 given, "Don't tell the Post Office who carried 10 out the work because, otherwise, they might end 11 up a witness in court"? 12 A. No, I don't recall that ever being said or 13 discussed. 14 Q. Was the facility available to obscure from the 15 Post Office who was carrying out the work, 16 checking the events analysis? 17 A. I don't know what got passed to the Post Office 18 as a result of the events analysis. I mean, 19 that all just went straight back to the Security 20 team and I don't know precisely what Penny then 21 passed on as part of the ARQ data. 22 Q. Do you recall whether you or others in the team 23 were told, "Don't worry, even though you're 24 responding to these informal requests for 25 analysis, we shield you by not revealing who 62 1 undertook the work when we passed the product 2 back to the Post Office"? 3 A. I don't recall any of us ever explicitly 4 worrying about that or thinking that it was 5 something to be concerned about. 6 Q. As a means of ensuring that what happened in the 7 Marine Drive case didn't happen again, that the 8 SSC was dragged into giving evidence? 9 A. I don't think it was something that we were 10 particularly thinking about. The checking of 11 the events was a task that needed to be done and 12 we did it and passed it back. I don't think we 13 considered possible consequences, although, in 14 the light of Marine Drive, maybe it's something 15 we should have been more alert to but, no, it 16 wasn't something we were talking about or 17 considering. 18 Q. After Marine Drive, I think it's right that 19 no-one from the SSC did give evidence again? 20 A. Yeah. 21 Q. Do you know how that came about? 22 A. I presume that was a decision by the management, 23 who, I think for various reasons, weren't happy 24 that I'd had to give evident in the first place. 25 Q. But you in the SSC carried on doing the 63 1 analysis -- 2 A. We did -- 3 Q. -- and passing it back to Security? 4 A. I became involved in Marine Drive because 5 I dealt with the original support call from the 6 branch, which did not go back to Security 7 because it went back to Marine Drive. Where 8 I -- I mean, after that, I think Gareth took 9 responsibility for making checks on other 10 branches for the transaction data for 11 litigation, where it was coming in through the 12 Security team, and I, occasionally, a few 13 occasions, possibly helped him with some of that 14 analysis but -- 15 Q. Were you doing so in the knowledge that your 16 identity would not be revealed to the Post 17 Office so there was no prospect of you having to 18 give evidence? 19 A. I think I always assumed that it would be Gareth 20 giving evidence. I don't think I considered 21 whether my identity or my involvement was being 22 hidden to protect me or anything. By that point 23 it just seemed to be the process that Gareth, as 24 it says here, had taken that responsibility. 25 Q. Do you know how that process came about? That 64 1 Gareth had taken responsibility? 2 A. No. 3 Q. Can we turn to the fourth issue then, please, 4 that can come down. 5 Back when you gave your evidence on 3 May, 6 you described an occasion when the Post Office 7 had wanted to insert transactions without the 8 branch being aware. I'll just read back the 9 questions and answers. The question was: 10 "Was there any method to alert others that 11 corrective action had been taken to insert data 12 or extra messages into a branch's account?" 13 You said: 14 "The ARQ data would contain both the 15 original transaction and the corrective 16 transaction, at the point at which they were 17 done. If the full, unfiltered data was 18 retrieved and inspected, then that would show 19 the comment, for example. Certainly, sometimes, 20 for counter corrections -- and it wasn't done 21 consistently -- but we often might use a counter 22 number that didn't exist to make it clear that 23 it was something out of the ordinary or 24 a username, including SSC, again to show that it 25 was something out of the ordinary. 65 1 "That wasn't done on this specific one and 2 I cannot remember whether that was because I was 3 specifically asked not to or I was just 4 producing a transaction that was absolutely 5 a mirror of the one that shouldn't have been 6 there in the first place and all I did was 7 change the signs of the values, effectively and 8 all I -- I left all the other data in there as 9 it was." 10 So remembering back what you were saying and 11 just to synthesize it, you were saying, 12 sometimes, a fictitious counter number was used 13 to mark out the transaction correction? 14 A. As I recall, yes. 15 Q. Secondly, you were saying but that wasn't done 16 consistently, ie the use of a fictitious counter 17 number to mark out the fact that SSC had made 18 a correction? 19 A. Yeah. 20 Q. Thirdly, you were doing that, or the SSC was 21 doing that, deliberately, ie using the 22 fictitious counter number, because you would 23 want to show that an SSC member had been making 24 the correction? 25 A. Yes. 66 1 Q. But, fourthly, there might be occasions when you 2 were specifically asked not to use the 3 fictitious one? 4 A. I don't think we were ever asked not to use 5 a fictitious one. 6 Q. It was just a passage of your evidence where you 7 said "That wasn't done on this occasion and 8 I can't remember whether that was because I was 9 specifically asked not to", which tended to 10 indicate that you were saying that you may have 11 been asked not to use the fictitious number? 12 A. I don't recall ever being asked not to do that 13 and I can't remember which specific instance 14 we're talking about here. Sorry. 15 Q. Overall, does it mean that it was possible for 16 members of the SSC to insert transactions using 17 the branch user ID? 18 A. Right, you're talking about user ID here now, 19 rather than counter number, but, yes, it was -- 20 I mean, the messages that we inserted could have 21 contained the branch user ID. 22 Q. Would it follow that standard filtered ARQ data 23 would not distinguish those insertions from 24 those that were, in fact, carried out in the 25 branch? 67 1 A. The standard ARQ data, yes. You might not be 2 able to see the difference. 3 Q. So transactions which appeared in the standard 4 filtered ARQ data, for example, in 5 Mr Castleton's case, with his ID user number 6 next to them, would not necessarily mean that 7 they were carried out by him? 8 A. It would have been possible, yes, for SSC to put 9 transactions in, that -- 10 Q. Using his ID? 11 A. Using his ID. 12 Q. Without leaving a fingerprint on the standard, 13 filtered ARQ data that that had been done? 14 A. Yes, I think that would have been possible. 15 Q. It shouldn't have been done but it could be 16 done? 17 A. Yes, I don't think there's anything that would 18 have prevented that. I don't believe that was 19 done but I can't say it's an impossibility. 20 Q. Thank you. 21 Can I turn then to Mr Castleton's case and 22 begin with what you did and didn't know, what 23 information you did have and what information 24 you didn't have, and the stage at which you 25 became involved. I'm going to try and do it 68 1 chronologically, including by establishing what 2 had happened before your first involvement, 3 which was, I think, on 26 February 2004, yes? 4 A. Yes. 5 Q. I'm afraid there's quite a long run-up to the 6 wicket here but I just want to see what had 7 happened before you became involved. 8 A. Certainly. 9 Q. Can we start, please, with the first call on 10 14 January 2004. You address this in your 11 witness statement, please. So that is 12 WITN00170200. 13 I'm told, sir, that's not on the system, 14 which is remarkable because it's Mrs Chambers' 15 38-page second witness statement. Maybe if 16 I just hand my copy of it to the operators to 17 see whether it's an error in my reading. 18 Sir, we probably need to take a break, then, 19 if that is indeed not on the system. 20 I apologise for this. Can we leave it that 21 we'll come back to you when it is on the system, 22 please. 23 SIR WYN WILLIAMS: Okay. That's fine. Do you want 24 me to sit in my chair, so to speak, for a few 25 minutes or can I wander around safely for at 69 1 least five minutes? 2 MR BEER: The latter, please, sir, and we'll get you 3 a message one way or another when we're to 4 reconvene. Thank you very much, sir. 5 (12.02 pm) 6 (A short break) 7 (12.19 pm) 8 MR BEER: Good afternoon, can you now see and hear 9 me? 10 SIR WYN WILLIAMS: Yes, I can, thank you. 11 MR BEER: Thank you. Apologies for that protracted 12 delay. 13 Can we look, please, at WITN00170200, and 14 can we go, please, to page 3 of your witness 15 statement and look at paragraph 10, please. 16 I should have taken you to the heading at the 17 top of the page "Involvement in relation to 18 Marine Drive Post Office and the litigation 19 against Lee Castleton". 20 So this is the section of your statement, 21 indeed the rest of the statement, which deals 22 with Mr Castleton's case. At paragraph 10, you 23 say: 24 "I have now been provided with a copy of the 25 NBSC call log for Marine Drive. This is not 70 1 something that was ever available to me during 2 my time at Fujitsu. As I have explained in my 3 first statement, there was a clear division 4 between business investigations conducted by 5 NBSC and system investigations conduct by SSC. 6 I can see now from this document that there was 7 a call about a discrepancy at Marine Drive on 8 14 January 2004 and another a week later on 9 21 January 2004, and the record shows NBSC as 10 assisting the postmaster to make checks at that 11 time." 12 So where you say, "this is the first time 13 that I've seen the NBSC call log", do you mean 14 you've been provided a copy of the document by 15 the Inquiry -- 16 A. Yes. 17 Q. -- and this is the first time in the 19 years 18 since the relevant events that you have seen the 19 call log? 20 A. I didn't see the call log at the time and I have 21 not seen it -- I had never seen it until it was 22 provided as part of the evidence set with the 23 Rule 9 request. 24 Q. So you say: 25 "This is not something that was ever 71 1 available to me during my time at Fujitsu." 2 To be clear, it wasn't available to you when 3 you investigated the issue at Marine Drive back 4 on 26 February 2004 -- 5 A. It was not available to me then. 6 Q. -- and it was not available to you when you were 7 asked to look at a wider range of issues in 2006 8 in preparation for the court case concerning 9 Mr Castleton? 10 A. No, it wasn't available to me then and I don't 11 recall being asked to look at wider issues at 12 that time either. 13 Q. We know that, for example, you were told, before 14 you gave evidence at the High Court, that you 15 may be asked questions about the Callendar 16 Square/Falkirk bug? 17 A. Yes, I was told that that was going to come up 18 and I should be ready to answer questions about 19 that. 20 Q. As part of preparation, you weren't shown a copy 21 of the NBSC call log? 22 A. No. 23 Q. You tell us in this statement that there was 24 a clear division between business investigations 25 conducted by the NBSC and systems investigations 72 1 conducted by the SSC. Is that the reason why 2 you in the SSC didn't see the NBSC call log? 3 A. I presume so. NBSC was a Post Office 4 organisation, totally separate from Fujitsu and 5 I don't think at the Helpdesk levels they -- 6 either side had view of the others' calls. 7 Q. The information we're going to look at in 8 a moment, concerning events on 14 January 2004 9 and the 21 January 2004, would that have been 10 helpful context for you to have had access to, 11 to the investigation that you were to carry out? 12 A. I don't think there was any additional evidence 13 that would have been of help to me in the NBSC 14 logs because I think that the Horizon Helpdesk 15 had captured certain -- the majority of the 16 information. Again, either -- I'm not sure now 17 whether they got some of this information by 18 talking to the NBSC agents or by talking direct 19 to Mr Castleton, but the Horizon Helpdesk had 20 captured, in various calls, I think probably all 21 the pertinent information about what 22 Mr Castleton was saying. 23 Q. Okay, we'll look at that as we go along. Can 24 you explain, in general terms, how you were 25 passed information from the HSH, the Helpdesk? 73 1 A. They would log the information on their 2 PowerHelp call -- I think that was what the 3 calls were called at that point in time -- and 4 then they'd route the PowerHelp call on to 5 a PEAK or PinICL stack, and that would 6 automatically raise -- I can't remember if it 7 was PEAKs or PinICLs at this precise point in 8 time -- and would paste the information -- the 9 information that had been recorded on the 10 PowerHelp call would also go onto the -- would 11 automatically go onto the newly raised PEAK that 12 would then be on the SSC stack. 13 Q. How would you be passed information from the 14 NBSC? 15 A. We were not passed information directly. It was 16 a matter of the HSH having recorded it on their 17 own call -- PowerHelp call that they had raised. 18 Q. If you wanted to ask for more information from 19 HSH, how would you do that? 20 A. If we felt that there was insufficient 21 information from HSH, then we would route the 22 call back to them saying we need some more 23 information. 24 Q. If you wanted information from the NBSC, how 25 would you do that? 74 1 A. We very rarely contacted anybody at NBSC. 2 I can't say we never did because I know there 3 are a few calls where I did talk to, I think, 4 somebody called Ibrahim but, I think in general, 5 it was HSH's responsibility to get a clear 6 picture of what the postmaster said the problem 7 was and then to route it on to us. I could have 8 talked -- contacted the postmaster myself also, 9 if I needed more information, but in this case 10 I didn't. 11 Q. So, essentially, the Fujitsu HSH, the Helpdesk, 12 were your agents, were your facility for 13 obtaining information from either the postmaster 14 or NBSC? 15 A. Yes. I'm not sure how often they would go back 16 to NBSC to ask for more information themselves 17 either, but certainly they would -- they were 18 expected to get a clear picture from the 19 postmaster before sending the call over to SSC. 20 Q. Thank you. Can we look, then, at this first 21 record of the call, then. 22 SIR WYN WILLIAMS: Sorry, Mr Beer, before we do 23 that -- and this may appear very pedantic -- 24 but, given the date on Mrs Chambers' second 25 witness statement, has she actually ever vouched 75 1 for its accuracy at the Inquiry? 2 MR BEER: That's a really good point! 3 SIR WYN WILLIAMS: I don't think she could have, 4 really. 5 MR BEER: Well, given that I couldn't have displayed 6 it if I wanted to, you're right to pick me up on 7 it. 8 SIR WYN WILLIAMS: It's all right. 9 MR BEER: With your leave, then, sir, shall we do 10 that now? 11 SIR WYN WILLIAMS: Yeah, I think it must be the case 12 that she hasn't done it. So we'd better do it 13 and now is as good a time as any. 14 MR BEER: Turn to page 34, please. Is that your 15 signature? 16 A. Yes, that is my signature. 17 Q. I don't believe we're in the territory of you 18 wishing to make corrections to the witness 19 statement? 20 A. No. 21 Q. Are the contents of it true to the best of your 22 knowledge and belief? 23 A. Yes. 24 Q. Thank you. 25 Thank you very much, sir, for your eagle 76 1 eyes. 2 Can we look, please, at the first record of 3 the call made by Mr Castleton, 14 January 2004. 4 It's LCAS0000365. Can we turn, please, to 5 page 18. 6 We can see here the NBSC call log, yes? Do 7 you recognise it now? 8 A. I recognise it having been sent it by the 9 Inquiry. I wouldn't have known what I was 10 otherwise. 11 Q. Just looking along the top line to orientate 12 ourselves, "Date" -- I think that's the call 13 taken, first column; the "Incident ID", the 14 second column; a "Brief Description" of the 15 event, third column; a "Detailed Description" of 16 the event", fourth column; "Resolution", fifth 17 column -- do you know what "KB" means? 18 A. No. I might guess "Knowledge Base" but that's 19 a guess. 20 Q. A cross-reference to the "Incident Log"; 21 a record of the nature of the call, the "Call 22 Type"; who the client was; the "Activity"; the 23 "Sub-Activity"; the branch code, the "FAD code"; 24 and the name of the Post Office concerned. Yes? 25 A. Yes. 77 1 Q. Can we go, please, to page 25. Can you see, 2 please, that this is an entry of 14 January 3 2004 -- 4 A. Yes. 5 Q. -- at the foot of the page. 6 A. Yeah. 7 Q. So this is the first relevant entry that we're 8 going to look at, the first contact between 9 Mr Castleton and the Post Office. The 14th is 10 a Wednesday, okay? Take it from me. We can see 11 the brief description is "Discrepancy", and then 12 the detailed description is a discrepancy of -- 13 and I think we can tell from looking elsewhere 14 in the document where the hash sign appears, 15 that's a pound sign, essentially -- 16 A. Yes. 17 Q. -- of £1,103.13 loss. Resolution, KB. If that 18 does mean "Knowledge Base", what would that 19 mean? 20 A. I've no idea. This was the NBSC system which 21 I had nothing to do with. 22 Q. We can see some cross-reference to an incident 23 log; the call type, "Horizon Balancing"; the 24 activity, "Cash Account Discrepancy"; 25 sub-activity "Discrepancy"; and then the branch 78 1 code and the office name. 2 A. (The witness nodded) Yes. 3 Q. If we just scroll down to the next page, we can 4 see there's nothing else. It moves on to 5 a different date, yes? 6 A. Mm-hm. 7 Q. Now, if we go back to that page, on 14 January, 8 what was your understanding of what should be 9 done where a postmaster reported a discrepancy 10 of, say, £1,000? 11 A. I would expect, and I don't know, that NBSC 12 would possibly help him to look through various 13 reports, and so on, to see if there could have 14 been any errors in recording data that might 15 have led to that loss. I mean -- 16 Q. Just to be clear, I realise, of course, you 17 didn't work in the NBSC. This was not your 18 function. The purpose of these questions is to 19 understand what you knew about what had been 20 done before a call came across your desk. 21 A. Yeah. 22 Q. So on the face of this, it looks like nothing 23 was done in relation to the report of £1,000 24 loss; would you agree? 25 A. I don't know what they would have told him, and 79 1 I think you have to bear in mind that, you know, 2 discrepancies at branches are not at all 3 unusual. Every branch will have had losses or 4 gains occasionally, which they will then have 5 tried to resolve. I mean, obviously making sure 6 that the cash has been counted correctly and 7 that what they've recorded on the system, which 8 they can check against the reports, is actually 9 what they have done in the office. 10 Q. Yes, and, on the face of it, there isn't any 11 record of that having been done on this 12 occasion? 13 A. It's not written down there but I've no idea 14 what the NBSC procedures were. 15 Q. But something is supposed to be done, isn't it, 16 when a postmaster reports a discrepancy of 17 £1,000? 18 A. I don't know what the procedures were. I would 19 still say that, you know, that the first thing 20 is to look for differences between what is on 21 the system and what actually took place in the 22 office, because that is the cause of 23 discrepancies. 24 Q. Okay. Can we move forwards a week, then, to 25 21 January 2004, and look at page 28. So this 80 1 is the following Wednesday, and you'll see that 2 it's a discrepancy, again; can you see that? 3 A. Yes, I can. 4 Q. The detailed description is: 5 "[Postmaster] has cash account discrepancy 6 of £4,294.67." 7 A. Yes, I see that. 8 Q. In the resolution is: 9 "[Subpostmaster] still has a loss and has 10 logged the call with suspense." 11 A. Yes. 12 Q. What was your understanding of the suspense 13 account? 14 A. That was somewhere to which they could move 15 losses or gains while they were being 16 investigated but I think they were meant to get 17 authorisation from a particular team in the Post 18 Office in order to do that. 19 Q. Was it your understanding -- I mean, it says 20 here "and has logged call with suspense" -- that 21 it was he, the postmaster, that had to log the 22 call with suspense or the NBSC logged the call 23 with suspense? 24 A. I don't have any knowledge of the process. 25 Q. But one of the two would refer the gain or loss 81 1 to a team dealing with suspense accounts; is 2 that right? You understood that? 3 A. I think that's probable but, really, this was 4 completely outside my knowledge. 5 Q. The next column recalls, after the text in lower 6 case, in uppercase: 7 "Checked through transaction logs with 8 [subpostmaster] and nothing showed except the 9 DDN." 10 Did you understand that to refer to 11 a discrepancy? 12 A. Yes. 13 Q. "Advised I will call him back tomorrow to see if 14 anything came to light when he balanced OOH ..." 15 Out of hours, or wouldn't you know? 16 A. Seems likely. 17 Q. "... and log call with suspense if necessary." 18 So this is before the postmaster, is this 19 right, has balanced a record of the NBSC saying, 20 "You're telling me you've got a loss. Let's 21 wait until you balance the next day and, if it's 22 still there, we'll log the call with suspense"? 23 A. Yes, I seem to remember that Mr Castleton tended 24 to start going through the balancing process on 25 the Wednesday afternoon and the 21st would have 82 1 been a Wednesday. 2 Q. Yes. 3 A. But he didn't usually complete the final balance 4 and produce a cash account until first thing on 5 the Thursday morning. 6 Q. But, if you'd been reading this at the time, you 7 would have understood this to be a postmaster 8 saying, "Look I've already got a discrepancy" 9 and they're saying "You need to sign off your 10 cash account the following day and see whether 11 it's there". 12 A. If he hadn't got as far as the final balance, if 13 he'd just done a trial balance, that would have 14 produced a discrepancy, but he could then have 15 gone back through to try to see if he could spot 16 any reasons for that discrepancy. He could then 17 have changed the cash declaration, or whatever 18 other figures, and done another trial balance. 19 So he's started the process to the point 20 that he's realised it's going to show -- it has 21 created a discrepancy but he hasn't actually 22 completed the rollover process. 23 Q. So "I'll call back tomorrow to see if anything 24 came to light", so that's the 21st. If we go 25 over the page to the 22nd, we are at the next 83 1 day. Brief description, "Discrepancy": 2 "[Postmaster] has a loss of £4,000. He was 3 in the office until 11.00 [pm] last night and 4 could not find anything." 5 A. Yes. 6 Q. "Resolution": 7 "Went through all the balance checks with 8 [the postmaster]. He had checked the rems in 9 and out, his cash stock and the P&A ..." 10 Do you remember what the P&A was? 11 A. Pensions and allowances? 12 Q. "... and was unable to find the loss. Advised 13 I would pass through to suspense." 14 So would you understand, at this point, that 15 the postmaster had signed his cash account off 16 by saying that there was a discrepancy, 17 essentially, between the cash and stock, which 18 Horizon showed he should have, and the cash and 19 stock that he, in fact, had -- 20 A. Yes. 21 Q. -- and that, the NBSC are saying, they are going 22 to pass through to the Suspense team, yes? 23 A. That's what they appear to be saying, yes. 24 Q. Yes. You said, I think, earlier that your 25 understanding was that the discrepancy, whether 84 1 it was a gain or a loss, would be placed in 2 a suspense account? 3 A. That was a manual action that, if authorised, he 4 could take, or I think it could be taken -- it 5 could be done, even if not authorised, but my 6 recollection is that they were meant to get some 7 sort of authorisation. 8 Q. If a postmaster was placing an amount in the 9 suspense account -- or was authorised to place 10 an amount in the suspense account -- and was 11 signing off the cash account, was it your 12 understanding that he was acknowledging that he 13 owed the money shown on the cash account, if it 14 was a loss? 15 A. I had no idea of what the legal position might 16 be on that money. 17 Q. Or was it your understanding that, despite 18 signing the cash account, the money was, in 19 effect, disputed and that's why it was in the 20 suspense account? 21 A. Again, I don't know about the business and the 22 legal side of it. All I would see would -- if 23 I looked, was an amount being moved from one 24 place on the system to another. 25 Q. I mean, to be clear, Mrs Chambers, I'm not 85 1 asking you for what your understanding of the 2 legal position was because I can't imagine you 3 ever address your mind to it. It's more what 4 you thought was signified by an amount being 5 held in a suspense account and was it that sum 6 is disputed? 7 A. Um, I'm not even sure if I knew that at that 8 point but then, yes, there's a loss and they've 9 parked it there. 10 Q. What did you think was the status of money then 11 held in the suspense account or did you think 12 nothing about that? 13 A. I don't think I thought about it, just that it 14 had been sort of a sum had been parked there. 15 Q. Parked: did you have an understanding of parked 16 for what purpose or parked until what? 17 A. I don't think I really thought about it. 18 Q. Can we move forwards, then, to 28 January 2004 19 and turn to page 32. So we're back to the next 20 Wednesday, yes? 21 A. Mm-hm. 22 Q. The 28th, the third week, the third Wednesday. 23 I'm reading under "Detailed Description": 24 "[Postmaster] says since broadband installed 25 every time he receives stock into [the] office 86 1 he is showing short by same amount. Has checked 2 to make sure remmed in properly but still 3 showing £2,500 short amount of stock. Rem needs 4 CB to check if accounting or system [problem]." 5 Are you able to decode for us what that last 6 part means? 7 A. No. 8 Q. Do you know what "CB" referred to? 9 A. No. 10 Q. "Resolution": 11 "Advised [the postmaster] that if he feels 12 this is a technical problem to call [the Horizon 13 Helpdesk] but after talking to [the postmaster] 14 he is entering all [transactions] ok, so this 15 could be the case. Advised him to balance and 16 roll for a definite figure then to call [the 17 Horizon Helpdesk]. If no joy call TP." 18 Do you know what "TP" was a reference to? 19 A. It might have been the Transaction Processing 20 team but I'm not certain. 21 Q. What did the Transaction Processing team do? 22 A. I don't know. I'm -- 23 Q. Were they part of Post Office or Fujitsu? 24 A. Post Office, I think. I mean, there was a daily 25 feed of figures at this point from Horizon to 87 1 a Post Office team via a system called TPS. And 2 so whether -- I'm not -- I really don't know but 3 it might have been a team inside Post Office who 4 were getting transaction data at some level sent 5 to them. 6 Q. But, again, is this the same as the previous 7 week, essentially the Post Office saying to the 8 subpostmaster "Let's not sort your problem out 9 now, do the end of week balance, fill out your 10 cash account tomorrow, see if there's still 11 an issue then"? 12 A. That appears to be what they are saying. 13 Q. Now, we've seen therefore here, in the NBSC 14 records, the postmaster was being told to call 15 the Horizon Helpdesk, the Fujitsu side of the 16 house, if there remained a problem. So can we 17 turn, then, to the Fujitsu side of the house to 18 see whether there's anything in the Fujitsu 19 records. Can we look, please, to FUJ00122322. 20 Are these a record of the PowerHelp calls? 21 You have to say something rather than nod your 22 head. 23 A. Sorry, yes. I didn't realise it was a question. 24 Q. Yes, it was: are these a record of the PowerHelp 25 calls? 88 1 A. Yes. 2 Q. Is the answer to that yes? Okay. These are 3 things that you would have had access to when 4 you came to look at the problem on 26 February 5 2004 but none of the documents we've looked at 6 already? 7 A. Yes. 8 Q. Okay. We can see that the call was opened on 9 28 January at 11.13. Can you see that in the 10 middle? 11 A. Yes. 12 Q. It was closed eight minutes later on the same 13 day, at 11.21? 14 A. Yes. 15 Q. We can see that the caller on the left-hand side 16 is said to be "Liam", I think that's supposed to 17 be "Lee". If you look at the "Problem Text", 18 which is just over halfway down: 19 "Caller states that discrepancies are going 20 through on the system. And this has been the 21 case for 3 weeks in a row. Week 1: £1,103 down. 22 Week 2: £4,230.97 down. Week 3 (today): 23 [approximately] £2,500." 24 Can you see that? 25 A. Yes, I can. 89 1 Q. Then all of that essentially matches what we've 2 seen in the NBSC logs, doesn't it, in terms of 3 the figures and the duration of the problematic 4 events? 5 A. Yes. 6 Q. Then if we go down, please, to the call activity 7 log. Thank you. The first line of text at 8 11.11: 9 "New call taken by Dane Meah: Caller states 10 that discrepancies are going through on the 11 system. And this has been the case for 3 weeks 12 in a row." 13 Then the three figures we've seen are set 14 out, yes? 15 A. Yes. 16 Q. "Caller states that these discrepancies have 17 been relevant to the level of stock currently 18 being held." 19 A. Yes. 20 Q. Then next: 21 "[Advised] caller that this problem will 22 need to be thoroughly investigated by NBSC 23 before the issue can be investigated as 24 a software problem." 25 A. Yes. 90 1 Q. "Transferred the caller to the NBSC so that the 2 incident could be investigated further." 3 Call closed, it's an NBSC issue, 4 transferred; can you see that? 5 A. Yes, I can. 6 Q. Was this something you came across frequently: 7 subpostmasters reporting losses or gains, 8 discrepancies, and being passed from one 9 organisation to the other with neither taking 10 responsibility? 11 A. I wouldn't say frequently but, yes, it did 12 happen. 13 Q. You told us, I think, on the last occasion that 14 you were concerned about people being bounced 15 from NBSC to the Helpdesk and back again? 16 A. Yes. 17 Q. And does this appear to be an early example of 18 it? 19 A. Yes, it appears to be an example of it. Yeah. 20 Q. He had reported, consistently, calls to the 21 NBSC; they'd not investigated them; they'd told 22 him to contact the Helpdesk at Fujitsu; he 23 contacts the Helpdesk at Fujitsu and they tell 24 him this has got to be thoroughly investigated 25 by the NBSC? 91 1 A. Yes. 2 Q. I think you can help us with this, you tell us 3 in your witness statement -- no need to turn it 4 up now -- the SSC wasn't involved at all at this 5 time because the HSH bounced it back to NBSC, 6 and had not created a PinICL or a PEAK? 7 A. Yes. 8 Q. Thank you. Can we move forwards to 29 January, 9 please, which is LCAS0000365, and page 33, 10 please. 11 So he's back with the NBSC now, yes. 12 A. Yes. 13 Q. The next day, on the 29th, "Cash account 14 discrepancy". Detailed description: 15 "Is showing a loss of £2,523.12. Says this 16 is the third discrepancy in as many weeks." 17 "Resolution": 18 "Checked through figures using transaction 19 log by mode and also amount looked at Rems 20 declaration and cash flow. No trace of the 21 discrepancy. Created call for Suspense Account 22 team." 23 A. Yes. 24 Q. So he's been to the NBSC three times. They've 25 said on the last occasion "Go to the HSH". He's 92 1 gone to the HSH. They say "This needs to be 2 fully investigated by the NBSC". He's gone back 3 to the NBSC. They've said, "We'll create a call 4 for the putting of the money in the suspense 5 account". 6 Does that appear to be accurate? 7 A. That appears to be what they're doing, yes. 8 Q. So, on the face of it, this call isn't being 9 investigated by the NBSC, the money is just 10 being put, or potentially being put, in the 11 suspense account again? 12 A. Yes, the NBSC are trying to help him find out 13 what might be the cause of the discrepancy by 14 helping him use the transaction log and checking 15 the REM declarations but they're unable to 16 pinpoint anything there that might have caused 17 the discrepancy. 18 Q. Can we move forward to later on the same day, 19 I think, page 35, so still on the 29th. 20 "Detailed description": 21 "[Postmaster] would like to have his 22 transactional archives looked at more closely to 23 try to identify what is going wrong with his 24 office. He is having a lot of losses over the 25 last three weeks and thinks there is a system 93 1 fault with his remittances." 2 A. Yes. 3 Q. "Resolution": 4 "I have followed KB instructions and sent 5 an email to Adele Kilcoyne so [postmaster] can 6 study his archives to try to identify what has 7 gone wrong. All details are in the log." 8 A. Yes. 9 Q. Is that a request that you would encounter often 10 in your many years working at Fujitsu, 11 subpostmasters wanting their/"transactional 12 archives" raked over? 13 A. This is an NBSC call log so I can't really 14 comment on that. I don't recall calls like that 15 coming through to SSC, in particular. 16 Q. You can't remember many calls or incidents 17 passed to the SSC, where the subpostmaster was 18 asking for, as it's recorded here, his 19 transactional archives looked at closely? 20 A. No, I don't recall receiving calls at SSC that 21 said that. 22 Q. Can we go back to the HSH, please, FUJ00122322, 23 and look at page 2, please. Still on the same 24 day -- sorry, page 2, thank you -- 29 January 25 2004. You can see call opened 10.26 and 10.31. 94 1 I should make it clear I'm not completely clear 2 as to the precise order of the last two 3 documents and this one. These are timed; the 4 other ones are not. 5 But, in any event, this is a record of 6 a call made the same day and we can see that it 7 lasted or the log was open for five minutes. 8 The caller on the left-hand side is recorded, 9 correctly this time, as Lee, the postmaster and 10 we can see the "Problem Text", a third of the 11 way down the page: 12 "[Postmaster] reports [I think 'he'] is 13 having problems on his system connected to rems. 14 Every time he rems in it leaves him with 15 a discrepancy and he has been to the NBSC and 16 back to us now and wants his system 17 investigating." 18 Then down to the "Call Activity Log": 19 "New call taken by Mary Rainbow: 20 [postmaster] reports he is having problems on 21 his system connected to rems, every time he rems 22 it in leaves him with the a discrepancy and he 23 been to the NBSC and back to us and now wants 24 his system investigating. 25 "Advice: advised the [postmaster] for this 95 1 to happen he needs to be re-referred from the 2 NBSC. 3 "Caller referred to the NBSC. 4 "Call [closed]." 5 Again, is that another example of what you 6 told us about last time, postmasters being 7 bounced between the two organisations? 8 A. Yes. 9 Q. Then lastly before lunch, can we look, please, 10 at page 3 of this document, moving forward to 11 13 February, which is a Friday. Caller, 12 Lee Castleton. "Problem": 13 "Marie @ NBSC -- [postmaster advised] his 14 system is doubling up cash declarations and 15 cutting off cheques. They still appear the next 16 day." 17 Then "Call Activity Log": 18 "New call taken by Tony Law: Marie @ the 19 NBSC -- [postmaster advised] his system is 20 doubling up cash declarations and cutting off 21 cheques. They still appear the next day. 22 "NBSC [advise] they have: checked that he is 23 cutting everything off properly." 24 Just stopping there, what do you understand 25 "cutting off" to mean in the context of cheques? 96 1 A. At the end of each day, they have to print off 2 a report of cheques. They are then -- the 3 cheques had to be posted off somewhere, I think, 4 with the report or a copy of the report as well, 5 and then there was a button that they had to 6 press on the system to say, "I've done that", 7 and that effectively sort of drew a line under 8 that set of cheques so that they wouldn't appear 9 on the next day's report. 10 Q. Thank you. So: 11 "NBSC [advise] they have: checked that he is 12 cutting everything off properly, cash figures 13 are being done properly. 14 "[Postmaster] has insisted on a system 15 check." 16 Skipping a line: 17 "Problem has been happening for 5 weeks. 18 "Every time stock has been remmed in they 19 have had a loss that night. 20 "Snapshots and [transaction] logs agree with 21 [the postmaster's] figures." 22 What would you understand that to mean: 23 snapshots and transaction logs agree with the 24 postmaster's figures? 25 A. Not entirely clear. I mean, a balance snapshot 97 1 was a report that they could print to get 2 a picture of the current system figures. But 3 I don't know precisely what that sentence is 4 meant to mean. 5 Q. "[Postmaster advised] NBSC have done: 6 [transaction] logs, Girobank logs, 7 reconciliations, rems in/out, stock holding. 8 "[Postmaster advised] there is a Sunday decs 9 but no one is on site to do this on a Sunday. 10 "5 weeks ago [postmaster advised] he did not 11 rem in any stock last week and balanced 12 perfectly. Every week that he does rem stock 13 in, the balance is wrong. 14 "Escalated to Heather Dryden. 15 "[Postmaster] states on the cheques listing 16 for the 11th it states all of cheques for the 17 10th as well. When I got him to go through his 18 cheques, [postmaster] stated that this happened 19 more than once but when going through his 20 cheques it had actually happened just the once. 21 [Postmaster] stated he was sure he cut off, so 22 advised I would log all details and call back if 23 it happens again. 24 "Call [closed]. [Postmaster] states his 25 cheques keep carrying over from the day before. 98 1 Advised [the postmaster] to call back if it 2 happens again ..." 3 Yes? 4 A. Yes. 5 Q. So we see that, in the course of this call, it's 6 recorded that the postmaster insisted on 7 a system check. That wasn't actioned at the 8 time, was it? 9 A. No, it appears that they followed up on the 10 specific thing -- problem with the cheques 11 listing and -- which I think they did actually 12 deal with correctly but they didn't then go back 13 to all the other issues that he was also 14 reporting. 15 Q. They didn't go back to those? 16 A. No. They only dealt with the cheques listing, 17 part of what -- the problems that he was having. 18 Q. They dealt with that by saying, "Let's see 19 whether it happens again"? 20 A. No, they -- I think that probably what they did 21 was get him to look at the events -- the log of 22 Riposte events, which he could look at on his 23 own system, which showed whether a report had 24 been cut off on a particular day. And, I mean, 25 I know from having looked at that same data 99 1 subsequently, that, yes, there was a day where, 2 although he was under the impression that he had 3 done a cut-off, it had not been done and that 4 explained why the cheques listing on the 11th 5 included the cheques from the 10th as well. 6 Q. You tell us in your witness statement that 7 a failure to cut off cheques wouldn't be 8 connected to the cash discrepancy issue? 9 A. No. It has no impact on the cash. 10 Q. Can you explain why, please? 11 A. Because this -- the cut-off just affected what 12 showed on the cheques report each day. As long 13 as they'd remmed out the cheques correctly, 14 which they had done, that meant that the value 15 of the cheques had been removed from the system, 16 as it should be. Cheques also didn't directly 17 affect discrepancies because a discrepancy is 18 a difference in the cash totals. 19 MR BEER: Thank you very much. 20 Sir, that would be a convenient moment 21 before we move to 25 February 2004. I wonder 22 whether we can break until 2.05. 23 SIR WYN WILLIAMS: Certainly. Yeah. 24 MR BEER: Thank you very much. 25 (1.04 pm) 100 1 (The Short Adjournment) 2 (2.05 pm) 3 MR BEER: Good afternoon, sir. Can you see and hear 4 me? 5 SIR WYN WILLIAMS: Yes, I can, thank you. 6 MR BEER: Thank you very much. 7 Good afternoon, Mrs Chambers. We had dealt 8 with calls in January and the early part of 9 February. We'd finished on 13 February. Can we 10 turn to another Wednesday for Mr Castleton, 11 Wednesday, 25 February 2004. It's FUJ00122322, 12 and to page 8, please. 13 So this is, is this right, a PowerHelp call 14 record for 25 February 2004? 15 A. Yes. 16 Q. We can see it was open for half an hour. 17 A. Yes. 18 Q. We can see that this time the caller was 19 Christine -- 20 A. Yeah. 21 Q. -- who I think was one of the clerks or a clerk 22 at Mr Castleton's branch. We can see the 23 problem stated under "Problem Text": 24 "Clerk reports that they have been having 25 problems on the system when balancing that seems 101 1 to be related to stock remmed in through the 2 week." 3 If we scroll down, please, to "Call Activity 4 Log": 5 "New call taken by Mary Rainbow ..." 6 So we're back with Mary Rainbow again: 7 "... clerk reports they have been having 8 problems on the system when balancing that seems 9 to be related to stock remmed in through the 10 week. 11 "Clerk reports that over the last seven 12 weeks they have had losses every week, at one 13 point they had a problem with cash on hand but 14 they was found to be an issue with ID numbers 15 and has been resolved." 16 A. Yes. 17 Q. "Clerk reports that last week when they rolled 18 over they put a loss into suspense account and 19 then took the unit back to trial balance and 20 came out with a zero net discrepancy to start 21 the new CAP ..." 22 Is that cash accounting period? 23 A. Yes. 24 Q. "... with. 25 "Clerk reports that she printed a balanced 102 1 snapshot on Monday and all looked okay but since 2 then they have remmed in some stock which 3 appears to have given them a loss. 4 "Advised the clerk that she will need to 5 check her stock position for last week and then 6 check her rems in summary, this should give her 7 the stock she should be holding. This figure 8 can then be compared with the stock showing in 9 adjust stock and this will highlight any 10 problems with stock on hand. 11 "Clerk reports they were told that checks 12 would be made on the system but she is unsure if 13 this information came from NBSC or HSH as she 14 has no names or [reference] numbers. 15 "Advised the clerk that we need her to take 16 the unit to trial balance this evening and come 17 back before she rolls over so we can take down 18 any detail she can give us. 19 "Clerk reports they usually report the 20 balance at about [5.30 pm] or maybe before. 21 "Spoke to Matt Saunders and he suggests call 22 should be passed to someone line, Heather Dryden 23 [that's presumably 'someone like Heather 24 Dryden'], if clerk does call back wanting 25 further assistance. 103 1 "The clerk reports they are working in 2 a shared AA stock unit. 3 "Clerk to call back tonight if further 4 assistance required. 5 "Clerk to call back if further assistance 6 required. Call [reference] taken. 7 "Call [closed]: clerk reports problems with 8 the balance that appear to be linked to stock 9 remmed in -- clerk will call back tonight if 10 further assistance required." 11 Can we then look to later that day, please, 12 page 10. So later on the 25th, we're now at 13 12.03 pm on the same day. I think this is the 14 NBSC calling the HSH; is that right? 15 A. It looks like it, yes. 16 Q. Looking at the "Call Activity Log": 17 "New call taken by Robert Congerton: NBSC -- 18 [postmaster] has called the NBSC regarding 19 problems with her balance. 20 "[Postmaster] called in at 11.03 today 21 regarding problems with her balance. Advised 22 NBSC that the agent advised the [postmaster] to 23 call back in tonight when they have a net 24 discrepancy on the cash account. 25 "Call [closed]." 104 1 Yes? 2 A. Yes. 3 Q. So then can we turn to page 14, please. This is 4 still the same day, when the call is opened, 5 Wednesday, 25 February at 5.33 pm. Again, can 6 we read the call activity: 7 "New call taken by Kuljinder Bhachu: 8 [postmaster] reporting they are getting large 9 discrepancies for the last few weeks. 10 "Looking at closed calls for this site, 11 there have been a number of calls logged 12 regarding discrepancies. NBSC have been in 13 contact with the [postmaster] and cannot find 14 any user error." 15 Just stopping there, do you know what NBSC 16 used to do in order to determine whether or not 17 a discrepancy reported had, as its cause, user 18 error? 19 A. As I said earlier, getting the postmaster to 20 check the transaction logs against any other 21 records that he had in the Post Office. 22 Q. "Spoke to Sandra @ NBSC ... regarding this 23 issue. Checked Tivoli events and health 24 checked. Site is health checking ok." 25 Can you help us with what a "health check" 105 1 was? 2 A. That the comms were all working okay, no sign of 3 any comms problems. 4 Q. "Critical event seen ... on 18 February 2004 5 ..." 6 A KEL reference is given. I'm going to 7 ignore the rest of it, until we pick up at 8 18.09: 9 "Spoke to [postmaster], who advises that the 10 problem with the [Cash Account] started ever 11 since the BT engineer came to move the BT box 12 for the preparation of ADSL [as it's called]." 13 A. Yes. 14 Q. "Username ... other balance users ... balance on 15 Wednesday after 5.30." 16 Then this: 17 "Could SSC please investigate why this [post 18 office] is experiencing large discrepancies ever 19 since BT engineer has moved box in preparation 20 for ADSL [installation]. KEL [reference] given 21 as possible problem." 22 You're later -- taking matters shortly, to 23 say that KEL reference was to do with something 24 else completely; is that right? 25 A. Yes, as I recall, yeah. 106 1 Q. "NBSC have said there is no user error. Thank 2 you." 3 Is this, therefore, the event which 4 generates the PinICL which triggers your 5 involvement? 6 A. Sorry, yes, you're not talking about a system 7 event; you're talking about the call being 8 handed -- passed over to SSC, yes. 9 Q. Yes, so the request for SSC involvement we see 10 at 18.14: 11 "Could SSC please investigate why this [post 12 office] is experiencing large discrepancies ever 13 since the BT engineer has moved the BT box ..." 14 A. Yes. 15 Q. Is that the task that you subsequently 16 undertook, investigating all of the 17 discrepancies that had been reported since the 18 BT engineer had moved the BT box? 19 A. I can't remember what the date was when the BT 20 engineer did move the box. 21 Q. It's described here as "seven weeks ago". 22 A. Okay, I -- 23 Q. I don't think you did conduct -- 24 A. No, I -- 25 Q. -- such a broad investigation? 107 1 A. No, I concentrated mostly on the last couple of 2 weeks, which was when I had most of the evidence 3 for. 4 Q. Given that this was the request, "Could SSC 5 please investigate why this [post office] is 6 experiencing large discrepancies ever since the 7 BT engineer had moved the box", why did you look 8 only at the last couple of weeks? 9 A. Because that was the data that was most readily 10 available and I think I assumed that, if there 11 was a problem, an ongoing problem, then I would 12 find evidence of it most easily in the most 13 recent weeks. 14 Q. What about if there was more than one problem? 15 A. Gosh, um -- 16 Q. A problem or problems in the earlier weeks, 17 different from any problem in the later weeks? 18 A. Yes, it's possible, then, if there was a problem 19 in the -- a different problem in the earlier 20 weeks, that my investigations might not have 21 found it. I -- obviously, even 18 months beyond 22 this, I couldn't remember precisely what my 23 investigation was. I knew I had very thoroughly 24 investigated a week or two -- 25 Q. My question is: why did you investigate only 108 1 a week or two? 2 A. As I said, because I felt that, if there was 3 an ongoing underlying cause, then I would find 4 that most readily in the most recent week. 5 Q. Would you not take this as your terms of 6 reference or your instruction, the request that 7 essentially dates back to when the first problem 8 was reported back on 14 January 2004? 9 A. I don't think I would necessarily have done 10 this. I was just doing a general investigation 11 to see if I could find the cause for the 12 discrepancies. 13 Q. But only the recent discrepancies? 14 A. That was my starting point, yes. 15 Q. Well, it was your ending point too, wasn't it? 16 A. Yes, it was my -- yes, given that I now cannot 17 remember precisely what I did over which stage 18 but, yes, if there'd been -- I probably did not 19 investigate thoroughly back towards the 20 beginning of January. 21 Q. Can we turn, please, to the PinICL, please, 22 which is essentially your originating 23 document -- 24 A. Yes. 25 Q. -- would that be fair? 109 1 A. Yes. 2 Q. That's FUJ00146165. In fact, it's a PEAK, not 3 a PinICL. We can see the PEAK number, 4 PC0099954, and we can see it was opened on the 5 day we've just referred to, 25 February 2004, at 6 6.19 pm; can you see that? 7 A. Yes. 8 Q. I think if we look at page 2, we can see that it 9 was assigned to you at 9.41. Can you see that 10 just below halfway down -- 11 A. Yes. 12 Q. -- and that's the following day -- 13 A. Yes. 14 Q. -- 26 February -- 15 A. Yeah. 16 Q. -- at 9.41? 17 A. Yes. 18 Q. Does that mean that you get it then or does it 19 just mean that it has been assigned and you 20 don't receive anything until you log on to 21 a system? 22 A. When I next looked at my PinICL stack, I would 23 have seen it but I don't know -- that wasn't 24 necessarily at 9.41. 25 Q. I think we can see that you had looked at your 110 1 stack by at least 10.17 -- 2 A. Yes. 3 Q. -- because you make an entry at 10.17, right at 4 the foot of the page -- 5 A. Yes. 6 Q. -- when you changed the description in the call 7 summary box? 8 A. Yes. 9 Q. I think others may ask you about that. So this 10 is the first record we see of you doing any work 11 on this event, yes? 12 A. Yes. 13 Q. Can we go back to page 1, please. What on this 14 page do you see when you open up this PEAK? 15 A. Just that they're reporting large discrepancies. 16 It says there's been a number of calls logged. 17 Q. Just stopping there, it says: 18 "[Postmaster] reporting that they are 19 getting large discrepancies for the last few 20 weeks." 21 That's not really accurate, is it, on the 22 basis of what we've seen? 23 A. Is that not accurate? 24 Q. Didn't it go back to 14 January? 25 A. Well, I'd have said that the last few weeks from 111 1 the 25 February probably does just about include 2 14 January. 3 Q. Was that phrase "last few weeks" what 4 conditioned your investigation to any look at 5 data for a week or two? 6 A. I don't think so. I think I then did look at 7 the closed calls and so then I did get the 8 picture and I could see that, yes, they had been 9 reporting problems since -- was it 14 January -- 10 when they had the £1,100 loss. 11 Q. Where would you have seen that they had been 12 reporting calls since 14 January from? 13 A. Was that not on the PowerHelp calls? 14 Q. Can you look, please, at FUJ00122322. The first 15 PowerHelp call is on 28 January -- 16 A. Yeah. 17 Q. -- and it's only in the NBSC call log that one 18 sees the records for 14 and 21 January -- 19 A. But this -- 20 Q. I'm so sorry, you'll see in the "Problem Text" 21 that it's recorded that it's been the case for 22 three weeks in a row? 23 A. Yes. 24 Q. Are you saying that you read this and realised 25 that the 28th wasn't the first report, it had 112 1 been going on for three weeks beforehand? 2 A. Yes, I think so. I mean, I read that now and 3 that's what I read, and I think I would have 4 seen that at the time. 5 Q. Yes. So why wouldn't you investigate the three 6 weeks beforehand? 7 A. Because the message store data would not have 8 gone back to 14 January. 9 Q. How far back would it have gone, when you were 10 looking at 26 February? 11 A. I'm hesitating because -- 12 Q. We've heard different dates in the Inquiry, so 13 I'm not going to suggest how long the message 14 store was retained for? 15 A. At one point, messages were kept for 42 days and 16 then that was reduced and they were then kept 17 for only around 35 days, I think it was. But 18 I can't remember when that change came in. 19 Q. We've heard both of those dates. Thank you. So 20 was it that you thought that, if there was 21 a system problem, it would be disclosed by 22 investigating recent weeks, that you limited 23 your investigation to recent weeks or was it 24 because the data wasn't available on the message 25 store? 113 1 A. I think, primarily, it was because I thought, 2 you know, if there was an ongoing problem 3 I would find the evidence for it in the recent 4 weeks and that was the obvious place to start 5 looking. 6 Q. Can we go back, please, to FUJ00146165, which is 7 the PEAK? 8 A. Yes. 9 Q. So you can see all of this data on the first 10 page, ie every entry that's made before you look 11 at 10.17 the following day? 12 A. Yes. 13 Q. Would your practice have been to read what's 14 here or were there certain bits that you know to 15 go to, without reading it line by line? 16 A. I certainly would have glanced over it. 17 I probably would have gone back and looked at 18 all the PowerHelp calls, as well, in case there 19 was more information on those that, for some 20 reason, wasn't on the PEAK. Certainly the PEAK 21 would only contain the information from the 22 latest, that -- the PowerHelp call on which this 23 was based. But I would have gone and looked at 24 everything on PowerHelp. 25 Q. Do you see anything else automatically, whether 114 1 by sort of hyperlink or other electronic 2 connection, when you open up the PEAK, or do you 3 have to go and find it yourself? 4 A. I have to go and find it myself. 5 Q. I think you tell us that, first off, you looked 6 for other PEAKs or PinICLs and you tell us in 7 paragraph 29 of your statement that there were 8 none? 9 A. Yes. 10 Q. That, would you agree, would be because none had 11 been raised because Mr Castleton had been passed 12 backwards forwards between NBSC and HSH a number 13 of times and nobody had referred the matter to 14 the SSC? 15 A. Yes, but I went looking for other PEAKs purely 16 because of that. I'd have seen that from the 17 PowerHelp calls anyway. I was also looking in 18 case there were any reconciliation calls or 19 system event type calls, for which PEAKs had 20 been raised, which I might not have found so 21 readily on PowerHelp. 22 Q. How would you search PinICLs or PEAKs, by 23 reference to the FAD code? 24 A. The FAD code, yes. 25 Q. So you're looking whether there's anything 115 1 branch specific? 2 A. Yeah. 3 Q. You tell us in your witness statement, there's 4 no need to turn it up, it's paragraph 30, that 5 you checked whether the branch appeared in any 6 reconciliation reports? 7 A. Yes. 8 Q. Can you explain what reconciliation reports are, 9 please? 10 A. There was a whole suite of checks that were made 11 comparing, for example, totals from the 12 counters, for the transactions each day, with 13 the totals that were -- had reached the data 14 centre to make sure nothing had got lost being 15 moved from when it was replicated from one to 16 the other. Yeah, there was a large suite of 17 checks. 18 Also, the cash account was recalculated at 19 the data centre, based on the copy of the 20 transactions there, and compared with the cash 21 account that had actually been produced at the 22 branch. So this was just checking the integrity 23 of the data, if you like. 24 Q. For what period did you check the reconciliation 25 reports? 116 1 A. I believe those reports were only kept for two 2 weeks. That's what I said, I think, back in 3 2006, when I probably knew, and so I just would 4 have checked all the reports as far as back as 5 they existed. 6 Q. Can we look, please, at LCAS0000112. This is 7 your witness statement in the proceedings. As 8 you say, it's dated 14 September 2006. Can we 9 look at the second page and look at paragraph 7 10 at the bottom. You say: 11 "On 26 February, I also checked for any 12 central reconciliation report entries for the 13 branch for the [past] 2 weeks which might 14 indicate a system problem." 15 So that's the two-week period? 16 A. Yeah. 17 Q. You don't say there "They were the only ones 18 that are available to me". 19 A. No, I didn't say that then. 20 Q. Were they the only ones available to you or did 21 you frame the request, namely only to look back 22 a couple of weeks? 23 A. No, I mean, the way I checked was just to look 24 for all occurrences of the FAD code in any files 25 in the folder that held the reconciliation 117 1 reports. So I checked everything that was 2 available at that time. 3 Q. Were those reports automatically generated by 4 the system? 5 A. Yes. 6 Q. That can come down, thank you. 7 Which part of the system were they 8 automatically generated by? 9 A. There was -- part of the overnight processing at 10 the data centre produced the -- well, compared 11 the sets of figures that it received from 12 various sources. 13 Q. Did they rely on Riposte as part of the process 14 for the compilation? 15 A. Only in so much the data that was aggregated for 16 the totals to be compared had to be -- I mean, 17 Riposte was the mechanism which read the message 18 store. So -- sorry, I'm trying to think about 19 this. 20 So yes, I mean, on the counters, the figures 21 that were used for comparison, there must have 22 been some sort of Riposte query, saying "Add up 23 this data", out of the message store. And then 24 I presume at the data centre it was also using 25 Riposte to access the data centre to access the 118 1 messages for the branch. 2 Q. If there were errors or discrepancies in the 3 compilation of these reports, would you have any 4 means of checking that that was so? 5 A. I mean, I think if there were errors on one side 6 or the other, then we would be seeing a lot 7 of -- you'd get a lot more entries on the 8 reconciliation reports because, otherwise, it -- 9 it seems pretty unlikely that you would get the 10 same Riposte error happening on a counter and at 11 the data centre at the same time, if -- and if 12 you had it -- had an error on one side but not 13 on the other then the figures wouldn't match and 14 you'll get an entry on the reconciliation 15 reports to be investigated. 16 And I've -- I think, very occasionally, 17 there were entries that were due to some sort of 18 inconsistency in calculation rather than 19 anything else but they were all investigated, as 20 were any reconciliation report entries. 21 Q. You tell to in your witness statement, and this 22 is paragraph on page 9 -- no need to turn it 23 up -- you looked at the HSH raised calls for the 24 branch on the PowerHelp system as you've 25 described? 119 1 A. Yes. 2 Q. We've seen those already. They go back to 3 28 January and the NBSC call log goes back 4 further than that? 5 A. Yes. 6 Q. But there's a reference on that first page to 7 three weeks that the problem has been occurring 8 for, yes -- 9 A. Yeah. 10 Q. -- with the amounts written down -- 11 A. Yes. 12 Q. -- on each of those three weeks? 13 A. Yes. 14 Q. I don't think you investigated those historic 15 balance discrepancies, did you? 16 A. At that point no, I probably didn't. 17 Q. In paragraph 34 of your witness statement -- 18 I wonder whether we could turn that up, please, 19 WITN00170200, page 11, please. Paragraph 34, 20 which is at the foot of the page, you say: 21 "As well as examining the message store 22 using a text file viewer, I probably extracted 23 all the transaction messages into a spreadsheet 24 so they could be more easily examined. 25 I certainly did this for the cash lines 120 1 (product 1). I cannot remember whether this was 2 for a week or longer, nor which week(s) 3 I checked." 4 A. Yes. 5 Q. Now, of course, when you were doing this you 6 didn't realise that you would be called to give 7 evidence about what you had done? 8 A. Yes. 9 Q. Is that fair? 10 A. Yes, that's fair. 11 Q. So you didn't keep a record of the work that you 12 had undertaken? 13 A. No, I -- 14 Q. At the end of the process, you summarised it in 15 an entry in the PEAK? 16 A. Yes, I would have kept the spreadsheet and the 17 message store in my own file store for some time 18 afterwards but I didn't still have it in 2006. 19 Q. So when it came 18 months later to make the 20 witness statement -- in fact, it's longer than 21 18 months later, it's two and a half years 22 later -- you didn't have your workings out 23 available to you? 24 A. No. 25 Q. If you'd known that you were going to be 121 1 a witness, whether a witness of fact or 2 an expert witness, would you have kept them? 3 A. Yes, I would have done. I could have attached 4 them to the PEAK, which would have been a safe 5 place to keep them, but we tended only to put 6 data on PEAK for calls that were then going to 7 be passed on to Development or whatever. 8 Q. Again, you say that you can't remember whether 9 you did this, ie extracting messages into the 10 spreadsheet from the message store, for a week 11 or longer. Why wouldn't it be for all of the 12 period, given that the concerns had been raised 13 a month and a half previously? 14 A. Yes, it might have been. I can't remember. 15 I know I -- I didn't want to say that I had done 16 more than I had done, when I couldn't truly 17 remember what I'd done or not. I know, 18 I absolutely certainly did it for a week. It 19 could well have been longer than a week but 20 I can't remember. 21 Q. If you had a problem revealed on the PEAK to be 22 of seven weeks' duration, wouldn't the natural 23 thing to do, if the data was available, to 24 examine the available data for that seven-week 25 period? 122 1 A. Yes, it would seem to be so, and I may have done 2 it but I cannot swear that I did do it. 3 Q. Again, the two weeks of reconciliation reports 4 you looked at, that was a limiting factor only 5 because they were the only reconciliation 6 reports available? 7 A. I checked all the reconciliation reports that 8 were available, yes. 9 Q. That two-week period wasn't the governing period 10 for the entirety of your investigation? 11 A. No. 12 Q. How much time did you take undertaking the 13 investigation? 14 A. Well, I'm aware I closed it before the end of 15 that day, so it was a few hours' work. 16 Q. If we look at FUJ00146165, second page, please. 17 If we scroll down to 10.17, we can see your 18 entry at 10.17. So you're certainly aware of 19 this PEAK by the time of that entry? 20 A. Yes. 21 Q. If we look at the bottom of the page, please. 22 We can see your entry of 3.16 pm that day -- 23 A. Yes. 24 Q. -- which is your last entry, essentially? 25 A. Yes. 123 1 Q. Would you have spent the five or so hours 2 between those two times only on this or would 3 you have been looking at PEAKs and other issues 4 too? 5 A. I was probably only looking at this. 6 Q. You said a couple of hours earlier, it looks 7 like it's about five hours? 8 A. Yes. 9 Q. Can we turn, please, to your witness statement, 10 WITN00170200, and look at page 4, please, and 11 paragraph 13. You say here that a call -- this 12 not one we've looked at, and you give the 13 number: 14 "... was not obviously referring to 15 discrepancies. It was logged when Mr Castleton 16 found the message 'Desktop intialisation failed' 17 on the counter 2 screen on the Monday morning. 18 Attempts to clear the problem by rebooting, and 19 by deleting the training message store, were not 20 successful so an engineer was sent to replace 21 the base unit with a new one. As part of the 22 standard base unit replacement process, a zipped 23 copy of the previous night's message store was 24 copied from counter 1 to the new counter 2, then 25 all subsequent messages were replicated across. 124 1 This is completed at around 2.20 pm and then 2 LCA001 was able to log on to counter 2. It was 3 not possible to carry out transactions on 4 counter 2 before then, (though see below)." 5 This issue you're talking about here, the 6 installation of the new base unit, was this part 7 of your investigation at the time, ie on 8 26 February 2004? 9 A. I don't recall that I did look closely at this 10 at that time. 11 Q. So this is something that you're telling us 12 now -- 13 A. I -- yes. 14 Q. -- by looking at documents now -- 15 A. Yes. 16 Q. -- rather than something you did back in 2004? 17 A. Yes, I looked at it, of course, in 2006 and I've 18 looked at it again now. 19 Q. These messages that you're talking about here 20 are dated 1 and 2 February 2004. Was the reason 21 that you didn't look back at those at the time 22 that they were outside the date range that you 23 were looking at? 24 A. No, I think it was because it wasn't relevant to 25 discrepancies, rather than that it was outside 125 1 the date range. 2 Q. So you would have looked at them and seen that 3 they -- that there was this issue about the 4 installation of a new base unit? 5 A. Sorry, when you say "them", what do you mean? 6 Q. Messages from Mr Castleton about the 7 installation of a new base unit? 8 A. No, that was a message that was displayed on his 9 computer -- 10 Q. Screen? 11 A. -- screen, that he saw, and so that was when he 12 logged a call about it with the Helpdesk, he 13 said that was what he reported was showing on 14 the screen. 15 Q. So why were you, when you looked at these 16 messages, content that they could have had no 17 impact on discrepancies when Mr Castleton was 18 saying it's since the installation of this new 19 base unit that the problem has occurred? 20 A. I'm not sure he was. He was saying it was from 21 the installation BT box? That message that he 22 saw on the screen meant that the computer was 23 not working at all, which is why it had to be 24 replaced. If it's not working, you can't be 25 doing transactions on it. I'm sorry, I'm not 126 1 quite sure where we're trying to go. 2 Q. If we carry on in your witness statement, 3 please, at 14, you say: 4 "Looking at the ... event messages, there 5 are none written on counter 2 between 1 February 6 and 2.24 pm on 2 February. All counters at the 7 branch regularly exchange handshake messages to 8 check they are all up and running; counter 1 9 realised it could not communicate with counter 2 10 and so wrote an event (and also displayed 11 a message to the user) that it was disconnected 12 from the network. I think the message that it 13 was disconnected is slightly confusing; the 14 network was ok (and counter 1 was in 15 communication with the data centre and could do 16 online banking transactions) but counter 2 was 17 effectively dead." 18 What do you mean by the messages produced 19 by, I think, Riposte were slightly confusing? 20 A. Because -- I put this section in because I was 21 questioned about it in 2006 and it was obviously 22 an area of concern to Mr Castleton and, because 23 I hadn't really covered this in my 2004 24 investigation, I thought it was worth trying to 25 spell out here what I can now see. So to go 127 1 back a bit, the -- there was an event that was 2 included in the ARQ event data that said node 1 3 is completely disconnected from the network, and 4 I think Mr Castleton asked me questions about 5 this in 2006 but, in fact, yes, the -- this 6 message was being produced on counter 1 because 7 it couldn't communicate with counter 2, and that 8 was because counter 2 had failed and was being 9 replaced. 10 Q. Can we look, please, at EXPG0000001 and turn to 11 page 174, please. You can see this has just 12 extracted into an expert report that the Inquiry 13 has received a KEL. Can you see it was raised 14 back in 1999 and updated in 2004? 15 A. Yes. 16 Q. The symptoms are: 17 "Riposte error demodify operation failed." 18 Is that I/O or L/O? 19 A. I/O. 20 Q. What is the "I/O completion weight operation"? 21 A. I assume I/O is input/output but this isn't 22 something I have any memory or knowledge of. 23 Q. Then some code is given: 24 "... error occurred within the cache 25 manager. The I/O completion wait operation 128 1 timed out ... The message server will be shut 2 down abnormally." 3 Then "Solution": 4 "Reboot counter. If message appears then 5 send engineer. If a single counter outlet 6 engineer is to replace the mirror disk first and 7 if the message reappears to then replace the 8 base unit ... the base unit is replaced ... 9 Contact the SMC to synchronise the message 10 stores ... See [another KEL] for further details 11 ... On a multi-counter outlet, replace the base 12 unit." 13 A. Yes. 14 Q. Is this KEL essentially explaining that, after 15 hardware replacement, engineers need to make 16 sure that they synchronise machines so that 17 transactions are not lost? 18 A. That's a single counter office. 19 Q. I'm sorry? 20 A. It's -- yes, so it's hard to read because you 21 haven't got it formatted -- it's not formatted 22 very well but it's "If a single counter office", 23 which is SCO, "the engineer must contact the SSC 24 to synchronise the message stores". 25 So the synchronisation is vital if replacing 129 1 the base unit at a single counter office. 2 Q. Do you recall discussing this issue in the 3 run-up to Mr Castleton's trial? 4 A. No, I don't, and he had a multi-counter outlet 5 anyway. 6 Q. Let's look at what was discussed, then. 7 A. I mean, this precise error isn't the error that 8 he was getting anyway. So ... 9 Q. I see. This is a different -- 10 A. This is a different, the probable disk problems 11 is the underlying problem but it was for a very 12 specific Riposte error, which was not seen at 13 Marine Drive. 14 Q. Got it. Let's look at the discussion that did 15 take place before the trial, then. POL00071165. 16 Can you see this is an attendance note of 17 a meeting at which you and a number of other 18 Fujitsu and Post Office witnesses were all 19 present? 20 A. Yes. 21 Q. Can you see that -- 22 A. Yes. 23 Q. -- dated 6 June 2006? 24 A. Yes. 25 Q. Some solicitors from Bond Pearce, a solicitor 130 1 from Hugh James, Brian Pinder, Peter Sewell, 2 Andy -- I think that's Dunks -- 3 A. Yes. 4 Q. -- D-U-N-K-S, rather than "Dunce" -- 5 A. Yes. 6 Q. -- Anne Chambers, Naomi Ellis, Gareth Jenkins -- 7 A. Yes. 8 Q. -- a "distinguished engineer of Fujitsu" -- 9 A. Yes. 10 Q. -- Gareth Ward of the Post Office, yes? 11 A. Yes, it was actually Naomi Elliott, who was my 12 manager's manager. 13 Q. Can we turn to page 6, please. Can we look at 14 the third main paragraph down: 15 "At approximately 3.30 am the PCs do 16 a refresh and occasionally you get glitches in 17 the system and this sometimes causes an 18 initialisation failure. An operational 19 integrity violation points to a refresh not 20 starting up properly. An engineer was sent out 21 to sort the base unit and node. It is worth 22 noting that all messages will be copied from the 23 other node. Potentially could lose some 24 information if the two PC nodes were not 25 communicating at the point of breakdown. There 131 1 is nothing to suggest there was a problem at the 2 end of the previous day. This can be checked by 3 looking at the end of day marker to 1 February." 4 Then "AC". I think, in that context, that 5 must mean you? 6 A. It must. 7 Q. "AC stated that if it is a problem can 8 forensically check the old hard drive but you 9 would need the postmaster to be aware of 10 transactions having been lost." 11 If this had occurred, would it be shown on 12 standard, filtered ARQ data? 13 A. If the transactions were lost, then they 14 wouldn't have got anywhere to be audited. 15 Q. So if the postmaster was not aware at the time, 16 and the synchronisation didn't work, the 17 transactions could be permanently lost? 18 A. Possibly, yes, but you can tell from the ARQ 19 data whether the synchronisation was working. 20 You can certainly tell from the underlying 21 audited messages where the synchronisation was 22 working. 23 Q. What about from the standard, filtered ARQ data? 24 A. You could certainly see whether somebody had 25 logged out, for example, at the end of the 132 1 previous day, which would sort of imply a clear 2 closedown. But -- sorry, I'm just trying to 3 remember. I think it would be hard to tell 4 100 per cent from the ARQ data whether this had 5 happened or not. 6 Q. Were, at this time, the hard drives in the Post 7 Office estate known to be subject of high 8 failure rates? 9 A. I'm not sure about high failure rates but, yes, 10 there certainly were problems. 11 Q. Was it regarded as a significant issue? 12 A. I don't know if it was or not. 13 Q. Was it talked about within the SSC, the number 14 of hard drive failure rates? 15 A. Not that I recall. There -- 16 Q. I'm sorry? 17 A. Yes, I was just going to say there were also 18 times when base units were replaced when, in 19 fact, it might have been possible to sort out 20 the problem in some other way, but it was 21 actually the quickest way to get a -- if, for 22 some reason, a counter was unable to restart and 23 run Riposte, it was actually quicker to get them 24 up and running again by sending an engineer out 25 to exchange the base unit, rather than by 133 1 digging down as much as you could potentially 2 have done into, you know, precisely what the 3 problem was. 4 MR BEER: Thank you, Mrs Chambers. 5 Sir, we're finishing at 4.00 pm today. 6 Might that be a time for the afternoon break 7 until 3.10? 8 SIR WYN WILLIAMS: Yes, that's fine, Mr Beer. 9 Thanks. 10 MR BEER: Thank you very much. 11 (2.55 pm) 12 (A short break) 13 (3.09 pm) 14 MR BEER: Sir, good afternoon, can you see and hear 15 me? 16 SIR WYN WILLIAMS: Yes, I can, thank you. 17 MR BEER: Thank you. 18 Mrs Chambers, in your statement, you deal 19 with the question of what would happen if 20 Mr Castleton was logged on to node 1 at the 21 branch during the time that node 2 was out of 22 action, and then node 2 came back into action, 23 he logged out -- sorry, he logged on to node 2 24 without logging off node 1. 25 A. Yes. 134 1 Q. First of all, can you explain in simple terms 2 what a node is? 3 A. A serving position, a counter, a separate -- two 4 separate computers. 5 Q. So he's initially logged on to node 1 -- 6 A. Yes. 7 Q. -- at a time that node 2, the other counter, is 8 out of action? 9 A. Yes. 10 Q. Node 2 comes back online? 11 A. Yes. 12 Q. He logs on to that node 2 -- 13 A. Yes. 14 Q. -- but without logging off from node 1? 15 A. Yes. 16 Q. You say, cutting through things, the result is 17 that any transactions that he had done on node 1 18 would travel with him for completion on to 19 node 2? 20 A. Any transactions that hadn't been settled on 21 node 1 would be moved over to node 2. 22 Q. Can we look, please, at LCAS0001300. Can we 23 look at page 337 in the bundle, please? 24 If you can go back a page. I think that was 25 it. Yes, I think that's 337, at the top there. 135 1 Can that be expanded, please? Thank you. On 2 this page, can you, in broad terms, without 3 doing it column by column, explain what is shown 4 by this document and where the data comes from? 5 A. Right. This appears to be part of an ARQ 6 extract, what's, I think, usually called the 7 transaction log. 8 Q. So just stopping there, this shows transactions, 9 does it -- 10 A. This shows transactions. 11 Q. -- as opposed to events? 12 A. Yes. 13 Q. Okay. Just in broad terms, tell us what the 14 difference between a transaction and an event 15 is? 16 A. A transaction has a financial value and is -- 17 will subsequently be included in the branch 18 accounts. An event is various things, showing 19 users logging on/off -- 20 Q. Printing? 21 A. Some -- yeah, AP and reports being printed. 22 Q. You said that this is ARQ data represented on 23 this page. Is this or could this be described 24 as raw ARQ data or has something been done to 25 get it into this format? 136 1 A. I -- the underlying data, and I don't know 2 precisely what "ARQ" means, but the underlying 3 archived messages are the raw messages as 4 written on the counter, which get replicated to 5 the data centre and then get funnelled off into 6 files for audit. 7 Q. How is that, I'll call that raw data, translated 8 into a spreadsheet such as this? 9 A. An extract is -- would be run against that raw 10 data, pulling out all transactions in this case, 11 which is where you've got -- I've forgotten what 12 the extract did but it pulls out all lines where 13 you've got a product number and various other 14 attributes. 15 Q. Is that pulling out -- can that be described as 16 parsing? 17 A. I think parsing is more the sense where you've 18 found the records that you like but then, out of 19 those records, you pull out various pieces of 20 information, for example, user, session ID. So 21 the pasting is the breaking up of a message and 22 just pulling out certain pieces of information 23 from that message. 24 Q. Those two parts of the extraction and 25 presentation, respectively, were they done by 137 1 software that had been written in the SSC? 2 A. No. The access to the audited data was -- could 3 only be done by the Security team and they took 4 total responsibility for that and the SSC 5 weren't allowed anywhere near the audit data. 6 Q. So the Security team had to get the data -- the 7 raw data, as I'm calling it -- from where? 8 A. There were very big files that were accumulated 9 day, after day, after day, and then they had 10 tools which SSC did not support, which enabled 11 them to extract the data for the particular 12 branch for a particular date range and then to 13 format it in this way and to split it into event 14 data and transaction data. 15 Q. When it came to you, was it already in this type 16 of format? 17 A. I didn't normally look at data in this format at 18 all. We didn't use this data. Except -- 19 I think the first time probably I ever saw data 20 in this format was for Marine Drive. 21 Q. Was for this case? 22 A. Yes. In 2006. 23 Q. So this first page, then, is looking at 24 transactions extracted from ARQ data? 25 A. Yes. 138 1 Q. Then if we skip over two pages at 339 -- and 2 just if that can be expanded a little bit so you 3 can see -- is that showing the events data? 4 A. That's the events data, yes. 5 Q. Thank you. 6 A. There would also have been messages that didn't 7 fall into either of those categories. 8 Q. Thank you. Looking at 339, first, please. 9 I can't quite see what's happened on the screen 10 there. Thank you. Can we find the line which 11 is 12.41.50. Can you see that? It's almost 12 exactly halfway down the page. 13 A. Sorry, yes. I can see it. 14 Q. I wonder whether that can be highlighted. 15 12.41.50. It's now about eight lines in. Just 16 the one below that. Thank you. Does that show 17 that at 12.41.50 pm, the user, LCA001, logged on 18 to node 1? 19 A. Yes, it does. 20 Q. Can we see the node in the second column from 21 the left? 22 A. That's the node, yes, on which the message was 23 written and that must have been the node to 24 which he logged on. 25 Q. On the same page, can we see at 2.24.58, node 2 139 1 was up and running and LCA001 logged on to that? 2 A. Yes. 3 Q. Can we go back, please, to page 337, and scroll 4 down, please. Can we see between 14.21.12 and 5 14.22.14, before LCA001 had logged on to node 2, 6 those credentials appear against a series of, 7 I think, 14 pension transactions? 8 A. Yes. 9 Q. If we just track those, 14.21.12, right down to 10 14.22.14, so about 14 columns on -- 14.22.14 -- 11 and one more. Thank you very much. 12 So before node 2 was up and running again, 13 at 2.24 pm, it's shown on these transactions, on 14 node 2, as appearing against 14 transactions 15 that are carried out in, I think, a period of 16 about 1 minute, 2 seconds; is that right? 17 A. Yes. 18 Q. Can you see in column F, in the session ID, 19 there is a series of numbers with "-1" after 20 them? 21 A. Yes. 22 Q. Do you understand from that that that is the 23 transactions travelling from node 1? 24 A. I wouldn't have -- well, yes, it's -- the 25 transactions were carried out on node 1. That's 140 1 why the session ID, which is the same for that 2 complete set, has got "1" in the third position 3 of it, after the branch code. So they were 4 carried out on node 1, but then finally at 5 14.22.14, the session -- ah, no. 6 Where's the -- the settlement's further down 7 somewhere, isn't it, I believe? Yes, we haven't 8 got a settlement in there at all. The confusing 9 thing about looking at the data in this format 10 is the time, it's the time that the transaction 11 was started, which isn't necessarily the time at 12 which the session was settled. 13 Q. If the transactions were made here with LCA001's 14 log-in, how could that happen before LCA001 had 15 logged in to node 2? 16 A. Because they were done on counter 1, which is 17 why the session ID and the transaction IDs all 18 have "1" in the third part of them. They were 19 carried out on counter 1 but they were not 20 settled before he logged on to counter 2. So 21 when he log on to counter 2, those transactions, 22 because they were still just on the stack and 23 hadn't been written to message store, so they 24 were transferred across to the second counter. 25 I think this is all explained very clearly in 141 1 John Simpkins' witness statement. 2 Q. I'm asking you for your view at the moment. 3 We'll get to Mr Simpkins, I think, after 4 Christmas. 5 A. Yes. My view, when looking at this, was that 6 yes, they were done on counter 1 but because 7 they weren't settled before Mr Castleton logged 8 on to node 2, they were carried across to node 2 9 and somewhere further down we'll have 10 a transaction with the same session ID which is 11 the settlement for that session. 12 Q. Can we scroll down, please, and see whether we 13 can find that? 14 A. It's that one at 14.28.17. 15 Q. So four down from the top of the page. 16 A. I've lost ... 17 Q. Can you explain what that shows? 18 A. That shows that at 14.28.17 on counter 2 the 19 session that had been started on counter 1 with 20 the session ID 1-899855 was settled on counter 2 21 to cash for £1,350.75. 22 Q. Thank you. That can come down. 23 It's right, isn't it, that each week the 24 subpostmaster would carry forward cash in hand 25 and the stock in hand, so that one set of 142 1 figures, from, say week 1 would feed into the 2 figures for week 2. 3 A. Yes. 4 Q. Can we, please, look at LCAS0000362, and 5 page 29, please. Just forgive me a moment 6 whilst I catch up in my hard copy papers. 7 (Pause) 8 Is this the cash account for week 41? 9 A. It appears to be, yes. 10 Q. We can see that in the top right-hand corner -- 11 A. Yeah. 12 Q. -- in respect of Marine Drive -- 13 A. Yes. 14 Q. -- Bridlington. Can we see that that week ended 15 7 January 2004? 16 A. Yeah. Cash account. 17 Q. Can we go to page 30, please. If we just go 18 back. I was taking the date from underneath 19 "Cash Account (Final)", "Week Ended" underneath 20 it? 21 A. That's the end of the accounting period but this 22 report wasn't actually printed and produced 23 until the -- 24 Q. Until the 8th? 25 A. Until the 8th, in the morning. 143 1 Q. So I wonder whether that can be highlighted 2 underneath the words "cash Account (Final)", the 3 words "Week Ended [7 January 2004]". Thank you. 4 Can we go to page 30, please, and scroll 5 down. Can you see the section of the cash 6 account dealing with cash and stock in hand. 7 A. Yes. 8 Q. Can you see that the cash in hand is £43,757.92? 9 A. Yes. 10 Q. Can you see that the stamps in hand are 11 £5,335.11? 12 A. Yes. 13 Q. So those figures ought to be the same when we 14 look at the transaction data in Horizon, yes, in 15 the ARQ data? 16 A. There wouldn't be -- that wouldn't be held as 17 transaction data in Horizon. 18 Q. It would be held as? 19 A. Opening figures, which aren't included in the 20 ARQ extract. 21 Q. Would they appear as "declare cash" or "declare 22 stamps" before the cash report account? 23 A. You would expect to see those figures or things 24 that added up to those figures in the event 25 data, yes. 144 1 Q. In the event data. Thank you very much. Can we 2 try and look for that, then, please, remembering 3 those figures: 43,757 and 5,335, yes? 4 A. Yes. 5 Q. Can we look, please, at LCAS0000224. I wonder 6 whether that can be expanded, so we can see 7 a little more of it, please. Thank you. Can 8 you help us as to what these data are? 9 A. This is an extract from the ARQ event data. It 10 looks like it might not be all the -- 11 Q. I missed the end of the sentence? 12 A. It's some lines, I think, from ARQ event data. 13 It's not a complete set of ARQ event data. It's 14 a subset. 15 Q. If we look for 8 January so if we scroll down, 16 please, and for 8 January 2004, can we see the 17 declare cash figure at 7.01.52 am? 18 A. Yes. 19 Q. The declare cash figure is £43,757.92? 20 A. Yes. 21 Q. The same figure? 22 A. I didn't write down the figure that we were 23 looking at before. 24 Q. Yes. 25 A. Yes. 145 1 Q. It's the same figure, yes. 2 A. Okay. 3 Q. That appears twice, declare cash total twice. 4 can you see that? 5 A. I can, yes. 6 Q. It doesn't matter for present purposes why that 7 is. Then a little while later, at 7.49.24, it 8 says, "Report Cash Account, Report Printed", 9 yes? 10 A. Yes. Sorry I'm rather puzzled by that data not 11 being in date order but never mind. 12 Q. Ie not in chronological time order? 13 A. Not in time order. The ONCH total above is 14 obviously from later that day so this data has 15 been manipulated in some ways. Sorry, I don't 16 mean that in a bad way but I'm just pointing it 17 out. But, yes -- 18 Q. But these should be chronological time order? 19 A. They would have been when it -- I would have 20 thought they would have been when extracted for 21 the ARQ event data but, yes. So you've got 22 a cash account at 7.49. 23 Q. So would that have been the printing of the 24 document that we just looked at? 25 A. I assume so, yes. 146 1 Q. So all okay, so far? 2 A. Yes. 3 Q. Can we look, then, at the declare stamp total 4 figures on 7 January at 17.06 and 17.41. If we 5 just scroll up a little. That's it, we've got 6 it, 17.06. 7 A. Yes, and then -- 8 Q. Can you see "Declare stamp total, £1,183.22"? 9 A. Yes, I can. Again, they're not in date order, 10 in time order. 11 Q. There's a previous one -- 12 A. Yeah. 13 Q. -- 35 minutes later, printed before -- 14 A. Yes. 15 Q. -- namely 7.41.30, "Declare stamp total, 0". 16 A. Yeah. 17 Q. There aren't any other declare stamp total 18 figures on this page, are there, before the 19 report cash account entries the following 20 morning? 21 A. Doesn't look like it, no. 22 Q. Those two figures, 1,183.02 and then 0, at 17.46 23 and 17.41, neither of those match the "5,335.11 24 figure for stamps in hand, as it appears in the 25 final cash account, do they? 147 1 A. They don't appear to, no. Could we just go up 2 slightly further to the previous cash account? 3 Q. Yes. 4 A. Mm. Yes. No, I was just -- from this data, no, 5 I cannot explain that. One thing is that you 6 could make stamps and cash declarations with 7 different declaration IDs and then they got 8 added together, so the 0 wouldn't necessarily 9 wipe out the 1,183.22 but I don't know why what 10 is on that cash account -- 11 Q. Is £5,335.11 for stamps? 12 A. Unless there was another declaration using 13 a different ID again, since the previous cash 14 account. But I can't -- 15 Q. It's not shown on here, is it? 16 A. It's not showing on there, no. 17 Q. So the value of stamps had somehow reduced, 18 either to zero or £1,183.02 -- 19 A. Yes, I can't explain that, looking at this. 20 Q. -- on these ARQ data, and that has not been 21 translated into the cash account for week 41, 22 has it? 23 A. That appears to be something that, at the 24 moment, I cannot explain, and could be an error. 25 Q. That would have disrupted the figures for week 148 1 42, as well, wouldn't it? 2 A. What was -- that might have given them a loss in 3 week 41, but it shouldn't have then had any -- 4 I wouldn't have thought it should have had any 5 knock-on effect but without looking at the data 6 and exactly what happened, I cannot tell. Did 7 they have a loss in -- 8 Q. Yes, this is the first week. The first week, 9 week 41, that Mr Castleton phones in and says, 10 "I've suffered a loss, an unexplained loss". 11 A. Yes, I was unaware of this. 12 Q. Did you look at these data? 13 A. I thought that in the run-up to the 2006 trial 14 I looked in great detail at all the data from 15 January and February but I don't recall seeing 16 this particular issue, whether I looked at it 17 and could explain it, I now don't know. As I'm 18 saying, obviously, there's something slightly 19 odd about this particular dataset, in that 20 it's -- it is an extract, and it does seem to be 21 a slightly odd order. 22 Q. Just go back to the previous page, please. 23 Sorry, the top of this page. Thank you. Is 24 this -- are these data produced by Fujitsu? Can 25 you tell from the character strings set up? 149 1 A. Yes, this is data that originally was produced 2 by Fujitsu, and you can see that you've got the 3 sort order there, so it should be -- attribute, 4 date and time is what it should be sorted on. 5 Now, just looking at it, you can see it is no 6 longer sorted on date and time. So I'm 7 wondering -- and, also, as I said, it's not 8 a complete set of event data in that there would 9 have been many other things -- you've not got 10 anybody logging on and off, for example, in 11 that. 12 So this is -- what was provided by Fujitsu 13 would have been a much bigger Excel spreadsheet, 14 and this is a subset, not a -- 15 Q. Sorry, somebody coughed there, I couldn't hear 16 that. This was? 17 A. This appears to be a subset of that data and 18 that subsetting would have been done by somebody 19 other than Fujitsu. And it's also been sorted 20 into an order which is not the order that it 21 would have been provided by Fujitsu. 22 Q. Do you recognise the format of the presentation 23 of these data? 24 A. Oh, yes, it's in the format that you would 25 expect, and those lines at the top are what you 150 1 would expect to see at the top of the ARQ event 2 data. But the content is not -- I mean, we've 3 got the full set of events for this period 4 somewhere in the evidence set and we could go 5 and look at that and see what was actually 6 presented by Fujitsu. 7 Q. That might be homework overnight. 8 A. I think it might be. 9 Q. But, on the face of it at the moment, these data 10 don't -- do not make sense to you, in that the 11 value of stamps has been reduced dramatically on 12 the ARQ data but not in the cash account for 13 week 41? 14 A. Yes. 15 Q. Can we look, please, at POL00112440 and look at 16 page 52, please. This is the first page of 17 a PEAK in relation to an entirely separate 18 branch. 19 A. Yes. 20 Q. You can see the FAD code there under the 21 summary. 22 A. Yes. 23 Q. It's not Mr Castleton's branch. 24 A. Yeah. 25 Q. So the summary is "Discrepancy with [cash 151 1 account] for the last few weeks". Can we read, 2 please, the first few entries under the 3 narrative: 4 "Call details have been taken from Andrew 5 Wise at NBSC on telephone number stated above. 6 [Postmaster] has a discrepancy with his cash 7 account for the last few weeks." 8 So this is 13 January, so it's four days 9 before Mr -- four or five days -- no, it's the 10 day before Mr Castleton makes his first report, 11 yes? 12 A. Yeah, yes. 13 Q. That was on 14 January? 14 A. Yes. 15 Q. "Postmaster has a discrepancy with his cash 16 account for the last few weeks." 17 A. Yeah. 18 Q. "The NBSC have advised that the [postmaster] can 19 only declare the holding amount or 0 not 20 a negative figure." 21 Then reading on: 22 "The NBSC have been through the checks with 23 the [postmaster] and feel there is a software 24 error as the [postmaster] has negative figures 25 which he would not have been able to enter." 152 1 A. Yes. 2 Q. "In week 37 the [postmaster] was showing 3 a surplus of £19.66 on the front page of the 4 cash account. 5 "In table 5 cash and stock he is showing 6 a surplus cash figure of £42,860.61 and 7 a surplus cheque figure of £116,248.50." 8 A. Yes. 9 Q. "The cash on hand was declared at £159,109.11. 10 In week 38 the [postmaster] had a surplus of 11 £49.22 on the front page of the cash account." 12 Yes? 13 A. Yes. 14 Q. Then, if we carry on reading, please. Just 15 scroll down. If you read to yourself, so 16 I don't have to read the figures out, what's 17 recorded in the second box from the bottom 18 there? 19 A. Mm-hm. 20 Q. Then over the page, please. Read the top box, 21 please. 22 A. Yes. 23 Q. Then scroll down, please. We see at 16.01 on 24 13 January, this call is assigned to you? 25 A. Yes. 153 1 Q. What did "ESC" mean? 2 A. Sorry? 3 Q. It says, "Assigning call to Anne Chambers in 4 ESC"; is that the SSC? 5 A. SSC or EDSC, which we were sometimes known as. 6 I don't know why -- 7 Q. Okay. It's not a different capacity you were 8 being assigned this call in? 9 A. No. 10 Q. You speak to the postmaster the next day. Can 11 we see that entry at the bottom? 12 A. Yes. 13 Q. "Spoke to [postmaster] to get some more info. 14 Before he balances he rems out his cheques, 15 although sometimes one or two more cheques are 16 taken before he balances the stock unit. He 17 declares cash then declares stamps then finally 18 declares stock. On the first declare stock line 19 he declares any rest home cheques he is 20 holding." 21 Can you help us with what "rest home 22 cheques" mean. 23 A. I had no idea when I picked up this call but it 24 was some sort of a cheque that he received that 25 I think then pensions for everybody in the rest 154 1 home got -- 2 Q. So like a care home? 3 A. Yeah. 4 Q. Okay, that's not a technical description, it's 5 a reference to something like a care home? 6 A. That was how the postmaster described it to me. 7 Q. "This results in a discrepancy between the 8 system check figure and the declared figure." 9 A. Yeah. 10 Q. "Something has changed in the counter code 11 recently which causes the discrepancy to be 12 recorded wrongly. So the cheque discrepancy 13 instead of being cleared is doubled and the cash 14 is wrongly adjusted." 15 A. Yeah. 16 Q. "However, I think the [postmaster] should not be 17 declaring his rest home cheques in this way, 18 even if he's always done so. Need to get advice 19 on this or preferably get NBSC to explain what 20 the proper procedure is." 21 A. Yeah. 22 Q. Was this a postmaster declaring these care home 23 cheques in a slightly unusual way but he'd been 24 doing it for years, yes? 25 A. Yes, he was doing something odd and -- but it 155 1 had never shown up as a problem on the system 2 but, because something had changed and suddenly 3 he started getting these very big numbers in 4 both cash and cheques showing in table 5 of the 5 cash account. 6 Q. Something had changed in January to start -- 7 A. Yeah. 8 Q. -- causing these problems -- 9 A. Yeah. 10 Q. -- with, essentially doubling up, hadn't it? 11 A. It wasn't quite doubling up but it was the way 12 that, if you rolled over with having declared 13 that you held a cheque, when it was treated with 14 the wrong sign, and then impacted the cash as 15 well. 16 Q. We see later in the PEAK -- I'm not going to go 17 to the detail, that the problem had been caused 18 by a Horizon fix that had been released in 19 November 2003; do you remember? 20 A. Yeah. 21 Q. What was the Horizon fix that had been released 22 in 2003, November? 23 A. I cannot remember -- no, I'm sorry, I can't 24 remember. 25 Q. Okay, but it was the release of this fix by 156 1 Fujitsu that had caused what had previously not 2 been a problem to become a problem for this 3 postmaster -- 4 A. Yes. 5 Q. -- and to cause discrepancies? 6 A. Yes. 7 Q. So it was the release of new software, the fix, 8 that was causing the problem? 9 A. Yes. 10 Q. Do you remember that this PEAK was signed off, 11 saying that the problem was the subpostmaster 12 not using the correct or standard procedure? 13 A. I don't recall that that was my conclusion on 14 the PEAK. 15 Q. Can we scroll down then, please. I think we can 16 see a continuation of your entry at the top of 17 the page there, yes? 18 A. Yes. 19 Q. On 15 January? 20 A. Yes. 21 Q. Then, if we carry on scrolling down, please. 22 A. "At that point I spoke to the PM and explained 23 it was a software problem." 24 Q. Yes. Later on the 15th, you email Julie Welsh, 25 who will pass it on to the Post Office, and you 157 1 asked the EPOSS Development to have a look at 2 it? 3 A. Yeah. 4 Q. Why were you asking the Development team 5 responsible for EPOSS to have a look at this? 6 A. Because EPOSS were -- that was the counter 7 software and it was a bug in their software. 8 Q. So it was a -- can I describe it this way, 9 a software bug in EPOSS -- 10 A. Yeah. 11 Q. -- that was the root cause? 12 A. Yeah. 13 Q. You describe it as a nasty problem? 14 A. Yes, it was nasty in that every week things got 15 worse and worse and worse for a branch that was 16 encountering this problem. 17 Q. Then scroll down, please. Some more process 18 entries on the rest of that page. Then, at the 19 foot of the page, "Fix for 97081 was to blame". 20 That's the November '03 fix, yes? 21 A. Sorry, that was the? 22 Q. That was the November 2003 fix -- 23 A. Yes. 24 Q. -- that was to blame? 25 A. Yeah. 158 1 Q. "Initialisation of the session, effect variable, 2 with [and then some code] was not initialised. 3 Alas, this will affect products such as cheque." 4 Yes? 5 A. Yes. 6 Q. Then scroll down, please. Some more process 7 entries, yes? 8 A. Yes. 9 Q. Forward to the 19th, scroll down, please. You 10 speak to the postmaster on the 19th. Is that 11 you? 12 A. No. 13 Q. That's somebody else, is it? 14 A. The call has been closed for some reason and 15 then reopened and so everything that was on it 16 before appears again. 17 Q. Okay. Oh, I see, yes. If we scroll down, 18 please, and again, and again, and again, and 19 again, and again. Just stop there. 20 Do you see at the top of the page an entry 21 is made: 22 "Phil needed this information for 23 a disciplinary with the [postmaster] tomorrow. 24 Advised Phil that this is basically user error 25 and 3rd line had advised the [postmaster] of 159 1 correct procedures." 2 A. Yes, that's not an update that I put on there 3 you can see, that's something that's been sent 4 across from PowerHelp. You can see HSH -- HSH1 5 is a PowerHelp user. So somebody on the 6 Helpdesk has looked at what I'd put previously 7 and has interpreted it, saying it was basically 8 user error, which was not my conclusion. 9 Q. It wasn't correct, in your view, was it? 10 A. No, it's not correct, no. 11 Q. It was a bug? 12 A. Yes, it was a bug. 13 Q. You hadn't advised third line of the correct 14 procedure? 15 A. I was third line. 16 Q. Yes. Sorry, you, third line, had not advised 17 the postmaster of the correct procedure? 18 A. I had spoken to the postmaster and I believed to 19 an auditor who was in the branch when I phoned, 20 and explained that I thought this was something 21 odd that they were doing but I had also said 22 that it was a software error. 23 Q. The further fix to deal with the bug that had 24 been introduced by the November 2003 fix was to 25 be part of the S60 release -- do you remember 160 1 that -- 2 A. Yes. 3 Q. -- which didn't proceed until July 2004; do you 4 remember that? 5 A. I can't remember the dates. 6 Q. Okay I'll give a cross-reference for that. 7 There's no need to search for it, it's 8 FUJ00155590 at page 4, with S60 being authorised 9 to proceed in July 2004. 10 A. Okay. 11 Q. When you came to Mr Castleton's case, do you 12 think that there was any coincidence of timing 13 or issues here, given that Mr Castleton 14 complained of difficulties with cheques? 15 A. I looked at Mr Castleton's cheque handling and 16 he certainly was not declaring cheques in the 17 same way that this branch had been doing. He 18 didn't use declare stock, he used adjust stock, 19 and it was very obvious for this particular 20 instance, which was the only one I ever saw with 21 this particular problem, because on your balance 22 account, the balance report or, indeed, in the 23 cash account, it gave extremely large numbers 24 increasing week-by-week for both cash and 25 cheques in the method of payment section on the 161 1 balance reports and in table 5 on the cash 2 account report. And they were not present for 3 Mr Castleton's branch. 4 Q. Did you speak with the -- anyone at Marine Drive 5 to discuss their cheques practices? 6 A. No, I didn't, because I -- they'd already talked 7 to the NBSC and to the Helpdesk about that, and 8 the only problem with -- that I could see with 9 their cheques was the one day where they hadn't 10 cut off the report, which wasn't an uncommon 11 problem. It was -- there were so many things 12 they had to do, it's not surprising that 13 branches very occasionally failed to do one of 14 the things that they were supposed to. 15 Q. What would you have thought if you had spoken to 16 the branch and they told you that cheques often 17 seemed to clear to cash? 18 A. I'd have gotten them to give me some examples. 19 So you're saying that when they received 20 a cheque in payment then it showed up on the 21 system as cash instead? 22 Q. Yes, and that they got to the stage where this 23 had happened more than once and that 24 Mr Castleton and the clerk would watch the other 25 do it, so that they had two pairs of eyes 162 1 watching what was happening when they had 2 a cheque transaction to complete? 3 A. So they were certain that they were pressing the 4 cheque button -- 5 Q. Correct. 6 A. -- that it was actually being treated as cash? 7 Q. Yes. 8 A. I would have advised them to get the screen 9 calibration checked. 10 Q. Why would the screen calibration have raised its 11 head with you? 12 A. Because if the buttons on the screen -- if they 13 were using the buttons on the screen -- were -- 14 if the calibration was slightly out, you could 15 press one button and it would have a slightly 16 different effect. It shouldn't have been that 17 far out and that, in fact, could have given all 18 sorts of problems. But that was part of the 19 standard Helpdesk stuff, I think, so that if -- 20 you know, if they said, "But we're pressing this 21 button and something else happens", then you 22 recalibrate the screen. 23 Q. You wouldn't have been able to tell any of that 24 from just looking at the Horizon data? 25 A. No, to me, if it's settled to cash, I would 163 1 assume that they had either pressed the fast 2 cash button or that -- the cash button, 3 I believe they were two different ones. 4 Q. Was that a problem sometimes, calibration of the 5 screen, so that where you pressed a tile on 6 the -- I'm calling it a tile, that might not be 7 the correct expression -- 8 A. Yeah. 9 Q. -- you know what I mean, an icon -- 10 A. Yeah. 11 Q. -- on the screen, because of a miscalibration 12 the function performed was not what was 13 intended? 14 A. That could be a problem but it would be pretty 15 consistent, so for everything you did it would 16 be out and I would have thought, although 17 I never worked in a branch, that it would be 18 fairly obvious fairly quickly to the user if 19 that were happening. 20 Q. But cheques clearing to cash may have fed into 21 what you saw when you examined the data, that 22 there were large discrepancies between daily 23 cash declarations and what the system expected 24 the cash to be? 25 A. No, because at the end of each day when they had 164 1 to go through their physical checks, they'd 2 print off the cheque report, which would show 3 the cheques that had been recorded for payments 4 on the system. They then had to compare that 5 with the physical cheques that they were holding 6 and if there were any differences. So if they 7 had a cheque and it went on the list of cheques, 8 they could do an adjustment which would adjust 9 between cash and cheques. So there would be no 10 loss or gain as a result of doing that. 11 Q. What if they did that at the end of the day, 12 where they could see that the cheque 13 transactions hadn't cleared to cheques, they'd 14 cleared to cash, what would happen to the 15 earlier misapplication of the cheque funds? 16 A. Nothing would happen to that, so you'd have your 17 earlier transaction, which would be, say, a bill 18 payment settled to cash, so you'd have -- those 19 lines would remain unchanged because once 20 they're written, they stay there. And then 21 further down at some point, you would see -- 22 I mean, there were various ways that the 23 postmaster could do it but one way would be 24 doings a stock adjustment. 25 So you'd see a line then with "mode SAP", 165 1 stock adjust positive, so your cheque going up 2 by whatever the amount was and the cash then 3 would come down by the same amount. 4 Q. Thank you very much. 5 A. There are examples of that in a very large 6 spreadsheet, which is available to everybody. 7 MR BEER: Thank you. 8 Sir, I don't know whether that's 9 a convenient moment for you. We've just reached 10 before 4.00 pm and it's a convenient moment in 11 the narrative. 12 SIR WYN WILLIAMS: Yes, of course. Thank you, 13 Mr Beer. So we'll resume again at 10.00 14 tomorrow. 15 Mrs Chambers, I got the impression that 16 you'd intended to do some homework, as it was 17 described, in respect of one line of questioning 18 this evening. Would you need to speak to anyone 19 about that or is that you just looking at the 20 documents, so to speak? 21 A. That's me attempting to look at documents on my 22 iPad, which may not be very easy. 23 SIR WYN WILLIAMS: Okay. I only ask because you 24 will appreciate that, unless I give you 25 permission to do so, you shouldn't speak to 166 1 anyone about your evidence. 2 A. I won't speak to anyone. 3 SIR WYN WILLIAMS: I take it you don't need my 4 permission, if you decide to do some homework. 5 A. Thank you. 6 MR BEER: Thank you, sir. We've set it for 7 ourselves, too -- or I've set it for Mr Blake! 8 SIR WYN WILLIAMS: So there will be competing 9 homeworks this evening. Very good. 10 MR BEER: Yes, thank you very much. 11 SIR WYN WILLIAMS: See you in the morning to hear 12 the result of it. 13 (4.00 pm) 14 (The hearing adjourned until 10.00 am 15 the following day) 16 17 18 19 20 21 22 23 24 25 167 1 I N D E X 2 Statement by MR BEER re Disclosure ............1 3 ANNE OLIVIA CHAMBERS (re-affirmed) ...........12 4 Questioned by MR BEER ........................12 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 168