1 Wednesday, 3 May 2023 2 (10.00 am) 3 MR BEER: Good morning sir, can you see and hear me? 4 SIR WYN WILLIAMS: I can hear you but I can't see 5 you. Now I can. 6 ANNE OLIVIA CHAMBERS (continued) 7 Questioned by MR BEER (continued) 8 MR BEER: Thank you very much. 9 Good morning, Mrs Chambers. Yesterday, you 10 gave evidence in the afternoon that ARQ data did 11 not consist of all keystrokes made by 12 a subpostmaster in a branch; do you recall? 13 A. Yes. 14 Q. You said that data was never captured. 15 A. Yes. 16 Q. That was page 160, lines 10 to 13 of the 17 transcript. Before I ask you some questions 18 about that, just by way of background for the 19 Core Participants, in the closing submissions 20 that the Post Office made in the GLO, the Group 21 Litigation Order trial, the Post Office said 22 this in their written submissions: 23 "Fujitsu does have access to a record of 24 keystroke activity. The record must be 25 downloaded from the counter rather than being 1 1 contained in the events log. The counter log is 2 called the POC log ..." 3 In his Horizon Issues judgment, Mr Justice 4 Fraser found that the audit data, as he 5 described it, recorded all keystrokes performed 6 in a branch by a subpostmaster. For those that 7 are following that, paragraphs 906, 911(1), 8 911(6) and paragraph 995 of his judgment. 9 That conclusion was reflected in 10 paragraph 15 of the Court of Appeal Criminal 11 Division's judgment in the Hamilton case, 12 paragraph 15, stating: 13 "Fujitsu held data called ARQ data which 14 contained a complete and accurate record of all 15 keystrokes made by a subpostmaster or 16 an assistant when using Horizon." 17 I want to look at some of the documents that 18 are referred to in support of what the Post 19 Office said in the course of the trial, in those 20 submissions. 21 Can we start, please, with POL00003233. 22 Thank you very much. Now, if we just look at 23 the top line of this, we can see this is a PEAK 24 273234. We can see under the progress narrative 25 that it was raised on 21 August 2018 so that's 2 1 two years after you left and then, if you see 2 the summary, it says the incident is a "Failed 3 Drop & Go Top Up", that doesn't matter for 4 present purposes. 5 A. Ct. 6 Q. If we go to page 3 of the PEAK, if we just, 7 sorry, look at the bottom of page 2 first so we 8 can see who is writing. It is Joe Harrison. 9 Did Mr Harrison work with you when you were in 10 the SSC? 11 A. Yes, he did, yes. 12 Q. Thank you. Was he a diagnostician too? 13 A. Yes, he was. 14 Q. He says: 15 "This is an instance of [and then he quotes 16 a KEL]. Counter 2 did receive the [and then 17 a message number] unsuccessful message but 18 debited the customer [then he quotes some text] 19 to the amount of [£30] anyway. 20 "As stated in the KEL 'This may be an issue 21 with script ... or a user error. The Drop&Go 22 scripts are supplied and maintained by ATOS. 23 Therefore please route calls to ATOS." 24 Then he says this: 25 "Here are the keystrokes and messages from 3 1 the counter, which might help ATOS." 2 Then if we scroll down and look at those, we 3 can see that he has seemingly cut and pasted 4 into the PEAK a series of text and can we look 5 at the cut and paste that has occurred. 6 A. Yes. 7 Q. Can you see the first entry and then the 8 majority of the remaining entries -- 9 A. Yes. 10 Q. -- refer to a button? 11 A. Yes, this when the postmaster touches the button 12 on the screen and that moves it then -- well, 13 you can just see that the button has been 14 pressed as it moves on to another screen. 15 Q. When you're talking about the button being 16 pressed, what are you referring to? 17 A. A virtual -- have you had the chance to see 18 a post office counter screen -- 19 Q. Yes, so a tile on the touch pad? 20 A. A tile on the touch pad, yes. 21 Q. So where it refers to "button", is that 22 referring, is it, to tapping a tile on the touch 23 pad? 24 A. Yes, or using the associated key on the 25 keyboard, so, to that extent, yes, the button 4 1 presses or the virtual button presses are 2 recorded but not every single keystroke. So we 3 can't see here that a name has been typed in or 4 that, you know, perhaps a name was typed in and 5 then deleted or anything like that. So perhaps 6 I misunderstood yesterday but I still say that 7 not every keystroke is recorded. But for HNG-X, 8 not for Legacy Horizon, we did explicitly ask 9 for this extra level of diagnostics which helped 10 us to see how the user was navigating the system 11 at any point. 12 Q. So breaking that down, you remember the example 13 you used yesterday of the £250 versus the 14 £2,500? 15 A. Yes. 16 Q. If the postmaster -- I think this was showing 17 some cash in? 18 A. Yeah. 19 Q. If the postmaster wanted to show that they had 20 received £250 in, and the system in the event 21 showed two thousand £500 in, would you be able 22 to tell, starting with Horizon Online that the 23 postmaster pressed a 2, a 5 and an 0 and then 24 Return or commit it to the stack, or whatever 25 the button was to be pressed, rather than 5 1 pressing a 2, a 5, and then 00? 2 A. You couldn't see that level of detail. 3 Obviously that information, whichever it was, 4 would be captured and then stored on the system 5 when the transaction was committed but then when 6 it's on the system, that would be the number 7 that I am seeing. So I wouldn't be able to tell 8 that, at the point it's actually being recorded 9 by the system, it is not precisely what the 10 postmaster had keyed. 11 Q. So if he said, "This system is showing that 12 I was showing a receipt of cash of £2,500, I did 13 not press a 2, a 5 and then 00, I only pressed 14 a 2, a 5 and then a single 0", you wouldn't be 15 able to tell from the keystroke data whether 16 that was accurate or not? 17 A. No. 18 Q. All that would say is that the system shows that 19 you pressed 2, 5 and 00 -- 20 A. Yes. 21 Q. -- because £2,500 is shown as cash coming in? 22 A. And that would also then be shown on the screen 23 to the postmaster, so if he felt that the number 24 was wrong, either because the system was now 25 displaying it to him wrongly or because he'd 6 1 miskeyed it, then you would expect that to be 2 corrected at that point. 3 Q. Then breaking it down a little further there, 4 you said that not all keystrokes were auditable. 5 A. Yes. 6 Q. What was the dividing line between those which 7 were and those which were not -- 8 A. Um -- 9 Q. -- ie what level of button was auditable? 10 That's a very imprecise question but I think you 11 know what I mean. 12 A. Yes, I think any of the buttons that controlled 13 the navigation around the system or where the 14 postmaster -- I mean, you can see the examples 15 on here, where the postmaster was given a choice 16 and had to choose "Yes" or "No". When they were 17 on the home screen and decided to go into 18 a particular area of code, um, that's -- 19 Q. Trying to go into a particular area of code? 20 A. Sorry, yes, when they chose -- sorry, that's 21 inaccurate. When they selected a particular 22 function, for example, Postage or Bill Payment, 23 other things would also be recorded early in the 24 process. If they scanned a barcode, that 25 barcode that had been read would be included in 7 1 the logs. 2 Q. So you could see the order of events -- 3 A. Yes. 4 Q. -- is that right? 5 A. Yes. 6 Q. You could see the Pathway that a subpostmaster 7 took? 8 A. For HNG-X, yes. And this was really useful for 9 us, for diagnostic purposes, because we were 10 able to see, you know -- we'll see that when we 11 look at some of the specific examples, but we 12 can see, yes, they started to do something and 13 then they used a particular button to move out 14 of it. Perhaps that's not something that would 15 normally happen but that doesn't mean it's 16 wrong. 17 And so that was very helpful to us for 18 diagnosing these problems because we could see 19 the perhaps less expected paths that were being 20 taken. 21 Q. You see four entries in, timed at 13.11.31, 22 there is the word or the character string, 23 "MSG10800: Check Parcels and Services Required". 24 Is that a record of a screen being displayed to 25 the subpostmaster, essentially a pop-up that was 8 1 displayed on the screen to the subpostmaster? 2 A. I think it was a question that he was asked at 3 that point by a message on the screen. I can't 4 remember exactly how it was would have been 5 displayed. 6 Q. What other ways of displaying it other than on 7 the screen were there? 8 A. Sorry, that was the only way. 9 Q. So where we see the "MSGs" on here, the 10 messages, is that a record -- I'm calling them 11 pop-ups, but essentially messages displayed on 12 a screen to a subpostmaster? 13 A. Yes. 14 Q. Again, the same question: were there some such 15 messages that were displayed and not auditable 16 or were they all auditable in this way? 17 A. I can't remember. I think they were all 18 displayed but I'm not certain. They were not 19 there for audit purposes, if you like; they were 20 there as a diagnostic aid, as I said. 21 Q. What's the difference? 22 A. Audit, I would feel is something that you would 23 return to later and say, "This is precisely what 24 happened and we have captured everything that 25 has happened". This data, although I think it 9 1 probably does meet those criteria, it wasn't 2 designed with that in mind. 3 Q. How do you know it wasn't designed with that in 4 mind? 5 A. Because I and a colleague asked the development 6 team when we had a meeting before, when HNG-X 7 was being developed, and we said "Gosh, it would 8 be really useful for us if we know what buttons 9 were pressed and we know what messages were 10 displayed". 11 Q. Was any of this available to you for Legacy 12 Horizon? 13 A. Not in the same form at all, no. You could get 14 clues from the messages in the message store but 15 it wasn't designed -- and there was a certain 16 amount of audit -- of diagnostic information, 17 I think we discovered yesterday in the audit 18 file, but this was very helpful at this level. 19 Q. This PEAK doesn't refer to -- Mr Harrison 20 doesn't refer to what he has cut in to the PEAK 21 as a POC log. Is this in fact an extract from 22 a POC log? 23 A. As far as I remember, yes. I don't clearly 24 remember all the filenames. 25 Q. What was a POC log? 10 1 A. A Post Office Counter log, a file that 2 I think -- I think there was one for each day, 3 I think they were possibly kept for a limited 4 time. It might only have been seven days, 5 I can't remember. They weren't copied off the 6 counter unless we needed to access them but they 7 were there for diagnostic purposes. 8 Q. You said they were only kept, you think, for 9 seven days. Do you mean kept on the counter for 10 seven days but available in an archive after 11 that time? 12 A. No, they weren't archived. 13 Q. They were not archived? 14 A. No, I can't now remember if it was seven days 15 they were kept on the counter or if it was 16 a longer period but it wasn't a very long time 17 and they were not taken off the counter and 18 stored anywhere else, unless somebody in SSC 19 went to get one for diagnostic purposes. 20 Q. Here Mr Harrison is cutting this in to the PEAK 21 on 21 August, if we just scroll up. 22 A. Yeah. 23 Q. Yes. 24 A. Yes. 25 Q. He's referring to events that happened on 11 1 31 July, so three weeks earlier? 2 A. So it was more likely then it was a month's 3 worth but I cannot clearly remember, I'm sorry. 4 Q. Was that by design, that they were only kept for 5 a limited period? 6 A. I imagine so because, obviously, they could get 7 fairly big and you didn't want to fill up the 8 counter file store more than you had to. 9 Q. Were they only available from the counter store 10 and nowhere else? 11 A. Yes. They -- yes, they weren't kept anywhere 12 else. Obviously, if somebody had gone -- had 13 looked at the same branch two weeks earlier and 14 happened to have extracted that log, they might 15 have it, but they would only be there -- be 16 anywhere else if somebody in SSC had 17 specifically extracted them. That was the case 18 certainly when I left in 2016. I can't say what 19 might have happened since then, of course. 20 Q. Can we look, secondly -- that can come down, 21 thank you -- at POL00001835. Thank you. 22 This is a PEAK 209755. Earlier in time, 23 you'll see that it was opened by Mr Parker on 24 15 April 2011. The summary doesn't tell us 25 anything, six lines in from the top, it just 12 1 gives the branch FAD code. If we look at the 2 second entry, if we scroll down, we can see what 3 the issue was: 4 "POL has a discrepancy with a postmaster 5 regarding a transaction in Huddersfield for 6 TPoS." 7 What was TPoS? 8 A. I don't know. 9 Q. "The branch thought that they [were] settling 10 the transaction below to debit card but it has 11 been 'automatically settled to cash'. Branch 12 thinks that something went wrong with their pin 13 pad -- debit card [transaction] declined but the 14 branch didn't notice." 15 Then some details for the branch are set 16 out: 17 "This was not noticed until the next day 18 when they balanced and they then pulled off 19 a transaction log and noticed the cash payment. 20 A TfS call for this was logged on the day after 21 the transaction ... and NBSC and HSD both told 22 the PM that it was user error. 23 "It has now been raised again via TPoS 24 introduction managers -- Fujitsu release 25 managers, etc. To provide a sanity check please 13 1 retrieve the counter log for node 7 on this date 2 and see if we can add anything?" 3 In short, an issue had been identified with 4 a subpostmaster trying to settle a transaction 5 to a debit card but it had automatically settled 6 to cash -- 7 A. Yes. 8 Q. -- and that was only noticed the following 9 day -- 10 A. Yes. 11 Q. -- when the subpostmaster tried to balance and 12 saw that the matter had been settled to cash? 13 A. Yes. 14 Q. Then if we can scroll down to Mr Allen's entry, 15 Dave Allen at the foot of the page there. Was 16 he a colleague of yours -- 17 A. Yes, he was a colleague. 18 Q. -- doing the same work as you? 19 A. Yes. 20 Q. He says: 21 "Immediately after selecting 'Sell Euros' 22 [message] 'Transaction Prompt' appears; this 23 states 'Transactions paid for using a debit or 24 credit card will require mandatory ID'. 25 "I note this isn't shown in the POC log for 14 1 the Huddersfield incident. 2 "Subsequently, the Clerk selected Method of 3 Payment -- 'Debit Card', whereupon [a message] 4 requests entry of the first 4 digits of the 5 card's PAN (the 'Debit Card Prefix'). 6 "After entering the debit card prefix, 7 [another message] 'Clerk Instructions' appears; 8 this states 'Do you wish to flag this 9 transaction as suspicious for anti-money 10 laundering purposes? If you select "Yes", you 11 must also complete [the] form [and a number is 12 given]' -- the [postmaster] answered 'No' to 13 this. 14 "After entering the Customer's name and ID 15 (passport) details, the Clerk is returned to the 16 home screen which shows the 'Total Due from 17 Customer' = £500.00 -- as would be expected. 18 "At this point there is nothing to stop the 19 Clerk settling to Fast Cash, even though 'Debit 20 Card' has been selected earlier in the 21 dialogue." 22 Then this: 23 "The POC log confirms that 'Fast Cash' was 24 indeed selected at this point. 25 "There is no evidence in the POC log of any 15 1 PIN pad interaction at any time during this 2 session and no evidence of any banking dialogue 3 in the counter message log, and no evidence of 4 the session being settled 'automatically' in 5 some way, rather than by action of the Clerk. 6 "The counter logs can't show us whether or 7 not the Clerk actually took £500 from the 8 Customer, in exchange for 540 Euros. 9 "Conclusion: the Clerk selected Debit Card 10 as the method of payment early in the dialogue, 11 but settled to Fast Cash at the end of the 12 Session." 13 Is this another example of being able to 14 access the buttons pressed and the messages 15 displayed that we saw in the previous PEAK, 16 albeit Mr Allen has not cut in to his entry on 17 the PEAK the text that supports what he has 18 said? 19 A. Yes, he was using the same type of information 20 from the POC log to give a narrative to what 21 seems to have happened. 22 Q. So is it right, then, that the documents we have 23 looked at show what selections, if I can use 24 that word, a subpostmaster has made and what 25 messages are displayed to the subpostmaster in 16 1 the course of the session they are engaged in, 2 as opposed to a record of every keystroke made? 3 A. Yes, and, yet again, I will say this is only for 4 HNG-X. 5 Q. Can we look, please, at an example of where you 6 have seemingly have had access to the POC log, 7 FUJ00085913. You'll see that this is a PEAK, 8 dated 14 October 2015 -- if we just scroll down 9 for the first entry -- with your name against 10 it. It, in fact, concerns Bug 4 that we're 11 going to look at a little later, the 12 Dalmellington bug? 13 A. Yes. 14 Q. If we see the summary, if we scroll up please, 15 "Horizon -- transaction discrepancies". If we 16 can skip, please, straight to page 5 of this 17 PEAK, and if we look in the -- sorry, page 3. 18 Can you see, right at the foot of the page we're 19 looking at here, it says: 20 "keystrokes: Back Office, Remittances and 21 Transfers, Delivery Scan your barcode"? 22 A. Yes. 23 Q. Can you help us, where is that information from? 24 A. I would have got that from the Post Office 25 Counter log. 17 1 Q. That's a similar sort of cut and paste by you 2 from the POC log into this PEAK? 3 A. Yes. 4 Q. If we go forward, then, to page 5 -- 5 A. No, actually, that's not -- hang on, that's not 6 me, because this is still an update that has 7 been put on by -- 8 Q. If we just go back to the foot of page 1. 9 A. Yes, this -- yeah, this bit that's highlighted 10 at the moment is information that's either been 11 provided by it looks like it might have been 12 provided by NBSC. 13 Q. Look at the foot of page 1. That's where this 14 entry begins, I think. 15 A. Yes. So this is information that has either 16 been -- that has been added by HSD, or whatever 17 they were called at this point in time, based on 18 information that they had received from NBSC. 19 Q. So did NBSC have access to the POC log then? 20 A. No. They must have asked the branch what they 21 had pressed to get into this situation. 22 Q. Just go back to that entry we were looking at. 23 If you look at the whole entry, that doesn't 24 look like it's the record of a conversation in 25 which a subpostmaster said what buttons they had 18 1 pressed -- 2 A. Um -- 3 Q. -- does it? 4 A. I would say, yes, it does. They have been asked 5 by the Helpdesk, one of the helpdesks, specific 6 questions and that is what they have answered. 7 Q. So where it says, "keystrokes" that's a record 8 of a subpostmaster saying it, is it? 9 A. Yes, because those are the -- sorry, those are 10 the buttons that he would press to do this 11 process. 12 Q. Then if we go forwards, please, to page 5, and 13 go to the bottom half of the page, please. We 14 can see entries from you from 14 October 15 onwards? 16 A. Yes. 17 Q. If we look at the third entry there, timed at 18 15.35.38, "Evidence Added", and then is that 19 a POC file reference code? 20 A. Yes, it is. 21 Q. What's that saying that you have done? 22 A. I have, by this time, extracted the POC file for 23 the day from the counter. I have examined it. 24 I made some comments on it, which are further up 25 the screen. 19 1 Q. Yes. 2 A. I have put it through the obfuscation process to 3 make sure that no personal data is visible to 4 unauthorised staff and then, once it was 5 downloaded, it was automatically attached to the 6 PEAK. 7 Q. Do we see the automatic attachment three 8 entries, four entries on, where there's 9 an underlined entry reading "8th Oct poc.log"? 10 A. Yes, I've added two different logs, one for 11 8 October and one for 1 October. 12 Q. So if we had the PEAK system available to us 13 now, that would be a hyperlink through to those 14 files, would it? 15 A. Yes, I don't know if those files would -- 16 underlying files would still exist or if they 17 were deleted after a certain length of time. 18 Q. Look at it the other way, then, back in 2015, if 19 you clicked on those, that would take you 20 through? 21 A. Yes. 22 Q. So what was the purpose of putting the 23 attachments in, in this way? 24 A. To make that available to fourth line support, 25 who were GDC by this point. 20 1 Q. If we go to the foot of the page, please the 2 second entry up from the bottom, you say: 3 "Routing to GDC [fourth line support, yes] 4 to investigate by user was able to press and 5 enter and settle the same 'rem in' basket 6 multiple times. I have not managed to reproduce 7 this." 8 So can you tell from that entry, and in the 9 absence of us having a POC log, the extent of 10 the data that you were able to see. 11 A. I was able to see the button presses and, if we 12 could just go up the page a little bit, I did 13 put an update on to say there that I could see 14 from the button presses that "Enter" had been 15 pressed several times -- 16 Q. If you keep going up, the second entry there at 17 17.42.11. 18 A. Yes. 19 Q. "I can see that the clerk pressed Enter 4 times 20 ..." 21 A. Yes. 22 Q. So thinking of the division that we made earlier 23 or the evidence you gave about the division 24 earlier, on what the POC log data did and did 25 not record, this seems to suggest that delivery 21 1 receipts were printed and then the clerk just 2 pressed "Enter" four times? 3 A. That's what the log showed, yes. 4 Q. So you could see a keystroke -- 5 A. I could see those keystrokes, yes. I -- yes. 6 You could see the -- and pressing that key would 7 then cause the screen to move to a different 8 screen, so it was -- these were navigational 9 keystrokes or keystrokes in response to 10 messages, and so on, you could see. 11 When you asking yesterday, I thought you 12 were asking about every key that was typed and 13 certainly that was not all recorded. 14 Q. So if, in my example of committing cash to the 15 account earlier of the £2,500 versus the £250, 16 if the clerk, after they had typed £250, had hit 17 "Enter" four times, would you be able to see 18 that? 19 A. Um -- it would -- it would depend precisely how 20 it was set up. You might be able to see "Enter" 21 being pressed but I can't be certain. I don't 22 know. 23 Q. What, if you can assist us, please, what on this 24 occasion allowed you to see multiple button 25 presses of the same nature? 22 1 A. I can't -- 2 Q. Is it the function being performed? 3 A. It's the function being performed. I can't 4 remember what the question was that they were 5 pressing enter in response to. I think it is 6 recorded somewhere. It may well be -- maybe it 7 was something along the lines of "Has the 8 receipts printed properly?" They pressed 9 "Enter" for yes, which should then have taken 10 them out of the process but, because there was 11 an error situation, it went backwards and then 12 printed a second delivery receipt and then they 13 were asked again, has it printed? It had, so 14 they pressed "Enter" for yes and, again, it 15 was -- this was an error situation but they were 16 pressing cases "Enter", which should have taken 17 them out of the process but it wasn't working as 18 it should. 19 Q. Thank you, that can come down from there. 20 That's the only questions I ask about that topic 21 from yesterday. 22 Can we go back to where we were from last 23 night and explore your contact with 24 subpostmasters. As we read yesterday in 25 paragraph 212 of your statement, you said that 23 1 the subpostmasters were not your clients. If 2 you spoke to a subpostmaster, did you give them 3 your name? 4 A. Um, I'd certainly give them my first name. 5 Probably not usually my surname. 6 Q. Did you give them a means of contacting you? 7 A. No. 8 Q. Why was that? 9 A. Because they were not meant to have direct 10 access through to third line support. 11 Q. How would they get access to you? 12 A. They could phone the helpdesk and ask that 13 a message be passed to me and that did very 14 occasionally happen. 15 Q. How very occasionally? 16 A. I don't know. Three or four times ever, 17 perhaps. 18 Q. In the 16 years? 19 A. Yes. It wasn't something that -- I mean, the 20 whole point of having a support structure is 21 that you've got the people nearer the bottom who 22 are actually beavering away, resolving the 23 problems and doing the investigations and 24 I think almost any support system you have 25 a certain amount of filtering with what direct 24 1 contact there can be. 2 Q. Was there a duty or an obligation on you to 3 speak to any subpostmasters or was it entirely 4 at your discretion, if you thought it might help 5 solve the problem? 6 A. It was at my discretion and I was slightly 7 surprised there didn't seem to be any guidance 8 given on that. 9 Q. Surprised at who? 10 A. Perhaps at the general processes but, you know, 11 I came into a team that was already up and 12 running, working in their way and when you're 13 doing that, coming in as somebody new, you 14 follow what everybody else is doing. 15 Q. We saw also yesterday that in paragraph 42(iv) 16 of your statement you said that the MSU was 17 responsible for liaising with the Post Office 18 via BIMS reports, if there were errors which 19 affected counter balancing? 20 A. If there were errors that affected the branch 21 accounts or client accounts, bills being paid, 22 information being fed through, they covered that 23 area as well and also banking transaction 24 discrepancies -- not discrepancies, anomalies. 25 Q. As counter balancing was your specialist area, 25 1 did that mean that you had more contact or 2 a greater relationship with the people in MSU 3 than others in the SSC? 4 A. Um, no, I think a lot of the counter calls -- 5 calls raised by MSU tended to be shared out 6 amongst the teams, so I think a lot of different 7 people would have had contact with them. 8 Q. Were the MSU involved in getting the Post 9 Office's approval for inserting or amending data 10 into branch accounts? 11 A. We couldn't amend data into branch accounts and, 12 no, they weren't. 13 Q. You said you couldn't amend branch accounts? 14 A. Yes. 15 Q. What do you mean by that answer? 16 A. You couldn't amend data that they had already 17 written. All that we could do was to insert 18 extra corrective transactions in the very few 19 cases where that was seen to be the best thing 20 to do to resolve a system problem that had 21 already happen. 22 Q. Were MSU involved in getting approval for 23 inserting extra corrective transactions? 24 A. No. 25 Q. Who was your point of liaison, therefore, back 26 1 to the Post Office to get approval for such 2 corrective amendments? 3 A. It went through whatever the particular change 4 control process was at that point and, in 5 practice, it would usually be the managers in 6 the Service Management Team who would talk to 7 people at Post Office. 8 Q. So who was your point of contact then, within 9 Fujitsu first? 10 A. Well, I would -- obviously, it changed over the 11 years. The formal way of doing it was for me to 12 fill in a form saying what was to be done, and 13 so on, and then there were people who had to 14 read that information and sign off that form. 15 In practice, I would probably talk to my 16 manager, a problem manager, one of the customer 17 service managers. It just depended who had been 18 involved with it. But there was a formal 19 sign-off process, as well, which would always 20 have included the SSC manager and one of the 21 customer service managers. 22 Q. How did you find out whether the Post Office had 23 approved the corrective amendments? 24 A. That would be added to the OCP, OCR, MSC -- 25 I can't remember all the acronyms -- but it was 27 1 part of the formal process that there had to be 2 a name and a sign-off on that. But I was not 3 responsible for actually going and seeking that 4 and making -- I just filled in the form to start 5 with and then other people were in charge of 6 making sure that the correct sign-offs were done 7 before I was then given the authorisation to do 8 a change. 9 Q. You said yesterday afternoon, right at the end 10 of the session of your evidence, that you knew 11 of cases where the Post Office did not tell 12 a subpostmaster that their financial data had 13 been altered remotely by somebody within 14 Fujitsu. That's at page 207, lines 20 to 24. 15 What was that knowledge based on? 16 A. Discussions, sometimes along the line of are 17 Post Office going to -- I wouldn't necessarily 18 be speaking directly to somebody within Post 19 Office for this, although I know there's one 20 occasion when I did, at least. But there were 21 several occasions where we'd say, "Will you 22 notify the branch or shall we?" And they'd say, 23 "No, we don't think it's necessary to notify the 24 branch". 25 Q. Why would they say or what reason did they give 28 1 for it not being -- 2 A. I don't -- 3 Q. -- hold on -- for saying it's not necessary to 4 notify the branch that their financial data had 5 been altered remotely by somebody within 6 Fujitsu? 7 A. That was their decision to make. I don't know 8 why they would make it. I would always have 9 been happier if the branch had been fully 10 informed. 11 Q. Why would you have been happier if the branch -- 12 A. Because I always thought -- 13 Q. -- hold on. The transcriber has to write down 14 what we say and it's easier if I get the 15 question out and then you answer. 16 A. Yes. 17 Q. I'm guilty of it as well, of interrupting you. 18 So did they give any reasons for not wishing 19 to inform the branch that their financial data 20 had been altered remotely? 21 A. I've seen it written down in one or two 22 instances, I think, because they didn't want to 23 let the branch know that there had been a system 24 problem. 25 Q. So deliberately keeping the existence of 29 1 a Horizon system fault from the subpostmaster 2 that it affected? 3 A. I think that certainly did happen on some 4 occasions. 5 Q. Were you uncomfortable with this? 6 A. Yes, I was, really. I just felt it would be 7 a lot clearer if everybody -- if the branches 8 knew when there had been a problem. I -- if 9 I spoke to a branch and there had been a system 10 problem then I would say, "There has been 11 a system problem". 12 One particular instance I can remember where 13 we -- I know the branch wasn't contacted was 14 where, as far as we were aware, the branch 15 was -- didn't know that the problem had 16 happened, it had been brought to our attention 17 because of an entry on the Reconciliation 18 Report, and so undoing what had been wrongly 19 recorded seemed like the best way forward and 20 they may well not have been aware that they had 21 had a problem in that case. 22 Q. When you refer to the "best way forward" do you 23 mean the open and honest way forward? 24 A. The way to resolve it perhaps with fewest 25 questions. 30 1 Q. Well, did it seem to you that, in this respect, 2 the Post Office was applying an approach, so far 3 as the subpostmasters were concerned, of the 4 least said to them, the soonest mended? 5 A. I can't speak for Post Office but I certainly 6 got the feeling they did not want the -- there 7 were occasions when they didn't particularly 8 want the postmasters to know about problems. 9 Q. Can we look at some documents, please, starting 10 with FUJ00142197. This is an email sent from 11 you to Gareth Jenkins, and Andrew Keil and Mik 12 Peach, on 10 December 2007. 13 A. Yes. 14 Q. If we read it together, you say: 15 "Gareth, 16 "We have a problem with a branch where 17 a single SC line was written for 100 Euros 18 (£484) with no settlement. 19 "This was in the middle of two RISP 20 transactions and I suspect it's another oddity 21 in the LFS counter code. 22 "Initially it caused a harvester exception 23 because some of the BlackBoxData info was 24 missing, but that was corrected (so has gone to 25 POLMIS?) and now the set of transactions for the 31 1 day don't net to zero, hence on the Incomplete 2 Summaries report. 3 "I don't know what to do about it. As it 4 stands, when they balance I think they will have 5 a gain at the branch. If we correct the POLFS 6 feed so it nets to zero it will not be in line 7 with the branch, and will probably cause 8 problems in future. 9 "This might be a case for writing 10 a corrective message at the counter but this has 11 not been a popular approach in the past." 12 Then you ask some questions. 13 A. Yes. 14 Q. You say that inserting a message was not 15 a popular approach in the past. Is this 16 a reference to what you were just describing or 17 is this a different issue? 18 A. This is a reference to Post Office not wanting 19 us to make corrections. 20 Q. So this is the same issue that we were just 21 discussing? 22 A. Yes, this is. 23 Q. But this isn't a communication between you and 24 the Post Office, between Fujitsu and Post 25 Office, this is an internal communication? 32 1 A. Yes. 2 Q. Did Post Office's desire not to reveal to 3 subpostmasters errors in the system have 4 an effect on the extent to which you did insert 5 corrective messages at the counter? 6 A. Um, I don't think the alternative to writing the 7 corrective message was doing absolutely nothing. 8 Something had to be done about this particular 9 problem because, as I said, in this case it was 10 going to cause them potentially a gain, and 11 they'd got the sort of equivalent of a -- they 12 would have the equivalent of a -- now, would 13 they? Yes, they would have had a receipts and 14 payments mismatch or a non-zero line on their 15 branch trading statement. Sorry, this -- I'm 16 trying to remember a long way back now. 17 Q. Yes. 18 A. Because they hadn't balanced, there was still 19 an opportunity where a corrective message at the 20 counter to cancel out this incorrect line would 21 have put them in the state that they should have 22 been in, so it seemed worth considering that. 23 Q. What I'm asking is it seems that, by at least 24 December 2007, the reluctance of the Post Office 25 to reveal to subpostmasters, through the use of 33 1 corrective action, errors in the system was 2 having a chilling effect on you within Fujitsu 3 about your willingness to do it? 4 A. Yes, I -- 5 Q. Would that be fair? 6 A. Um, I mean, there's the other position, which is 7 that, you know, writing a corrective message, 8 SSC making changes to counter accounts, you can 9 understand why there was quite a reluctance to 10 give us permission to do that as well. 11 Q. Why? 12 A. Possibly because, at some levels, it was thought 13 that we didn't have the ability to do that. 14 I don't know. I cannot speak for Post Office. 15 Q. Can we look, please, at FUJ00087194. This, 16 I think, is related to the email that we just 17 saw. 18 A. Yes. 19 Q. Just looking at the whole page first, can you 20 describe what this document is? 21 A. Sorry, can I have a drink and a cough. 22 Q. Yes, of course. 23 A. This one of the change procedure documents, so 24 an OCP which I filled in what has been proposed, 25 why the change is justified, when it'll be done, 34 1 more details as to precisely what will happen, 2 and then I'd already talked to Gary Blackburn at 3 Post Office about it, so this is obviously after 4 the discussion that I had with Gareth. 5 And then further down we can see that 6 approval has been sought from Post Office 7 through the formal route and there should also 8 be sign-off by my manager. 9 Q. Thank you. So if we just read through it 10 together: 11 "Write corrective bureau message for [then 12 the branch code is given]. 13 "A single ... message ... was written in 14 error on 26th November ... selling 1,000 US 15 dollars, with no corresponding settlement line. 16 To remove the effects of this message at both 17 the branch and on POLFS, we will need to insert 18 a new message to negate the effects of the 19 original message. 20 "Justification: If the change is not made in 21 the counter messagestore (before the stock unit 22 is balanced on Wednesday), the branch will have 23 an unexpected gain of £484 (or thereabouts ...), 24 and a receipts and payments mismatch. This gain 25 would have to be resolved at the branch. There 35 1 would also be an inconsistency between the 2 branch and POLFS to be resolved. By correcting 3 the problem locally, the branch may not be aware 4 of the problem, and there will be no 5 inconsistency between the branch and POLFS." 6 You set out when it's planned for. You set 7 out some extra detail. Then you say: 8 "The message will include a comment to show 9 it has been inserted to resolve this problem 10 (this will not be visible to the branch)." 11 Skipping a paragraph, you say: 12 "Neither the new nor the old message will be 13 included in data sent to POLFS." 14 So I think this is a record to show that, 15 despite the misgivings in the email exchange we 16 looked at earlier, authorisation had been given. 17 But you record twice on this document that, by 18 doing it this way, the branch will not be aware 19 of the problem and that the message will not be 20 visible to the branch. Why was it important to 21 record those two things? 22 A. Just so it was known that that was the case. 23 It's not saying that none of it would have been 24 visible to the branch. They would have been 25 able, if they'd printed their transaction log, 36 1 they would have seen the first transaction and 2 they would also see the equal but opposite 3 transaction. They would see that but they would 4 not have seen the comment -- 5 Q. Who had done it? 6 A. -- of who had done it. 7 Q. Why was it important to record that the "who had 8 done it" will not be shown to the subpostmaster? 9 Why were you writing that on here? 10 A. Um, just in case anybody at some point in the 11 future wanted to know. I just tried to -- you 12 know, I wrote down as full a description as 13 I could of what was happening and so, if there 14 was a question at some point, we would know this 15 particular fact. 16 Q. In writing it, were you giving some reassurance 17 to POL "Don't worry, this won't be shown to the 18 branch. They won't see what's going on here"? 19 A. I don't recall that being my intention at the 20 time. I certainly wasn't doing anything to try 21 to specifically hide it from the branch. 22 Q. Wasn't that the effect of what you were doing, 23 though? 24 A. I don't think I could have added anything on 25 that would -- could I have made it obvious to 37 1 them in some way? I'm not sure. 2 Q. Wouldn't telling the branch assist them in 3 future -- 4 A. Yes. 5 Q. -- in that if there had been a recurrence that 6 was not picked up, then they might understand 7 better how it had happened? 8 A. A recurrence would have been picked up by the 9 same things that picked up this one. They 10 hadn't reported "This is a problem already". If 11 it had happened again, it would have been picked 12 up by the same mechanism that picked it up this 13 time. 14 Q. So are you saying that it's best not to worry 15 them with a fault in the system? 16 A. I wasn't making the decision as to whether the 17 branch should be informed or not. But, yes, by 18 doing it in this way, maybe I was thinking, "Oh 19 good, we can just get it sorted out before they 20 balance, they don't need to be bothered by it". 21 That probably -- you know, if I had realised 22 I was going to be questioned about it so long 23 afterwards, I might have possibly made 24 a different decision but that's the decision 25 I made back in 2007. 38 1 Q. Did the Post Office tell you to undertake this 2 correction in a way that did not reveal this 3 information to the branch? 4 A. I don't recall them specifically saying that. 5 Q. Or did you do it in that way, as a matter of 6 choice, because you knew that that's what your 7 client would want? 8 A. I cannot remember and I haven't seen any 9 documentation as to whether I had a conversation 10 with Gary Blackburn as to whether he was going 11 to contact the branch about this or not, and 12 I don't know what he said in reply. I think 13 I probably would have asked him that question 14 but I can't remember. 15 Q. I mean, is what we see here -- you undertaking 16 the corrective transaction in a way that does 17 not reveal the way in which the corrective 18 transaction has been undertaken and who has done 19 it to the postmaster -- reflect the view that 20 you received from the Post Office, that it was 21 important not to reveal to subpostmasters any 22 hint that there were issues with the reliability 23 of Horizon? 24 A. I don't think I took this action for that 25 reason. 39 1 Q. Albeit the effect of your actions was not to 2 reveal to a subpostmaster the person and the 3 means by which the corrective action had been 4 undertaken? 5 A. That was the result of what happened, given that 6 Post Office chose not to talk to the postmaster. 7 Q. Can we look, please, at POL00023765. This is 8 a PEAK from 7 December 2007; can you see that? 9 A. Yes. 10 Q. From the summary, the issue is with a branch and 11 a branch FAD is given, "POLFS Incomplete 12 Summaries Report". You become involved in this 13 later. 14 A. Yes. 15 Q. Can you recall or explain what an incomplete 16 summaries report is? 17 A. Where the transactions, which had been for 18 a day, for a branch, were harvested to be sent 19 on to POLFS, which was their financial back end 20 system. If the transactions didn't net to zero 21 then they would not be sent and we would have to 22 investigate, you know, why there was an issue. 23 Q. If we go over the page, we can see, I think, you 24 attaching some files, is that right, on 25 10 December? 40 1 A. Yes. This is the same branch as before. 2 Q. Yes. 3 A. Yes. 4 Q. We can see on the 11 December a couple of files 5 or links to files, entitled "Details of how 6 POLFS feed was corrected" and "Correction made 7 to counter messagestore"? 8 A. Yes. 9 Q. Again, are they hyperlinks to documents -- 10 A. Yes. 11 Q. -- that we -- I don't think we have those. But 12 anyway, if we go to the foot of the page, 13 please, and look at Andy Keil's entry. Was he 14 a colleague of yours in SSC? 15 A. Yes, he was. 16 Q. He notes at 17.19.46: 17 "Worth noting that the branch did not have 18 any issues with the mismatched transactions 19 because this was fixed before they did the roll. 20 The branch is not aware of this and it's best 21 that the branch is not advised." 22 A. Yes. 23 Q. Again, is that a further reflection of a culture 24 within the SSC of it's best not telling the 25 branches where such corrective measures are 41 1 undertaken to their financial data by the SSC? 2 A. I think it's just reflecting that, in this one 3 specific case, Post Office had said that they 4 did not want to -- they were not going to 5 contact the branch. 6 Q. You said "in this one specific case". 7 A. Yes. 8 Q. You said earlier in your evidence and last 9 night, that you were aware of cases where the 10 Post Office did not tell a subpostmaster that 11 their financial data had been altered remotely 12 by somebody within Fujitsu. You're not 13 suggesting that this was the only example of it, 14 are you? Rather, this is reflective of that 15 wider practice, is it not? 16 A. This is the call that I had in mind when giving 17 those answers. Very, very hard to remember now 18 but I think, as time went by, we were aware that 19 Post Office certainly did not always want to 20 tell the branches of faults, and so on. But 21 I wouldn't say that this was fixed within SSC. 22 As I've said before, if a branch had raised the 23 problem themselves and we were talking to them 24 and it -- we knew it was a system error, then, 25 yes, we would say so. 42 1 Q. What explains the difference of approach, then, 2 if the -- 3 A. Because the branch may not have been aware of 4 this issue. It had only been -- they hadn't 5 reported it as a problem. It had only been 6 picked up on our internal reports. 7 Q. Did you feel uncomfortable with this? 8 A. Yes, I did. I would -- I think I said earlier, 9 I would rather that the branch had been involved 10 in the discussions, so they knew what was 11 happening. 12 Q. Is this another case of you just doing what was 13 common practice and that which your client 14 wished you to do? 15 A. I don't think it's that unreasonable to do what 16 your client wishes you to do. As to whether it 17 was common practice, this, you know, the whole 18 process of making counter corrections was pretty 19 unusual. It was not something that was 20 happening every week, every month. They were 21 very, very few and far between. So this was 22 what our client wanted at the time. Perhaps it 23 was me anticipating what our client might or 24 might not want to do. But, personally, I would 25 have been much happier if the branch was aware 43 1 what was being done. 2 Q. Did the Post Office ever give any good or 3 substantial or honourable reasons for not 4 wishing for this material to be revealed to the 5 subpostmaster? 6 A. I'm not sure that they gave us our reasoning -- 7 gave us their reasoning in that way, no. 8 Q. Was it a case, then, that they were -- the 9 reason was the least the subpostmaster knows 10 about errors in the system, the better? 11 A. I think you have to ask what Post Office what 12 their thoughts on that are. But I would say, 13 yes, I did get that impression at times. 14 Q. How and from whom did you get that impression at 15 times? 16 A. I think possibly once or twice I was on 17 a conference call about a system problem with 18 Post Office people, and I think I've seen at 19 least one document where it's minuted that they 20 don't want -- they didn't want to give 21 opportunities for fraud, if postmasters became 22 aware of certain issues. 23 Q. Can you just explain how revealing to 24 a subpostmaster that a corrective action had 25 been made to correct a bug in the system would 44 1 give an opportunity for a subpostmaster to 2 commit an offence of fraud? 3 A. I wasn't talking about corrective actions there, 4 I was talking more about overall discussion of 5 system problems that had occurred. I don't 6 recall that ever being said. In fact, I'm sure 7 that wasn't ever said in any discussion as to 8 a single corrective action at a branch. 9 Q. We've seen some evidence that people such as 10 Penelope Thomas, Andrew Dunks, Brian Pinder 11 produced ARQ branch data for the purposes of 12 proceedings. Was there any method to alert them 13 that corrective action had been taken to insert 14 data or extra messages into a branch's accounts? 15 A. If they had looked at all the PEAK calls for 16 a branch, they might have seen those but I don't 17 know if that was part of their process. The 18 OCR -- the ARQ data would contain the -- both 19 the original transaction and the corrective 20 transaction at the point at which they were 21 done. 22 If the full unfiltered data was retrieved 23 and inspected, then that would show the comment, 24 for example, that I mention was added in this 25 one. Certainly sometimes for counter 45 1 corrections, and it wasn't done consistently, 2 but often we might use a counter number that 3 didn't exist to make it clear that it was 4 something out of the ordinary, or a username, 5 including SSC, again to show that it was 6 something out of the ordinary. 7 That wasn't done on this specific one and 8 I cannot remember whether that was because I was 9 specifically asked not to or I was just 10 producing a transaction that was absolutely 11 a mirror image of the one that shouldn't have 12 been there in the first place, and all I did was 13 just change the signs on the values, 14 effectively, and I left all the other data in 15 there as it was. But I cannot properly remember 16 my reasoning. 17 Q. What was the purpose of using a fictitious 18 username? 19 A. To make -- well, if it had "SSC" in it to make 20 it clear that it was not done by somebody at the 21 branch. 22 Q. Did you always use SSC or did you use other 23 fictitious usernames that did not identify the 24 SSC as having made the change? 25 A. It would always have been something that was 46 1 very clear that it -- I -- as I say, I can't 2 remember without an example if it would have 3 been something like SSC999, which would have 4 been a valid username, or something else, but it 5 wouldn't have "Fred12" or something. It would 6 have been something to draw attention to it, not 7 to try to hide it. 8 MR BEER: Yes, thank you. 9 Sir, that might be an appropriate moment for 10 the morning break, as I move next to some 11 examples of bugs, errors and defects. 12 SIR WYN WILLIAMS: Yes, by all means. How long do 13 you think is appropriate? 14 MR BEER: Until 11.30, please, sir. 15 SIR WYN WILLIAMS: Yes, fine. 16 (11.12 am) 17 (A short break) 18 (11.30 am) 19 MR BEER: Good morning, sir, can you continue to see 20 and hear me? 21 SIR WYN WILLIAMS: Yes, I can. Thank you. 22 MR BEER: Thank you very much. 23 I keep promising to get on to bugs, errors 24 and defects but I've still got to cover 25 something that I rather skipped over, 47 1 Mrs Chambers. 2 Can we go back, please, to POL00023765. 3 This was the PEAK that we just looked at about 4 the corrective action. 5 A. Yes. 6 Q. If we can just look at the foot of page 2, 7 please, we've got the message or the entry by 8 Andrew Keil that we looked at in the morning 9 session at 12 December, 17.19.46? 10 A. Yes. 11 Q. "Worth noting that the branch did not have any 12 issues with the mismatched transactions because 13 this figure before they did the roll [the 14 rollover]. The branch is not aware of this and 15 it's best the branch is not advised." 16 So is that recording that by 12 December, 17 the fix had been applied? 18 A. Um, I assume so. Yes. I mean, it was in the 19 OCP when it was due to be applied. 20 Q. Yes. If we just go over the page to an entry 21 that I didn't take you to, your entry on 22 14 December at 16.13.37. You say: 23 "As detailed above, the two POLFS incomplete 24 summaries ... have been resolved. 25 "The counter problem which caused the first 48 1 issue has been correct by inserting a message 2 into the messagestore, for equal but opposite 3 values/quantities, as agreed with POL ..." 4 Then you give the OCP reference. 5 A. Yes. 6 Q. "As a result of this corrective action, the net 7 effect on POLFS is zero, and POLFS figures are 8 in line with the branch. POLMIS received both 9 the original message and the corrective 10 message." 11 But then you say this: 12 "Once the problem was corrected, there 13 should have been no impact on the branch. 14 However it has been noted that the stock unit of 15 BDC had a loss of [£]1,000, which was generated 16 after the correction was made. We have already 17 notified Gary Blackburn at POL (email attached). 18 This appears to have been a genuine loss at the 19 branch not a consequence of the problem or 20 correction." 21 So by 12 December the corrective fix had 22 been applied concerning a loss of $1,000. After 23 that correction had been effected, a stock unit 24 showed loss of $1,000. It was only generated 25 after the correction was made and you're saying 49 1 that this appears to be a genuine loss of the 2 branch and nothing to do with the correction. 3 A. Yes, I have obviously been thinking about this 4 quite a lot. The loss was only generated when 5 they balanced so that's why it showed at that 6 point, they hadn't balanced before then. 7 I think my conclusion that it wasn't 8 a consequence of the problem may have been 9 wrong. It wasn't a consequence of the 10 correction. I know that Mr Justice Fraser 11 considered some of this and there was -- I'm 12 afraid we haven't got the documents in front of 13 us, but his view was that there had been two 14 different corrections done and one of them was 15 for the wrong amount, and I can -- I disagree 16 with that strongly, in that the correction that 17 he thought was for the wrong amount didn't 18 affect the branch accounts at all. That was the 19 correction to the POLFS back end feed. 20 But yes, the branch then had a loss in this 21 stock unit. One possibility was that they had 22 done a balance snapshot or something during the 23 week and realised that, actually, they had got 24 $1,000 more than they expected in that stock 25 unit and had taken it out of there and put it 50 1 into the main safe to see what happened. 2 Another is that I'm now wondering if this line 3 that was incorrectly written as an "SC", serve 4 customer, should actually have been another of 5 these RISP lines, which was reversing a rem out, 6 and so whether it now -- can I get this right? 7 Yes, that would -- if it was the case, that 8 would have had this effect. 9 But I agree now, certainly given those 10 circumstances, it would have been far, far 11 better to have talked to the branch at that 12 point to try and work out whether they did have 13 a genuine loss at the end of the day, whether it 14 was something that they then could resolve 15 themselves. I'm not aware that they ever phoned 16 in about it. I don't know if Gary Blackburn, 17 who was aware of this, ever contacted them or 18 checked to see if they did have any lasting 19 problem, but no, this -- it was not as a result 20 of the correction, but it wasn't the state that 21 we wanted to end up in. 22 Q. On what basis did you, in the light of what 23 you've just said, conclude that this was 24 a genuine loss at the branch? 25 A. Um, I don't know. I mean, because I had checked 51 1 very carefully and I could see that my 2 correction had done precisely what I intended it 3 to, which was to remove this rogue SC line, 4 which should not have been written. It's only 5 now I'm wondering if, when it was written, it 6 should actually have been another RISP line but 7 I can't prove that at this point. 8 Obviously, if the branch had raised 9 another -- a call saying that they'd got 10 an unexpected loss, "What on earth has been 11 going on", then that would have been 12 investigated and followed through but, to the 13 best of my knowledge, they didn't. 14 Q. Did it occur to you at the time that the amount 15 of the correction, the value of the correction 16 that you had made, was equal to the value of the 17 loss that was now being shown? 18 A. Yes, of course it did. Which is why I checked 19 and double checked and triple checked. 20 Q. And therefore might be a relationship between 21 the two? 22 A. Yes. 23 Q. Isn't what you've just said though, to put the 24 burden back on the branch, to say they need to 25 complain again, they've got to go through the 52 1 whole rigmarole of going to NBSC again? 2 A. They hadn't actually complained at all. 3 Q. No. 4 A. They didn't raise the original call. 5 Q. Okay, they would have to go through a rigmarole 6 for the first time, then? 7 A. Yes. 8 Q. Okay, we'll move on. I'm going to ask you about 9 as many of Bugs 1, 2, 3, 4, 5, 6, 7, 8, 10, 19 10 and 23 as identified in the appendix to the 11 judgment of Mr Justice Fraser in the Horizon 12 Issues trial as time allows today, and then I'll 13 revert to the process when we meet again for the 14 Phase 4 evidence. 15 I'm not going to rely on his findings for 16 the purposes of asking you questions, not least 17 because we have more material than was 18 apparently made available to Mr Justice Fraser. 19 Just so you understand, what I'm going to try to 20 do is firstly seek to understand in general 21 terms what the nature of the relevant bug was, 22 in a very high-level summary, then identify the 23 issues that I would like to try and explore with 24 you, then run through the material in 25 chronological order that concerns that bug, and 53 1 then explore any issues that are left that 2 haven't been addressed. 3 I'm not going to explore the bugs in 4 chronological order, simply do them 1, 2 and 5 following. 6 A. Yeah. 7 Q. Before we get to that, would you agree that not 8 all errors in the Horizon System were caught by 9 the automated processes set up by Fujitsu to 10 detect errors? 11 A. Yes. 12 Q. You tell us in your statement -- no need to go 13 there, it's paragraph 41 -- that: 14 "From around 2007 a real-time monitoring 15 system was developed by the SSC to alert us to 16 system-wide problems, for example a large number 17 of debit card transactions failing. This system 18 was tweaked and expanded over the years." 19 What was the name of that system? 20 A. The SSC monitor? Um, I can't properly remember. 21 Q. Who monitored it? 22 A. We took it in turns. We had an SSC monitor 23 monitor. 24 Q. Was that one person a day or a shift? 25 A. Yeah, a day. 54 1 Q. Who developed it? 2 A. I think John Simpkins probably did most of it 3 but then, if other people had good ideas of how 4 things could be monitored, they got sort of 5 added into it. It was more monitoring of the 6 back-end systems, not the counters themselves 7 although, obviously, banking transactions, and 8 so on, were going all the way through the 9 system. 10 Q. When you say monitoring of the back-end systems 11 are you referring to POLFS there? 12 A. No, I'm -- 13 Q. You're referring to Fujitsu back-end systems? 14 A. Fujitsu back-end systems. 15 Q. So how did it operate? What was its coverage? 16 A. I cannot remember any -- much detail at all. It 17 was just wherever there was a useful source of 18 information, perhaps, for example, about the 19 number of debit card transactions going through, 20 they would have a response code on them to say, 21 if it had been a successful payment or 22 otherwise, we could monitor how many were going 23 through a particular point in the system with 24 some sort of failure/error code. And if it 25 exceeded a certain threshold, then it would go 55 1 red instead of green and that would encourage 2 somebody to see what was going on. 3 Q. So it was essentially a sort of pattern 4 analysis? 5 A. Yeah, for that particular instance. 6 Q. Can you help us, other than failed debit card 7 transactions, what, if anything, else it picked 8 up? 9 A. Banking transactions, which were actually 10 a separate system. I can't now remember the 11 details, I'm sorry. 12 Q. There wasn't such a system before 2007; is that 13 right? 14 A. Not sort of trying to pick up problems before 15 anybody had reported to them -- them to us in 16 some other way, yeah. 17 Q. But, in any event, the system didn't itself 18 proactively identify all bugs, errors and 19 defects? 20 A. Not at all, no. 21 Q. Was Fujitsu essentially reliant on, therefore, 22 a problem occurring within the live estate 23 causing a discrepancy or a loss, and the 24 subpostmaster raising it through the NBSC or the 25 Horizon Support Desk? 56 1 A. We had all the reconciliation reports that ran 2 overnight, so that was the main way of finding 3 financial inconsistencies on the system. 4 Q. So there was the reconciliation reporting 5 system? 6 A. Yes. 7 Q. Did it nonetheless remain the case that the 8 majority of bugs were picked up through 9 subpostmaster initiated action? 10 A. Um, I mean, obviously there were problems to be 11 investigated throughout the whole system, all 12 the back-end stuff as well, but if we're talking 13 specifically about counter balancing problems, 14 which were only a very small proportion of the 15 overall calls that we were handling, um, then 16 I would say it was probably about 50:50 17 inconsistencies being reported by -- on the 18 reconciliation reports or branches reporting 19 that they had a problem in a particular area. 20 Q. You said that counter balancing was only a small 21 proportion? 22 A. Oh, yes. 23 Q. But it was a very significant issue for the 24 subpostmaster concerned -- 25 A. Yes, of course. 57 1 Q. -- potentially? 2 A. Yes. 3 Q. Did you realise that at the time, did you 4 acknowledge that at the time, that the 5 consequences for a subpostmaster may be very 6 extreme indeed? 7 A. I don't think we -- certainly, as I think I said 8 yesterday, I didn't realise initially that -- 9 how -- really how the Post Office subpostmaster 10 structure worked and that they were financially 11 responsible. Obviously, some of problems would 12 have been at the bigger Crown branches, which 13 Post Office were responsible for. And there was 14 always this huge difficulty in distinguishing 15 where a problem is caused by something in the 16 system and the -- certainly more than just 17 a possibility that it is caused by some 18 inaccuracy of processing at the branch itself, 19 the user input. 20 Q. Did you and others in the SSC treat counter 21 balancing issues any differently because of 22 a recognition that the consequences for 23 a subpostmaster may be very direct and personal? 24 A. I don't think that would mean that we would 25 necessarily give it -- you know, sort of put it 58 1 to the very top of the heap. You could argue 2 that it's actually extremely important that 3 a branch or a whole series of branches can 4 trade. If they're not able to trade, that is 5 also -- has serious consequences for all of 6 them. 7 If the entire estate can't do banking 8 transactions that obviously also has a severe 9 impact on the whole estate and so, to some 10 extent, I think those type of problems may have 11 been seen as more important -- not more 12 important but would possibly require faster 13 action than a discrepancy call from a single 14 branch. 15 I mean, I do see now that, yes -- I am well 16 aware of the impact that these problems have 17 had. But it was so hard to distinguish between 18 business issues and potential system issues, and 19 we would look for every possible sign of 20 a system issue. But if there wasn't one, 21 without knowing what had actually taken place at 22 the branch, you can't do more. 23 Q. Would your view have been different as to the 24 relative importance accorded to bugs, errors and 25 defects that may have affected the ability of 59 1 the system to continue to trade, ie financial 2 issues, on the one hand, and issues that may 3 affect the continued employment or suspension, 4 civil proceedings against, criminal 5 investigations and criminal proceedings against, 6 subpostmasters, on the other, if you had known 7 more about how the Post Office had treated the 8 subpostmaster contract as meaning that 9 subpostmasters were responsible for all losses? 10 A. Yes. I feel we should perhaps have been warned 11 if the result of us looking at a single call 12 over a single day, or whatever, was going to -- 13 could result in action being taken against 14 a postmaster with, I don't know -- I don't know 15 how much extra investigation was ever done. 16 Q. In the early days, say between 2000 and 2006, 17 did you not realise, therefore, that the 18 conclusions that you reached, the nature of the 19 investigations that you undertook that preceded 20 them and which you wrote up on a PEAK, could 21 result in the next day a subpostmaster being 22 suspended and locked out of their branch? 23 A. No, I don't think we did realise that. 24 I assumed there would be a huge amount more 25 investigation and double checking of the figures 60 1 and everything else. 2 Q. Double checking by whom? 3 A. I assume people in Post Office would be doing 4 that. 5 Q. Can we turn to Bug 1, please, the payments and 6 receipts mismatch bug. Can we start with my 7 sort of summary of it. Would you agree with the 8 following summary of the bug: firstly, it was 9 a Horizon Online bug that occurred in 2010? 10 A. Yes. 11 Q. Secondly, it occurred when a subpostmaster tried 12 to roll over a stock unit with a discrepancy? 13 A. Yes. 14 Q. Thirdly, the system would ask the subpostmaster 15 if they wanted to transfer the discrepancy to 16 the local suspense account? 17 A. Yes. 18 Q. If the subpostmaster cancelled the rollover, the 19 discrepancy was zeroed from the location cache 20 but nothing was written to the branch database? 21 A. Yes, I believe that's true. 22 Q. If the subpostmaster then tried to roll over, 23 the stock unit would be rolled with the corrupt 24 local cache missing the discrepancy? 25 A. Yes. 61 1 Q. That would therefore create a receipts and 2 payments mismatch? 3 A. Yes. Although I think that receipts and 4 payments mismatch wasn't actually picked up 5 until the end of the following period. 6 Q. The issues that I would like to explore with 7 you, please, are, firstly, why it appears that 8 only significant action was taken in relation to 9 this bug from September 2010 onwards when, 10 firstly, the PEAKs in relation to it had been 11 raised in February 2010 and, secondly, 12 Mr Jenkins appears to have been aware of the bug 13 in May 2010 when he noticed a Windows NT event; 14 and then, secondly, what was done to ensure that 15 all branches that may have been affected by the 16 bug had been properly identified. 17 A. Right. I've got several things to say in 18 response to that. Firstly, from everything I've 19 seen about this bug, I was not involved in the 20 investigations in September. So, really, 21 everything I'm going to say is based on what 22 I have read since. I have no memory of it. 23 I haven't seen any evidence that suggests it 24 was -- that it did occur before September. 25 I know there were couple of receipts and 62 1 payments events which Gareth flagged, and 2 there's an email about that earlier in the year, 3 and there was also at least one other problem 4 that occurred during the HNG-X pilot, which was 5 roughly the first six months of 2010. But they 6 were different underlying causes and I'm not 7 aware that this specific problem, which resulted 8 in a receipts and payments mismatch had been 9 seen or reported before September. 10 Q. That's a very helpful general answer. Can we 11 look at material then, the chronology of events. 12 There are about ten steps in the process that 13 I would like to ask you about, but there are 14 about another 20 steps in the process but I'm 15 going to ask other witnesses about those or 16 suggest that they're adequately established 17 through the documents themselves. 18 Can we start, please, with FUJ00081064. Can 19 you see that this is PEAK 0194381. 20 A. Yes. 21 Q. It was opened on 10 February 2010? 22 A. Yes. 23 Q. You can see from the summary "Counter APP"; what 24 does "APP" mean? 25 A. Application? 63 1 Q. Total receipts £250,016.45, total payments 2 £200,016.45. Then if we see from the first 3 entry that summary is included, so a £50,000 4 discrepancy; do you see that? 5 A. Yes. 6 Q. So this is showing a mismatch, is this right -- 7 A. Yes. 8 Q. -- between receipts and payment -- 9 A. Yes. 10 Q. -- of £50,000? 11 A. It's reporting a mismatch, yes. 12 Q. Now, I don't think you, as you have said, ever 13 became involved in this PEAK, so far as I can 14 see; is that right? 15 A. I can't remember unless I go down the -- 16 Q. Yeah, if the operator could just scroll through, 17 please, you'll see I think your name doesn't 18 appear on it. 19 A. Okay. So it's been sent off to GDC, who are 20 providing fourth line support. 21 Q. Yes, and if we scroll down, please, I think we 22 can see that your name is not on it. 23 A. Yes, okay, I do now remember this. I mean -- 24 Q. This document? 25 A. I remember seeing this document before. Yes. 64 1 Q. We can see, if we go back up to the top of the 2 first page, that this becomes 3 "KEL ballantj1759Q"? 4 A. Yes. 5 Q. We can see that under the "All references" 6 section, yes? 7 A. Yes. 8 Q. Can we look then, at "KEL ballantj1759Q", that 9 is POL00029425. This was created, we can see, 10 by your colleague John Ballantyne on 12 February 11 2010 -- 12 A. Yes. 13 Q. -- and last updated by you on 17 May 2011? 14 A. Yes. 15 Q. The way a KEL is written, you can't actually 16 tell what Mr Ballantyne originally wrote and 17 what you changed subsequently; is that right? 18 A. You can't see on here. The old ones were kept 19 but I've no idea if they still exist. 20 Q. So the text on here, we can't see what was his 21 work and what's your work? 22 A. No, no. 23 Q. I don't suppose you now recall what changes you 24 made? 25 A. I may recognise some of my -- 65 1 Q. Your style? 2 A. -- style, but I'm not sure. 3 Q. You'll see it cross-refers, in about the tenth 4 line there, back to the PEAK we just looked at, 5 yes? 6 A. Yes. 7 Q. If we scroll down, please, it states the 8 problem: 9 "This event is generated when the payments 10 and receipts totals do not match on one of the 11 counter balancing reports. This indicates 12 a software error or data corruption." 13 A. Yes. 14 Q. So it continues: 15 "[This] has been caused in the past by ..." 16 Then three possibilities are set out, yes? 17 A. Yes, yes. 18 SIR WYN WILLIAMS: I'm sorry to interrupt you, 19 Mr Beer, but I've had a message to say that I'm 20 no longer on the screen. I'd just like to 21 assure anybody who is looking that I'm still 22 here and the problem with me being on the screen 23 is being seen to. 24 MR BEER: Yes, thank you, sir. We're going to be 25 looking at lots of documents at the moment so 66 1 you wouldn't have been seen, in any event, 2 because when we look at a document, you 3 disappear. 4 SIR WYN WILLIAMS: That's all right, then. That's 5 fine. 6 MR BEER: The solution is set out: 7 "SMC/counter eventing team: raise a B 8 priority call and send to SSC if you see this 9 event, unless it is from a training counter ... 10 "SSC: Instances of this error must be 11 investigated. If the error is as a result of 12 a new problem, please add the details to the 13 list of causes above. 14 "The branch accounts may need to be 15 corrected. See [another KEL] for advice on how 16 this has been done for a previous problem." 17 What do you understand "The branch accounts 18 may need to be corrected" to mean? 19 A. I don't know now and, when I covered this in my 20 witness statement, I hadn't seen the "wrightm" 21 KEL. I have now and it doesn't cast any light 22 on it so I'm sorry but I don't know why that's 23 there. 24 Q. You made a point in your witness statement 25 I would need to see "wrightm". We're going to 67 1 look at the "wrightm" KEL in a moment. 2 A. Yeah. 3 Q. So you don't understand what that means? 4 A. Unless it's to -- referring to the corrective 5 action that may or may not have been taken for 6 the September bug, where they pressed cancel at 7 a certain point. 8 Q. Who was this direction to correct the branch 9 account addressed to? 10 A. It's saying it to SSC, I believe, but don't 11 think I -- I'm just about certain I did not put 12 that in there. So I'm not entirely clear why it 13 is there. 14 Q. By what method would you identify which branch 15 accounts need to be corrected? 16 A. Once you had a full understanding of the 17 specific problem and its consequences. 18 Q. By what method would they correct the branch 19 accounts? 20 A. It would depend to the problem and its 21 consequences. 22 Q. In your witness statement -- no need to turn it 23 up -- paragraph 66, you say: 24 "Post Office would have been informed of 25 each instance. I am not sure whether this was 68 1 via a BIM or some other route. Fujitsu would 2 not have contacted branches directly unless the 3 branch had raised the call in the first place." 4 By that, are you saying that the Post Office 5 would have been made aware of each of the 6 individual cases where this issue affected 7 a subpostmaster or are you saying that the Post 8 Office would be informed that there was 9 a systemic problem? 10 A. Um, as I said, when I wrote this section I was 11 working a little bit blind, given that I had no 12 direct involvement with this. We wouldn't have 13 told Post Office about the office snapshots 14 problem. Actually, they probably were told 15 about it because everything in the pilot was 16 closely monitored, but the office snapshot one 17 there, that was false reporting of a receipts 18 and payments mismatch because it didn't take the 19 transfer into account. 20 Obviously, the stuff that had to be done for 21 the September problem was a major problem, which 22 was all followed through at the time. 23 Q. Followed through by? 24 A. I wasn't involved but I believe you've got some 25 more documents about it. 69 1 Q. So in that passage in your witness statement, in 2 which you said, "Post Office would not have been 3 informed of each instance" -- sorry, "would have 4 been involved (sic) of each instance", you're 5 not sure whether this was via a BIM report or 6 some other route, "Fujitsu wouldn't have 7 contacted the branches directly unless the 8 branch had raised the call in the first place." 9 Is that essentially a reflection of the 10 division of approach that you described to us 11 yesterday, ie what determined whether or not you 12 contacted a branch or not? 13 A. Yes. I believe so. 14 Q. Ie it depended on whether the branch had 15 initiated the issue? 16 A. Yes. 17 Q. Can we turn then to the wrightm...J KEL that you 18 said in your witness statement you needed to 19 look at. That's FUJ00081608. 20 This is the wrightm33145J KEL -- 21 A. Yeah. 22 Q. -- that we saw referred to in the KEL that you 23 had last updated on 17 May 2011. 24 You'll see that this KEL is not raised until 25 23 September 2010 -- 70 1 A. Yes. 2 Q. -- which is seven and a half months -- I hope 3 I've got the maths right on this occasion -- 4 since the PEAK that we were looking originally 5 at 10 -- 6 A. Yes, but that original PEAK was the office 7 snapshot problem, not the same problem that 8 happened in September. 9 Q. Why do you restrict the previous PEAK to only 10 the office snapshot problem? 11 A. I don't. It was originally raised for the 12 office snapshot problem but then when there were 13 other issues that could cause receipts and 14 payments mismatches, it was useful to include 15 them on there so that somebody subsequently 16 checking that same error message could see what 17 had happened in the past and what was -- and it 18 did say if any new problems come in with this is 19 symptom, it will need to be investigated again. 20 Q. So the sentence that we saw in the ballantj KEL 21 can't have included originally the 22 cross-reference to there is KEL -- 23 A. No -- 24 Q. -- because this KEL didn't exist at that time -- 25 A. No, of course not. 71 1 Q. -- when it was written? 2 Okay, we'll come back and look at this KEL 3 in detail at a moment. If we just go back to 4 the chronology, then, because this isn't raised 5 until September. I just want to see what had 6 happened in the interim. Can we look at 7 FUJ00081062, please. This is an email chain, 8 I think, all dated 6 May, certainly the part 9 that I wanted to refer to. If we look at the 10 bottom of the page, please. Thank you. If we 11 scroll up so we can see who it's from and to. 12 Thank you. 13 It's from Mr Jenkins to you on 6 May 2010, 14 yes? 15 A. Yes. 16 Q. Was subject line of "Receipts payments 17 mismatches"? 18 A. Yes. 19 Q. He says that he's noticed NT counter events 20 which look like receipts and payments 21 mismatches? 22 A. Yes. 23 Q. Yes? Why was he emailing you? 24 A. Because I was a useful person who would know 25 what was going on in SSC and could check whether 72 1 calls had been raised for them. 2 Q. Sorry, could check? 3 A. Whether a PEAK call had been raised for these 4 two instances. 5 Q. Why would Mr Jenkins contact you in particular, 6 rather than the other 24? 7 A. Because I was a helpful person. 8 Q. More helpful than anyone else? 9 A. Probably. 10 Q. Okay. He continues "Jon", and who is that? 11 A. Jon Hulme, who was, I think, in charge of the 12 counter development team at that point. 13 Q. "... that there were also raised from the Office 14 Snapshot erroneously ..." 15 I think should that read "that these were 16 also raised from the Office Snapshot 17 erroneously"? 18 A. Probably. 19 Q. "... but that PEAK [and a number is given] was 20 fixed in [a fixed code] which should be Live." 21 A. Yes. 22 Q. "Have you been made aware of these or had any 23 calls? I don't know if there is a KEL for SMC 24 to pick up any such events and raise calls -- 25 there certainly ought to be ..." 73 1 Can you help us, what is an NT counter 2 event? 3 A. When the counter application would check at 4 various points at the end of the balancing 5 process to make sure that receipts and payments 6 were equal and, if they weren't, it would flag 7 that in various ways. One of the ways it 8 flagged it was by creating an NT counter event, 9 which would be written to the application event 10 log, which was one of the files we were talking 11 about yesterday. 12 Actually, no, now we're on HNG-X, it was 13 very slightly different with the file that had 14 the events in, I think. But anyway, it's the 15 same sort of thing. And these events would have 16 gone from the counter through the Tivoli stream 17 to be -- hopefully to be monitored for and 18 checked by the SMC, whose job was to look for 19 these sort of events or any other unexpected 20 events. 21 Q. He, Mr Jenkins, says in his last line there that 22 he doesn't know if there's a KEL to pick up such 23 events and raise calls. Now, there was, of 24 course, a KEL. 25 A. Yes. 74 1 Q. We know that there was the KEL ballantj1759Q? 2 A. Yes. 3 Q. Why would Mr Jenkins not know about a KEL that 4 had been in existence, by my calculations, three 5 months by that time? 6 A. His job was not support. He didn't necessarily 7 use the KEL system. He wasn't responsible for 8 raising them or particularly using them. 9 Q. What was the Development team's access rights to 10 KELs? 11 A. He wasn't, strictly speaking, part of the 12 Development team but, yes, the Development team 13 had access to the KELs. 14 Q. What was Mr Jenkins's access rights to the KELs? 15 A. I don't know. I can't now remember if he did 16 have access to them or whether he -- it was just 17 easier to ask me, probably. 18 Q. He speaks, essentially, of a system being made 19 or needed to raise calls. What's that 20 a reference to? 21 A. Well, part of the process of looking out for 22 this type of error was that SMC would -- were 23 meant to be monitoring for this type of error, 24 and, if they saw one, then they should raise 25 a call -- it wasn't PowerHelp by then, but 75 1 whatever it was -- which would then get passed 2 on to PEAK for SSC to investigate. 3 Q. At the top of the page you reply, copying 4 Mr Parker in. You say: 5 "Gareth. 6 "... there is a KEL [then you give the 7 reference] which tells the SMC to raise a call 8 if they see this event. 9 "I haven't noticed any calls (but I haven't 10 been doing that sort of call recently). I do 11 have a PM-raised call from a few weeks back 12 which I need to look at (the mismatch was only 13 for a few pence so it has gone to the back of 14 the heap)." 15 Was there a heap -- 16 A. Yes. 17 Q. -- ie a mountain of unresolved systems issues 18 that you had to work your way through? 19 A. Er, yes, we were very, very busy at this time 20 during the HNG-X pilot. HNG-X was being used at 21 about -- I can't remember if it was 250 or 500 22 branches and, as you'd expect for any new 23 system, despite having gone through very 24 expensive testing, once you let several hundred 25 branches have a try, they found paths that 76 1 couldn't have been gone through during the test 2 process. 3 So I can't remember what other sort of call 4 I had been doing but, yes, I had been busy. The 5 postmaster-raised call, I think I say in my 6 witness statement, I shouldn't have left it that 7 long, even if it was only for a few pence, but 8 it would have been -- the effect on the branch 9 wasn't significant but it definitely needed 10 looking at and it hadn't just been closed down. 11 It was waiting. 12 Q. Were any of these receipts and payments 13 mismatches picked up by the reconciliation 14 process? 15 A. No, because the events were now being used 16 instead of the reconciliation process for this 17 specific type of error. 18 Q. But, on this occasion, it was a postmaster who 19 had raised the mismatch, not the NT events? 20 A. The call that was on my stack, which I have no 21 memory of now and haven't had sight of, was 22 raised by the postmaster, yes. 23 Q. Was that the case, that even though Fujitsu 24 systems were supposed to pick up things like 25 this, errors were often flagged for the first 77 1 time by a subpostmaster? 2 A. Um no, I don't think that is usual. I mean, 3 I don't know now whether there had been 4 a SMC-raised call for that call that was on my 5 stack which hadn't then been linked with it. 6 I haven't got that information. 7 Q. Can we move forwards, please, and look at PEAK 8 PCO203864, which is at FUJ00081586. 9 If you see, this was a PEAK raised on 10 2 September 2010 and it concerns a mismatch of 11 a smaller amount of money, £11.20. 12 A. Yes. 13 Q. Yes? 14 A. Yes. 15 Q. Can we turn, please, to page 2 and look at your 16 entry for 18.52.00? 17 A. Yes. 18 Q. You say: 19 "Joe, this is important because it means 20 that their accounts don't net to zero due to 21 some sort of system error -- not user error. 22 Similar to a receipts and payments mismatch. 23 Garrett had a call about a problem with 24 incomplete summaries recently, worth checking 25 whether that was the same branch." 78 1 A. Yes. 2 Q. What are you referring to there? 3 A. You mean the problem with incomplete summaries? 4 Q. Yes. 5 A. That was this reconciliation report which 6 reported on any branches where the day's 7 transactions didn't net to zero. So the branch, 8 if it was the same branch, they might have had 9 that problem one day and then, at some point in 10 the future when they did their balancing then -- 11 and produce their branch trading statement, then 12 this situation that this call is about with the 13 trading position not being zero would be 14 reported and I can't remember if that was on one 15 of the reconciliation reports or if it was 16 an event again. 17 Q. Did you think here is a version of the payments 18 and mismatch bug that we saw earlier in the year 19 doing its work again? 20 A. No, I don't think so because I thought -- we're 21 missing some evidence here. The earlier 22 problems, we know about the wrongly reported one 23 during the office snapshot. Nobody has shown me 24 the PEAKs that were subsequently raised for 25 those two events that Gareth reported. I am 79 1 absolutely certain that, him having flagged it 2 up, that would have been followed up on pretty 3 quickly. But we haven't got those calls for me 4 to look at to give you any explanation of. 5 So, as far as I was concerned, when I saw 6 this call coming in, I found it alarming. Not 7 because I knew there was already a problem in 8 this area but because it looked like there might 9 be something new. 10 At this point, September 2010, the rollout 11 of HNG-X to the entire estate was in progress. 12 I'm not sure how far through it had got but now, 13 instead of a few hundred branches, we are now 14 probably onto several thousand branches, with 15 the opportunity to find some new error paths, 16 and so on. So I was obviously concerned that, 17 yeah, we've got a problem here and it wasn't 18 because I knew of existing problems. I thought 19 it was quite likely that there was a new 20 problem. 21 Q. Can we go then to the KEL that we looked at 22 earlier, FUJ00081608. Looking at the top, we 23 can see that it was raised by Mr Wright on 24 23 September 2010 but was last updated by Cheryl 25 Card on 1 April 2016, both SSC members; is that 80 1 right? 2 A. Yes, and there have been ten versions of it. 3 Q. Yes. We're looking at the tenth version. It 4 describes the receipts and payments mismatch bug 5 rather well, so if we can just read it together 6 under "Symptoms": 7 "When a clerk balancing the stock unit the 8 rollover screen is eventually displayed, and the 9 clerk then presses the Preview or Print button 10 produce the Trial Balance ... The counter then 11 returns to the rollover screen. 12 "Having checked the report, the clerk then 13 presses the Rollover button, and in normal 14 circumstances is given the choice of rolling to 15 a new Balance Period or a new Trading Period. 16 "If the clerk chooses to roll to a new 17 [Trading Period], the net discrepancies are 18 present, then the system asks whether the clerk 19 wishes to transfer the net discrepancy to local 20 suspense, or else cancel the rollover ... 21 "If the clerk presses Cancel, the system 22 returns to the rollover screen and he/she can 23 press Print or Preview or Rollover or Cancel 24 back to the Stock Balancing menu." 25 Then there's a reference to another KEL. 81 1 If we read the solution at the foot of the 2 page. A reference data fix was released in 3 November 2010 under a PEAK, and the number is 4 given: 5 "Now that the fix has been deployed, if 6 Cancel is pressed on [number given] then the 7 discrepancy is not cleared. 8 "A Workaround (prior to fix): 9 "If the Clerk presses Cancel on [message 10 number given], then to avoid the bug they must 11 press Cancel again to return to the Stock 12 Balancing menu. 13 "Unfortunately the workaround cannot be done 14 after the problem has occurred at the office! 15 In this case the branch accounts will need to be 16 corrected. 17 "Please advise branches to continue rolling 18 over stock units and the office as normal. It 19 is not necessary to wait for the correction to 20 be applied before rolling into a new TP." 21 A. Yes. 22 Q. The workaround suggests, is this right, that 23 that was applied in the period before November 24 2010, before the fix was released? 25 A. The workaround is really just saying which 82 1 button the clerk would need to press to avoid 2 the problem. You didn't read through the 3 problem section on the screen, which is actually 4 where it describes the sequence of button 5 presses that got you into this situation. But, 6 yeah, the workaround was no good unless you were 7 very well aware of what was going to happen. 8 Q. So it's not really a workaround at all, is it? 9 A. No, no. 10 Q. Because it couldn't be done after the problem 11 had actually occurred? 12 A. No. 13 Q. So it's not a workaround at all? 14 A. It's not a workaround, no. 15 Q. That's because it would always be the case that 16 the problem would come to light after the 17 occurrence in the office? 18 A. Yes. 19 Q. So, is this right: until the fix was applied, 20 Fujitsu were relying on subpostmasters to call 21 in? That was essentially the only step that was 22 being taken? 23 A. Um -- 24 Q. There was nothing proactive done? 25 A. I cannot remember. I wasn't involved, but 83 1 I think, in all the various documents that we've 2 seen, there was a lot of talk with -- between 3 Fujitsu and Post Office as to how to sort this 4 out, to resolve any discrepancies. In fact, in 5 this case, the branch were losing their 6 discrepancies, so they made a loss. This 7 actually lost their loss. If they made a gain, 8 they lost that as well. 9 But I believe, but it's in a lot of this 10 other documentation somewhere, that steps were 11 taken by Fujitsu to find all occurrences of this 12 problem and then with Post Office to decide what 13 to do about them. 14 Q. What about this: as you rightly said, the 15 problem section of this KEL described a sequence 16 of button presses by a subpostmaster resulting 17 in this receipts and payments mismatch, yes? 18 A. Yes. 19 Q. What about sending a notice out to all 20 subpostmasters saying, "We've got a bug in our 21 system, don't cancel rollovers when you've got 22 a discrepancy because it will cause a receipts 23 and payments mismatch"? 24 A. Yes, um -- 25 Q. A bit like a sort of product safety recall or 84 1 a warning notice to everyone that's using 2 a system, "We've got something wrong with our 3 system. Don't do this, otherwise it will cause 4 an issue"? 5 A. That would have possibly caused more confusion 6 at 12,000 branches than the problems caused at 7 the -- I can't remember how many it was but 8 I think it was fewer than 100 that were actually 9 affected by the problem. But, yes, that would 10 be something to consider doing. But that would 11 be up to Post Office to communicate to the 12 branches. 13 Q. Were you ever aware of such a discussion 14 occurring in relation to this issue, this bug, 15 or any other bug, "Let's tell people" -- 16 relatively simple on this occasion -- "don't 17 cancel rollovers when you've got a discrepancy"? 18 A. That would cause more confusion because they 19 would not want to roll over with a discrepancy 20 that they disagreed with, so you would have to 21 word it very carefully and there was a way of 22 them cancelling -- it was just a very specific 23 point that they had to not continue to avoid -- 24 or, sorry, not cancel to avoid the problem. It 25 wasn't the only way they had of backing out to 85 1 of the process. 2 Sorry, to get to your question -- 3 Q. What about the broader issue -- 4 A. Yeah, um -- 5 Q. -- of accepting that there's a problem with the 6 system and telling the subpostmaster community 7 about it? 8 A. Yes, I cannot definitely remember. That would 9 not have been up to Fujitsu to make that 10 decision. We had no means of communicating 11 directly with all the subpostmasters. Post 12 Office could send messages that would appear on 13 the screen at the start of day but that was 14 totally within their control as to what they 15 were -- wanted to communicate with their 16 postmasters. 17 Q. I'm not suggesting, let me be clear, that this 18 should have been something that Fujitsu took on 19 itself to do. 20 A. Mm. 21 Q. It was a service provider to a client. I'm 22 asking whether you were aware in your 16 years 23 of ever any discussion about that occurring, 24 "Rather than correcting things behind the scenes 25 and not telling subpostmasters about them, we 86 1 actually say there's a bug in the system"? 2 A. Um, I can't remember. I wasn't usually involved 3 at discussions at that sort of level for 4 problems that would affect a significant number 5 of branches. 6 Q. In your witness statement, paragraph 54, you 7 say: 8 "I am asked whether there were any written 9 or unwritten practices, policies or procedures 10 to restrict what information about a bug or 11 potential bug could or would be shared with 12 others, either for limited periods or 13 indefinitely. I was not aware of any such. If 14 I spoke to a postmaster about a problem and 15 I identified it had been caused by system error, 16 I would say so." 17 Again, the revelation to a subpostmaster of 18 a system issue was dependent on you speaking 19 directly to the subpostmaster. 20 A. Yes. 21 Q. I think you told us earlier that that happened 22 very infrequently in your 16 years? 23 A. No, I said what happened very infrequently was 24 making corrections to the branch financial data. 25 I certainly would have spoken to postmasters 87 1 most weeks, perhaps not quite as often as that. 2 It would depend on the sort of calls that I was 3 handling, but, yes, it wasn't that unusual to 4 speak to a postmaster. 5 It wouldn't always be to say there was 6 a system problem because sometimes I would be 7 speaking to them for some other reason. 8 Q. Was there any guidance or policy on whether or 9 not you should reveal to subpostmasters system 10 faults with the Horizon System? 11 A. No, I was never given any guidance on that. 12 Q. It was a matter of individual discretion for 13 you? 14 A. Yes, but I and my colleagues certainly would 15 say -- I would hear them on the phone talking to 16 postmasters and I've seen quite a few PEAKs, and 17 so on, where it says, "Spoke to the postmaster, 18 explained it's a system problem". So that was 19 being done. 20 Q. Given that you have just said that you did it 21 and you were aware of other colleagues in SSC 22 sitting near or around you doing it, how did 23 that sit with what we discussed earlier: the 24 Post Office's reluctance to reveal system 25 errors, as you described it? 88 1 A. Yes, that seemed to be the policy that they took 2 on some of these bigger issues that were 3 affecting more branches. But within SSC, we 4 were never, ever trying to hide the fact that 5 there were system problems. 6 Q. Can we turn, please, to POL00028898. This is 7 PEAK 0204765. You'll see that it's opened on 8 25 September 2010. The summary is, having given 9 the branch code and a message number, "non-zero 10 trading position on office rollover". 11 If we look at page 2, please. Scroll down 12 to the entry for 15.16.30, an entry by your 13 colleague Cheryl Card. She says: 14 "The problem occurred on [15 September] when 15 stock unit 02 rolled over. This was originally 16 reported as per [the KEL that we've read] in 17 call [then a PEAK number is given], but for some 18 reason the call was closed without being 19 investigated. 20 "There is a known problem with the use of 21 the Cancel button during stock unit rollover. 22 This is fully described in [the other KEL we 23 looked at]. A fix is currently being worked 24 on." 25 A. Yes. 89 1 Q. Then if we go over the page, you'll see from the 2 second entry from the top, the call has been 3 assigned to Mr Jenkins on 27 September, for 4 advice on how to correct the branch accounts. 5 A. Apparently, yes. 6 Q. Can you assist, why was this still occurring? 7 A. Because the fix hadn't been made yet. 8 Q. If we go back to page 1, if we look at the call 9 status at the top, the "Priority" status at the 10 top, it's described three lines from the top on 11 the right-hand side as "Non-critical", yes? 12 A. Yes. 13 Q. If a PEAK was given this status, "C -- 14 Non-critical", was that taken into account in 15 a service level agreement with the Post Office 16 when working out penalty clause thresholds of 17 payments by Fujitsu to the Post Office? 18 A. I don't know. 19 Q. Were you aware of a service level agreement 20 which contained essentially liquidated damages 21 thresholds, depending on the status of calls as 22 between A, B and C? 23 A. I don't think so. Only in as far as I said 24 yesterday: I knew that some -- priority 25 financial calls did have to be done within 90 1 certain lengths of time to resolve the financial 2 side of it. But no, I mean this was presumably 3 raised as a C priority by the helpdesk, unless 4 anybody changed the priority subsequently. That 5 didn't mean that SSC wouldn't pick it up quickly 6 and investigate it. 7 Q. So the priority status didn't affect the 8 priority with which the SSC dealt with the PEAK? 9 A. No, not necessarily. 10 Q. What was the purpose of attributing a priority 11 status? 12 A. If it was a priority, it would definitely be 13 looked at quickly, but that doesn't mean that 14 the rest Cs went to the back of the heap, 15 necessarily. Obviously something with 16 a non-zero trading position would be looked at 17 fairly quickly, I would think. I can't see how 18 quickly it was given to Cheryl, unless you 19 scroll down. 20 Q. I think she first picked it up on the 27th. 21 A. Right. 22 Q. If we scroll to the second page. 23 A. Yeah. 24 Q. I think her first entry is on the 27th. Scroll 25 down, please. 91 1 A. That's so -- well, yes. Yes, so -- and without 2 knowing which day of the week it is and so on, 3 but yes. So it came into SSC on the 25th and 4 then the investigation started on the 27th, by 5 the look of it. 6 Q. If we go to the third page, please, and look at 7 the third entry down: 8 "The branch accounts will need to be 9 corrected. PEAK [and then a number] has been 10 sent to development for advice as to how to 11 correct the accounts." 12 Then do you see there's some text copied in 13 and, amongst other things, the severity given 14 there is as critical? 15 A. That's the severity of the event. These have 16 been -- these two entries are from the NT 17 events, which are being monitored centrally. 18 Q. Why might an NT event that has been attributed 19 the severity of "critical" be assigned priority 20 status C, of "non-critical"? 21 A. Because when the call was raised, it wasn't 22 actually raised for one of these events. 23 Q. Would not the fact that this NT event had been 24 recognised to be linked to the call that had 25 been made a cause of recategorisation of 92 1 priority? 2 A. It might have made sense for somebody to have 3 increased the priority of this bug from C, but 4 we can see, from all the other documents and 5 calls that were coming in with this problem, 6 there were a lot of people working on it by now. 7 It was not one little C priority call at the 8 back of the heap with people at the helpdesk 9 adding things to it and nobody looking at it. 10 The investigation was well under way. 11 Q. I'm going to skip over much of the 12 correspondence, documents and emails from 13 September, October and November 2010 concerning 14 the bug and its revelation to the Post Office, 15 because they mainly concern Mr Jenkins' actions. 16 But can we go, please, to a document from 17 mid-November 2010, FUJ00081214. This is 18 a series of emails. Can we start, please, with 19 the third page. Just look at the bottom of the 20 second page to see who it is from and to. 21 From Antonio Jamasb, a Post Office employee, 22 the branch IT service manager, to Saheed Salawu 23 in Fujitsu. Did you know him or her? 24 A. Him. He might have been Steve Parker's boss at 25 this point. 93 1 Q. Sorry, I missed that. 2 A. He might have been Steve Parker's boss at that 3 point, but I'm not at all sure. 4 Q. If we scroll up a little bit, we can see 5 Mr Salawu's sign-off block in his signature. So 6 you can see what his role was. 7 A. So Service Delivery Manager. So I was wrong 8 about him being Steve's boss. 9 Q. Anyway, going back to the text of Mr Jamasb's 10 email: 11 "... I have a conference call on Monday with 12 senior stakeholders within POL. 13 "I need a full update for Receipts and 14 Payments. 15 "I need: 16 "Up-to-date spreadsheet of branches affected 17 and what the discrepancy is. 18 "Up-to-date list of branches/counters yet to 19 have fix. 20 "Any calls logged with HSD re issue. 21 "A summary from Fujitsu stating why we have 22 no other integrity issues with Horizon, and why 23 we couldn't see this issue. 24 "Sorry to drop this on you." 25 In relation to the fourth of Mr Jamasb's 94 1 requests or demands, would you agree that that's 2 a reasonable question for a customer to ask 3 their contractor? 4 A. It would be a difficult question to answer, 5 possibly. I've no idea. I wasn't involved, as 6 far as I'm aware, in this investigation at the 7 time. 8 Q. But would you agree, stepping back, that this is 9 a reasonable reassurance for a customer to seek? 10 A. I think it's a reasonable reassurance for 11 a customer to seek. 12 Q. Asking "Please tell us, Fujitsu, why we have got 13 no other integrity issues with Horizon?" Part 14 one. 15 Part two: "Why was it we couldn't see this 16 issue?" 17 Can we go to -- scrolling up, please, 18 Mr Salawu forwards it to some others within 19 Fujitsu. Can you help who they were, Mike 20 Woolgar and Neneh Lowther? 21 A. I think they were other what's obviously called 22 a Service Delivery Manager at this point. 23 Q. Second paragraph: 24 "I know Mike was running with this but there 25 should be information that can answer the 95 1 queries. It's a good test of how effective our 2 update process works." 3 Then scrolling up still further. Mr Woolgar 4 emails Messrs Simpkins and Jenkins: 5 "... are you able to provide answers to the 6 questions from POL ... yesterday?" 7 He will deal with the third one and then go 8 to page 1, please. Mr Jenkins reply, he adds 9 Mark Wright in, who I think would have been your 10 manager by then or team leader. 11 A. He was my team leader, yeah. 12 Q. Mr Jenkins says: 13 "I think it is Mark from SSC that has been 14 running with this rather than John. 15 "Attached is an email he sent to POL with 16 an update yesterday. I think that addresses 17 points 1 and 2 ... 18 "As for point 4, then that is probably down 19 to me. In simple terms I don't think we can 20 make such a statement." 21 You'll remember what the request was, "Can 22 Fujitsu tell us why we have no other integrity 23 issues with Horizon and why couldn't we see this 24 issue?" 25 He says: 96 1 "I don't think we can make such 2 a statement." 3 He continues: 4 "What we can do is check through what known 5 integrity issues we have and also make the more 6 general statement that when integrity issues 7 arrive, then they do leave a trail enabling them 8 to be identified and their scope to be 9 ascertained. 10 "John/Mark: are you aware of any other 11 integrity issues we have not yet fixed? I can't 12 think of any off the top of my head." 13 At this time, would you have answered the 14 question in the same way as Mr Jenkins? That 15 you couldn't say that there are no integrity 16 issues with Horizon? 17 A. Yes, I don't think I -- I think I would have 18 answered it in the same way. I said earlier 19 I thought it was a very difficult question to 20 answer, and I -- yes, I would go along with what 21 he says. 22 Q. In relation to what else he said, would you 23 agree that the best that could possibly be said 24 was that there were, in fact, known integrity 25 issues with Horizon? 97 1 A. Well, it's to check through what known integrity 2 issues we have. 3 Q. Yes. 4 A. Yes. 5 Q. Would you agree that, when an integrity issue 6 shows itself, it leaves a trail? 7 A. Yes. 8 Q. So you would have answered this in the same way? 9 A. I think it was way above my pay grade to be 10 answering that type of question. 11 Q. So returning to the issues, then, it seems that 12 significant action was taken in relation to the 13 bug in September 2010; is that right? 14 A. As far as I know. I wasn't involved. 15 Q. You have, I think, answered my question already 16 that PEAKs had been raised from February 2010 17 onwards and your answer to the question "Why 18 wasn't action taken in relation to those PEAKs", 19 was that they concerned a different issue. 20 A. I think it's highly likely that they concerned 21 a different issue. We have not seen it, so 22 I cannot say definitively either way. 23 Q. But Mr Jenkins appears to have been aware of 24 this bug, a receipts and payments mismatch bug, 25 which caused the Windows NT 902 events from 98 1 May 2010? 2 A. No, he was aware of -- that there could be 3 various causes of receipts and payments bugs. 4 We haven't seen anything that links those two 5 that he flagged with the same -- with the 6 receipts and payments particular issue caused by 7 the "prev" button. 8 Q. In relation to the hundred or so branches that 9 you mentioned earlier -- 10 A. I've no idea how many it was. It would be -- 11 it's written down somewhere. 12 Q. To your knowledge, what action was taken 13 proactively to tell them of the existence of 14 this bug? 15 A. I don't know. 16 MR BEER: Thank you. Can we move to the Callendar 17 Square/Falkirk bug. 18 It might be a good opportunity to take 19 a break and reconvene at 1.50, sir? 20 SIR WYN WILLIAMS: By all means, yes, that's fine. 21 So 1.50, thank you very much. 22 MR BEER: Thank you very much. 23 (12.48 pm) 24 (The Short Adjournment) 25 (1.50 pm) 99 1 MR BEER: Sir, good afternoon can you see and hear 2 me? 3 SIR WYN WILLIAMS: Yes, I can, thank you. 4 MR BEER: Thank you very much. 5 Good afternoon; Mrs Chambers. Can we then 6 turn to the Callendar Square/Falkirk bug, Bug 2. 7 In very simple terms, an explanation of the bug 8 could be as follows, would you agree with it: 9 firstly, it was a big that afflicted Legacy 10 Horizon? 11 A. Yes. 12 Q. It started in about 2000? 13 A. Yes. 14 Q. It was caused by a lock in the Riposte software? 15 A. Yes. 16 Q. You give helpful information in slightly more 17 detail in your witness statement, which I'd ask 18 to be turned up, in paragraphs 73 and 74, which 19 are on pages 23 and 24. In 73 you say: 20 "Within the SSC we referred to the 21 underlying problem as the Riposte Lock problem." 22 That's instead of the Callendar Square or 23 Falkirk bug? 24 A. Yes. 25 Q. "Normally Riposte messages were automatically 100 1 replicated between counters so each counter held 2 an identical set of all transaction and 3 reference data relating to that branch. But 4 occasionally one counter would fail to accept 5 any messages from other counters. This usually 6 seemed to be triggered by something early in the 7 declaration or balancing process. Repeated 8 application events were generated which were not 9 visible to the user. The event storm and 10 failure to replicate ..." 11 Just stopping there, what do you mean by 12 "the event storm"? 13 A. The repeated application events. Every few 14 seconds the same event was generated and we 15 referred to that as a storm. 16 Q. "... would persist until the counter was 17 rebooted or ClearDesk was run?" 18 What was ClearDesk? 19 A. That was the process that ran at some points in 20 the early hours of the morning to restart the 21 counter application. 22 Q. Thank you. Then over the page to 74: 23 "The counter would still be able to serve 24 customers but would appear to be working 25 normally, but anything done on other counters 101 1 after the event started would not be visible. 2 Reports printed on the counter would not include 3 transactions done on other counters so those 4 transactions might be re-entered. Incorrect 5 discrepancies could be reported if the money was 6 in the till but the transactions weren't 7 included in the balance. Transfers between 8 stock units might be accepted in twice, causing 9 a discrepancy and a receipts and payments 10 mismatch. Single counter branches could not 11 have this problem." 12 That can come down, thank you. The issues 13 I would like to explore with you, so you know 14 where we're going, are firstly exploring the 15 explanation for what was done to address the bug 16 in the early 2000s; when Fujitsu was first aware 17 that the Riposte lock could cause a balancing 18 issue; who was aware of that issue; why it was 19 allowed to remain outstanding until 2006; and 20 did the fix, known as S90, work fully. 21 Okay? So if we can just go through the 22 chronology of those events and try to pick it up 23 as we do. 24 A. Yes. 25 Q. Again, this isn't a complete chronology, there 102 1 are about 15 documents that I want to ask you 2 about but there are a very large number of 3 additional documents and steps in the 4 chronology. Can we start, please, with 5 FUJ00017986. This is a PinICL, 00127251, and 6 you'll see that it was opened on 2 July 1998. 7 A. Yes. 8 Q. The opening summary at the top left-hand side is 9 "Riposte error: Failure to get lock". 10 A. Yes. 11 Q. I appreciate this before your time by couple to 12 years, in the SSC but, looking through this 13 PinICL, is it right that this appears to be 14 early evidence of the Riposte lock? 15 A. It is an instance of a single Riposte lock 16 error. There's no mention in there of repeated 17 events, which were the -- it was when you got 18 the repeated events that you might then also 19 have additional problems -- it might then affect 20 the replication. One single event, we never had 21 any evidence that that caused any long-term 22 problems. 23 I'd also say that I don't believe this is 24 anything to do with the counter software. 25 I don't think it was this process. 103 1 "B_LD_CD_DEL" looks like one of the bulk loaders 2 that would have been running on one of the 3 Horizon back-end systems. 4 Q. So following from that, if we look at the last 5 few entries on page 4, if we look at six lines 6 up, Mr Bell's entry: 7 "I have not seen this problem since the test 8 rig was updated to Riposte 216. 9 "Also the network has been changed so I'm 10 closing this call." 11 Does that tend to suggest that because the 12 problem was not seen as at 5 October 1998, the 13 PinICL was therefore closed as an isolated 14 example? 15 A. Yes, it was an example of a single Riposte error 16 which -- I can't tell from this but there's no 17 evidence that this one single error -- and, you 18 know, you do get errors and your systems have to 19 cope with this. There may be a good reason why 20 it failed to get the lock and it reported it. 21 So it's an instance of that particular error 22 message, and, yes, they didn't see any more of 23 it on the test rig, so they closed the call. 24 Q. When you and other SSC staff, much later or many 25 years later on, were investigating the extent of 104 1 the Riposte lock bug and the duration of it, 2 would this kind of PinICL have been available to 3 you? 4 A. I don't recall how long PinICLs were kept for. 5 It might have been there but, looking at that, 6 it bears absolutely no relevance to a counter 7 balancing problem found some years later. 8 Q. So if you had had access to it, you would have 9 dismissed it as irrelevant? 10 A. Yeah, as -- 11 Q. Can we move forwards then, please, to 12 FUJ00031913. We're now on 5 November 1999, so 13 again before your time, but it's another PinICL 14 opened concerning another Riposte lock, yes? 15 A. Yes, it is, which happened at a particular time 16 of day for a particular process. 17 Q. Reading through the PinICL, would you agree that 18 it appears that the SSC took no substantive 19 action in relation to the lock and, instead, it 20 was simply closed on 11 November 1999? 21 A. Assuming it's closed further down, it was 22 something -- ClearDesk failed to create training 23 object, that is the overnight processing 24 starting things up again, it's trying to create 25 something to do with the separate training 105 1 service that ran on the counter. It got, it 2 looks like, a single timeout message and also 3 this "error occurred" message. It doesn't show 4 any lasting ongoing problem. It wouldn't affect 5 replication in any way, so if it was just 6 closed, I'm not surprised. 7 Q. You're not surprised it was just open and closed 8 very quickly? 9 A. Yeah. 10 Q. Can we move forwards, then, to the year 2000, 11 FUJ00059049. You were in post by this time? 12 A. I'd been there about three weeks, I think. 13 Q. Yes. We'll see that this is a KEL raised by 14 Mr Ballantyne on 2 November and then closed by 15 you in 2005? 16 A. Not closed: last updated. 17 Q. Sorry, last updated by you. If we see the 18 "Solution", please, further down the page. 19 A. Yes, could I just say, sorry, before we go down, 20 this -- where they're getting the error messages 21 committing the discrepancies, and so on, this 22 suggests that the underlying problem which is 23 happening, it's not just preventing the 24 replication between counters; it's actually 25 a problem on this counter itself, where Riposte 106 1 is trying to write messages into the message 2 store on this particular counter and it's not 3 able to do so. 4 Q. So I'm not following for the moment what the 5 point is or the distinction you're making. 6 A. I think when I described the problem in my 7 witness statement I said the problem was the 8 failure to replicate messages from -- that were 9 being done on one counter onto this counter that 10 they were balancing on. 11 This particular description, which I don't 12 think I'd had the opportunity to remind myself 13 of when I wrote that, it's clear that they are 14 doing the work on this one counter and then, 15 when, in the balancing process, it's trying to 16 write the declarations and the discrepancies 17 into the message store, it's unable to do that, 18 presumably because this lock is held and so it 19 can't write into its own message store. 20 Q. So this could afflict a single counter; is that 21 what you're saying? 22 A. It possibly could, yes, but because it can't 23 commit or do any of these things, you're 24 actually not going to be able to complete your 25 balance process on the counter in the state that 107 1 it's in at the moment anyway. But yes, it -- 2 reading it now, I think it might affect a single 3 counter. 4 Q. I see. In terms of the solution, just reading 5 the "Solution" to yourself, is it right that in 6 essence the solution was to advise a restart by 7 the subpostmaster and to stop balancing if they 8 were doing the balance? 9 A. That is the solution that was being given at 10 that time, yes. 11 Q. How long was that solution to stop balancing and 12 restart? 13 A. That was for a long time because it took until 14 2006, I think, for us actually to get a fix for 15 the problem. 16 Q. So was that the operative advice for about 17 a six-year period? 18 A. Yes, it would have been. 19 Q. Can we look, please, at the PEAK to which this 20 KEL is associated, 0056922, and the PEAK is 21 FUJ00070841. Now, again, you're in post by this 22 time but for a very short period of time. I'm 23 not showing you the PEAK because you're 24 mentioned in it. It's another in the line of 25 documents evidencing the existence of the bug. 108 1 Can we look at page 4, please, at the foot 2 of the page, and Mark Jarosz's entry at 3 10.35.00, thank you. He says: 4 "My assessment of what happened is that on 5 Wednesday 1st Nov at 18.32.13 a lock was 6 acquired on the run table which was not 7 released. This had the subsequent effect of 8 causing [I think that's 'many'] Riposte API 9 calls to fail and hence the applications 10 connected to Riposte could not function 11 reliably. I would speculate that the probable 12 cause was a thread silently failing but we have 13 no way of proving this. 14 "I will check with Escher to confirm my 15 assessment is reasonable and if not further 16 update this PinICL. 17 "In the meantime I would recommend that in 18 future occurrences a restart of Riposte should 19 be attempted prior to rebooting NT. 20 "If the frequency of occurrence of such 21 an event becomes significant ([more than] 1 per 22 month) then we will need to create 23 a reproducible case." 24 Can you explain what you understand 25 Mr Jarosz to be saying there? 109 1 A. Jarosz. Some process has failed but it's not 2 being picked up, he doesn't know what that will 3 be. I don't think I can explain that very 4 clearly, technically. He was going to check 5 with Escher, who owned the Riposte code, because 6 this was the underlying product that Horizon was 7 built around, which Fujitsu didn't support, 8 and -- 9 Q. The last point I'm interested in particularly: 10 if the frequency of occurrence becomes 11 significant, which he defines as meaning more 12 than once per month, "we'll need to create 13 a reproducible case". 14 A. Yes, so if it keeps happening more than once 15 a month, then we're going to have to see if we 16 can reproduce the problem, which actually we 17 never managed to do. 18 Q. Who would monitor whether such incidents, 19 concerning the Riposte lock, were occurring at 20 a rate of greater than or less than once per 21 month? 22 A. I'm not sure that anybody was monitoring that at 23 that point. 24 Q. How would the proposal or conclusion or outcome 25 that Mr Jarosz has arrived at there be carried 110 1 into effect then, or wouldn't it? 2 A. I don't know. 3 Q. If nobody was monitoring the frequency with 4 which such events occurred, it couldn't be, 5 could it? 6 A. No, it couldn't. I don't know if any of my 7 colleagues back then, who were aware of this 8 problem, took it upon themselves to do such 9 monitoring but I certainly didn't do it, because 10 I hadn't been there very long and wasn't really 11 aware of this problem at that point. 12 Q. More generally, was there a system within the 13 SSC of logging disparate PinICLs and PEAKs 14 together so that some sort of meta analysis 15 could be carried out? 16 A. Not to my knowledge. I don't know if that's 17 something that the SSC manager did have any -- 18 did have any systems in place for, but I'm not 19 aware of that. 20 Q. Were you ever required or requested to 21 contribute to such a system, either in its 22 design or providing data to it? 23 A. There certainly were occasions when I would do 24 my own checks for similar calls happening. But 25 I don't recall it being something that I was 111 1 ever instructed to do. 2 Q. So there was nothing to stop one PEAK coming in 3 to one person in the SSC and another coming in 4 to one of the other 24 people in the SSC about 5 the same subject matter, and person 1 not 6 knowing about person 2's PEAK? 7 A. That could happen. I mean, what we did 8 sometimes do is, on the KELs say "Record further 9 instances here", so then we did get a bigger 10 picture but that wasn't part of the process that 11 anybody told us to do. 12 Q. It was a bit hit and miss, I think it's fair to 13 say? 14 A. It could be hit and miss, yes. 15 Q. So when one looks at the KELs, one doesn't see 16 a list of all the associated PEAKs? 17 A. You would on some KELs for particular problems. 18 Q. But not on many others? 19 A. Not on many others. 20 Q. Were you conscious of this within the SSC at the 21 time, thinking "I'm working my way through my 22 heap, my stack of tickets, I'm getting them in, 23 dealing with them, getting them out, and there 24 could be somebody else who's working a different 25 shift to me, somebody home working", I don't 112 1 know -- probably not home working in the SSC -- 2 A. (Unclear). 3 Q. -- but working a different shift, working on the 4 other side of the room, "and they could be 5 exploring precisely the same problem and we 6 don't know about it"? 7 A. That is possible. Certainly if it was the 8 same -- at the same point in time, we'd almost 9 certainly notice just because you would be 10 keeping an eye on the other calls that were open 11 and you would see if somebody had a similar 12 call, or the pre-scanner would say, "Ooh, I've 13 just given one that looks like this to so and 14 so", but if they were, you know, several weeks 15 apart you would not necessarily make those 16 links. 17 Having said that, in -- certainly, we did 18 talk to each other, and so we often did have 19 a pretty good idea of other things that were 20 happening so you would get some sort of an idea 21 of, you know, "Oh, there's another one of those 22 sort of problems", but it wasn't being formally 23 measured or managed. 24 Q. There was no system in place? 25 A. No. 113 1 Q. Can we look, please, at one of your KELs at 2 POL00030325. This was a KEL, "AChambers330S", 3 raised by you about a month into your time, yes? 4 A. Yes. 5 Q. On 27 November 2006, last updated by Mike 6 Croshaw on 20 October 2006. If you scroll down, 7 please, looking at, without reading them out, 8 the "Symptoms" and the "Problem", this looks 9 like another similar example of the Callendar 10 Square bug, doesn't it? 11 A. The KEL was originally raised for -- very 12 specifically just for a single occurrence of the 13 event at particular point in processing during 14 the LFS end of day processing, when it wouldn't 15 have affected -- that's not part of the counter 16 balancing process. Where it goes on to say 17 "sometimes a storm of these events occurs", that 18 that is later what we have called the Callendar 19 Square bug. 20 Q. The Riposte lock? 21 A. Yes. 22 Q. It said under "Solution" that: 23 "A single event can be ignored. 24 "Do not pass further instances to SSC unless 25 there appear to be side effects." 114 1 Why was the KEL signed off in this way? 2 A. Because a single event is just "Oh dear, Riposte 3 has had a slight problem, it's obviously 4 recovered from it", in that we only have the one 5 event, we haven't got ongoing problems. So if 6 there don't appear to be any side effects, then 7 it doesn't need any further investigation. 8 Q. Was there a concern that too many issues were 9 being sent up to the SSC? 10 A. I'm not aware of that particular statement in 11 that particular KEL causing that particular 12 problem. 13 Q. But the KEL is meant to discourage, isn't it, 14 passing instances up to the SSC? 15 A. If just single event has occurred. 16 Q. Wouldn't you want to know where single events 17 had occurred if they were occurring as single 18 events across the estate? 19 A. Not if they -- um ... 20 Q. Wouldn't that help identify the problem? 21 A. It might have done if there was a problem caused 22 by these events being raised. I realise we're 23 getting into a state where -- a chicken and egg 24 situation here. But, yes -- 25 Q. I'm going to ask you about the chicken and egg 115 1 situation in a moment. 2 A. Yes, I certainly wasn't trying to -- I don't 3 believe that KEL has been written in that way to 4 necessarily stop anything being sent to SSC for 5 any of these events. It was more written in the 6 first place for the single event, I believe. 7 Q. But this is addressed to HSH, isn't it, the 8 solution, to the Helpdesk? 9 A. Yes, it is, saying a single event can be 10 ignored. 11 Q. Was there sufficient skill and expertise within 12 the Helpdesk team to identify whether or not 13 a call related to a single event or was, in 14 fact, one of a series of events? 15 A. You could -- 16 Q. How would they know? 17 A. Because when -- if HSD or SMC were monitoring 18 the events, they would see each event coming in 19 as a separate entity from a specific branch. 20 Q. What a branch calling in? 21 A. No, this isn't a branch 'phoning in, this is the 22 automatic feed they get through Tivoli of the 23 events, the NT events that are being raised on 24 the counter. 25 Q. How would those lower levels of support identify 116 1 if there were what is described as "side 2 effects"? 3 A. Um, if further processes started raising other 4 events, if we're talking about the events, you 5 know, you might get one event saying, you know, 6 that a lock is held, and then other processes 7 might then generate events because they couldn't 8 do what they were meant to be doing. I mean 9 a lot of events were being raised all the time 10 from a lot of different processes, not that many 11 critical events. There were different 12 categories of events. But we certainly didn't 13 expect every single event being raised by the 14 system always to be individually investigated. 15 Q. Whether or not it required to be individually 16 investigated or not, wouldn't it be important 17 knowledge for the SSC to have, as to the 18 existence of these individual events, as you 19 called them? 20 A. Quite honestly, I don't think we would have been 21 able to do anything about them, except to look 22 at it and say "Well, we can't see that it's 23 caused any knock-on event on any other counter 24 process". The only thing that perhaps it might 25 then have had to go off to Escher again to say, 117 1 "Can you investigate why these are getting 2 these?" But you do get unexpected errors 3 happening on systems. Systems have to be 4 written in a way where they can cope with 5 unexpected errors at this sort of level. 6 Q. Some of the errors were causing what you've 7 described as side effects and some were not. 8 A. Um, I'm not sure if we ever -- I'm -- yes, it's 9 hard, very hard now, to go back and say -- you 10 know, if we had investigated a single event, we 11 would have had to have looked at the bigger 12 picture of which other processes had been 13 impacted. Certainly this one, just after 14 midnight, during LFS end-of-day processing, 15 it -- I think that one possibly did affect some 16 counters that night, sorry, but not the 17 balancing just the LFS process, which didn't 18 affect branch balancing in any way. I don't 19 think I can add much more here. 20 Q. Okay, can we look at version 2 of this KEL, 21 please, which is FUJ00059141? 22 A. Which version was this that we were looking at? 23 Q. This is version 1 we're looking at the moment. 24 If we scroll up, we can see the top of it. Now 25 we're looking at version 2. 118 1 A. Right, so that. 2 Q. Scroll up. That's version 1. 3 A. Version 1 must have been the update by Mike 4 Croshaw in 2006. 5 Q. Yes. 6 A. Which is confusing but the KEL system changed 7 I think at that point and that's why it went 8 back to version 1, so there would have been 9 earlier versions as well. I think. 10 Q. Okay, I'm not going to explore that any further. 11 Can we go to version 2 then. Again, it shows 12 correctly the first date that you raised this 13 KEL, 27 November 2000, and we can see that this 14 version 2 is last updated by your manager 15 Mr Parker on 14 June 2010. Can we take from 16 that that this is confirmation that these 17 lock-type problems were continuing to be 18 experienced throughout the time that Riposte was 19 in operation? 20 A. Firstly, I'd say I'm wondering if that's some 21 sort of administrative update, given that we 22 were just about to go live on to HNG-X in 2010 23 at that point, and go fully live. 24 Um, I think we very occasionally did 25 continue to see just a single event at odd 119 1 times, but not the event storms that were 2 happening during the balancing process, which we 3 saw up to 2006. 4 Q. Again, was there any method by which that data, 5 from which I've called a meta analysis, could be 6 conducted, was retained? 7 A. Well, all the events were retained, somewhere, 8 for some length of time. 9 Q. Yes. In a system but not being looked at. 10 A. Yes. 11 Q. I'm asking whether anyone -- where a bug was 12 identified like this, whether anyone in SSC or 13 elsewhere, for example in order to go to Escher 14 to say, "Look this is a continuing problem", 15 retained dataset that they could go to Escher 16 with and say, "This is a continuing problem it's 17 been going on for 10 years". 18 A. I know that in 2006, after the fix for the event 19 storms went live, I did monitor for some time 20 after that to see if the event storms stopped 21 and the event storms did stop, but there may 22 have been the very, very occasional single event 23 still happening. But it was the storms that 24 seemed to cause the message store either not to 25 replicate or it not to be possible to write to 120 1 them. Otherwise, if you've just got a single 2 problem the processes would retry and it would 3 work on the retry. 4 Q. Can we turn, please, to FUJ00083548. This is 5 an email exchange between Mr Jenkins and Mark 6 Jarosz and Brian Orzel. If we just read through 7 the message at the top of the page together, 8 from Mr Jenkins: 9 "I've had a look at this event log and 10 I don't think there's anything to really worry 11 about. Migration appears to have completed OK 12 and the outlet is running fine on CI4." 13 Was that a release, CI4? 14 A. Yes, I believe so. 15 Q. "I've seen number of 'verification failures' 16 during migration before and I believe that they 17 are to be expected during the various loads of 18 Riposte before the counter reboot. 19 "However, I'm curious as to why we get the 20 three errors mentioned in the PinICL. They 21 occur at 20.26 on 9/11/00. All are identical: 22 Facility MessageProcessor ... Error 94 'An error 23 occurred while attempting to destroy 24 a checkpoint run. Timeout occurred waiting for 25 lock'. They seem to occur during the Riposte 121 1 index rebuild immediately after the migration of 2 the 'real' message store. I assume that they 3 are benign, but would appreciate confirmation 4 from Mark before closing the PinICL." 5 I'm not going to look at the rest of the 6 message. For this kind of thing that Mr Jenkins 7 was doing, would you know that this kind of 8 thing was going on, ie the things that he is 9 describing? 10 A. Well, he's referring to a PEAK, 57478, and 11 mentioning three identical errors. I don't 12 think that's the PEAK that we were looking at 13 before. 14 Q. It's not, no. It's a separate one entirely. 15 A. Right, so I think this is a different problem 16 that we're looking at here. But a different 17 problem but, yes, again, you've got the same 18 underlying -- 19 Q. The same underlying cause? 20 A. -- event -- 21 Q. Yes. 22 A. No, not the same -- not necessarily the same 23 underlying cause but the same events have been 24 generated and in this time it's during the phase 25 of the index rebuild, and this is looking like 122 1 it's part of the migration. So when branches 2 are first moving on to Horizon. So, yes, some 3 Riposte errors were looked at. 4 Q. Can we move on, please, to FUJ00083574. Look at 5 the message in the middle of the page, again 6 from Mr Jenkins. Different PinICL, 57957. He 7 says: 8 "This PinICL is related to 56922 which you 9 looked at couple of weeks ago." 10 That's not the one that we were just looking 11 at. This is a third one. 12 A. This looks like we're back to a problem in the 13 LFS space, which is what that KEL of mine was 14 talking about. 15 Q. Yes, and he says: 16 "I've had a look through the message store 17 and the event log and have noticed that at the 18 time of this failure [12.02 am, essentially] 19 that there is an LFS background task running." 20 A. Mm. 21 Q. He says, next paragraph: 22 "I suspect that it is significant that the 23 Riposte error is 10 secs after the BLOB is 24 written ..." 25 What was the BLOB? 123 1 A. I think it was a Binary Large Object. 2 Q. What does that mean? 3 A. A big amount of data that was so big that it had 4 to be written into a whole set of messages. It 5 was too big to fit into one individual message. 6 Q. He then says, next paragraph: 7 "As the PinICL says, this seems to be 8 happening fairly frequently." 9 Yes? 10 A. That's what it says, yes. 11 Q. Next paragraph: 12 "I do think we need a definitive statement 13 from Drew ..." 14 Do you know who Drew was? 15 A. No. 16 Q. "... as to whether this event is benign or what 17 problems we could have when it happens. Could 18 it be due to an application error? Do we need 19 to get more info on when these problems occur. 20 It is clear that the circumstances in this case 21 are very different from those in the original 22 PinICL." 23 To your knowledge at this time, we're late 24 November 2000, was this problem happening fairly 25 frequently? 124 1 A. I don't know. I would say again, this isn't the 2 problem with the repeated events that was 3 affecting balancing. 4 Q. The reply at the top of the page, Mark Jarosz 5 says: 6 "From your description [Gareth] it sounds as 7 though we potentially have a recipe for 8 a reproducible case." 9 Can you assist us with what a reproducible 10 case is? 11 A. One where you could try the same process and 12 reproduce and make the problem happen in 13 a consistent way. 14 Q. What would be the purpose of forcing the problem 15 to occur? 16 A. Well, if you can reproduce it then you stand 17 a much better chance of, firstly, finding the 18 root cause and, secondly, being able to test to 19 show that you have fixed the problem. 20 Q. Can we move forwards, please, to FUJ00083583. 21 We can see Mark Jarosz's reply to Mr Jenkins of, 22 I think, 1 December 2000. It must be 23 an Americanised date, I think. 24 "Hi Gareth, 25 "I can confirm (having checked with Drew) 125 1 that a timeout of this sort is likely to be 2 benign in the sense that it should not result in 3 a message store corruption." 4 "At this stage, can you remember whether you 5 in the SSC were told about this investigation 6 having taken place? 7 A. I have no recollection of this and, again, we're 8 still talking about problems in the area of the 9 LFS agent, which is nothing to do with counter 10 balancing. 11 Q. So at the moment would you draw a distinction to 12 say this isn't really directly relevant to the 13 bug that we're looking at? 14 A. Yes. 15 Q. Therefore, you didn't need to know about it? 16 A. Personally, probably not, no. There was 17 an awful lot to learn about when I first 18 started, as you can imagine. 19 Q. Yes. Can we move forward a couple of years 20 then, please, and look at FUJ00083633. You'll 21 see that this is a PEAK, 0083563 opened on 22 7 November 2002 and can we look at your entry, 23 please, at 16.27.00, towards the bottom of the 24 page. 25 You say: 126 1 "Following critical event generated on 2 various FADs ..." 3 That's by various branches, is that right, 4 or about various branches? 5 A. Yes, at various branches, not raised by various 6 branches. 7 Q. Yes, about various branches, thank you: 8 "The call summary is now: 9 "Many 'run map' critical events on various 10 [branches]. 11 "Response: 12 "These events were investigated in the past 13 ..." 14 You give a PEAK number. 15 "But the call was closed on the basis that 16 the errors were no longer occurring. 17 "Analysis of the events in the last month 18 shows 2,132 of these events. In many cases 19 there is just one, or a small number on 20 effective counters, but [and then you give 21 a branch number] generated over 900 in one day, 22 and 191323 over 100 [presumably 100 days] ..." 23 A. No, that would be 100 events. 24 Q. Over 100 events? 25 A. Sorry, that's -- the "191323" is a branch code, 127 1 a FAD code. 2 Q. I've got it. So there are two branches you're 3 referring to? 4 A. Yes. 5 Q. The first one generated over 900 in one day? 6 A. Yeah. 7 Q. The second branch over 100 in one day? 8 A. Yes. 9 Q. Thank you. I understand. 10 Then if you look down the rest of the PEAK, 11 the trail seems to go cold, nothing happens. 12 A. Yes, I mean, I asked for the call to be raised 13 so that I could do some background investigation 14 into these events, which I had noticed, and 15 I was concerned about, because they were 16 happening. But there was a call that I sent to 17 Development or was sent to Development at around 18 the same time for them to investigate. 19 Q. But there appears to be a four-month gap between 20 your entries of 7 and 8 November, and your entry 21 of 24 February? 22 A. Yes. It was something I was doing as 23 a background task because I was concerned about 24 the events. 25 Q. But what was happening in the meantime to these 128 1 two branches that you mention here? 2 A. I can't remember now if I looked -- I mean, the 3 branch -- the events would only have -- they had 4 happened on those particular days. The events 5 didn't keep having the same -- they didn't keep 6 affecting the same branches over and over again, 7 and whether I did look back to see if the events 8 had had any impact on the branches on those days 9 or whether they had raised calls to say they 10 were having issues, I now cannot remember. 11 Q. But if you had have done that you would have 12 noted it on this PEAK, wouldn't you? 13 A. Um -- 14 Q. The branch you'd had at 900 events in one day 15 and the other branch you'd had over 100 events 16 in one day? 17 A. Um, yes. I mean these events didn't necessarily 18 mean that they did have balancing problems. It 19 just meant they could, in the majority of cases. 20 It didn't -- 21 Q. You don't know until you look, do you? 22 A. That's true. And I don't -- I cannot say now 23 whether I did look at those at that point. It 24 may be that, you know, at this point -- I'm not 25 sure when we first realised that it was tending 129 1 to happen more when they were doing their 2 balancing and therefore it might have an effect, 3 but, yes, it obviously was important to look 4 and, as I say in my witness statement, I'm not 5 at all happy about how this was handled over the 6 years. 7 Q. When you say that, you mean by Fujitsu? 8 A. Yes, by Fujitsu, by SSC. By all of us. We 9 could have done better. 10 Q. In relation to your part in that, you made 11 a record on 7 November that, in the last month, 12 there were 2,132 of these events and you 13 highlighted two branches, one where 900 events 14 had happened in one day and another where over 15 100 events had happened. If you had 16 investigated whether any of those events had 17 caused discrepancies, you would have noted it 18 down on this PEAK, wouldn't you? 19 A. Yes, if they had been aware that there were 20 discrepancies, that they had persisted through 21 the event storm and had managed to balance and 22 it looked as if they did have discrepancies as 23 a result, then I probably would have recorded 24 it. 25 If I had looked and seen that either they'd 130 1 realised they'd had a problem, had phoned for 2 advice and had rebooted and then done their 3 balancing, I probably wouldn't have recorded 4 that. But since I didn't record anything, 5 I don't know. 6 Q. Were you essentially relying on subpostmasters 7 to spot a problem and call it in? 8 A. I certainly think at this time, with this 9 problem, we did assume that they would notice 10 that -- either that they were getting error 11 messages, as we saw in one of the PEAKs you just 12 showed me, where they got error messages because 13 they couldn't commit the declarations or 14 whatever else. So they'd definitely know they 15 had a problem in that case. 16 In other cases, we certainly did get some 17 where they phoned in and they said "We're 18 balancing, our figures are all over the place", 19 and then the Helpdesk would advise them to 20 reboot and then it would be okay when they 21 restarted. 22 Q. Can we move forward again, please, to look at 23 FUJ00083651. This is a PEAK, 0086212, opened on 24 24 January 2003. Can we look, please, at your 25 entry on 29 January at 11.31, which is on the 131 1 bottom half of the page, tab. You say: 2 "It looks as if there was a problem with 3 last week's balance -- cutoffs and some final 4 OBCS transactions were done on counter 3., then 5 balancing continued on counter 4, but this did 6 not seem to know what had been done on counter 3 7 (there were many underlying Riposte timeout 8 messages). The transactions were ended again; 9 I need to ascertain whether they were sent to 10 TIP twice and whether the postmaster is out of 11 pocket. 12 "Have spoken to [postmaster] who confirms 13 there were problems and is worried that they may 14 continue this week. I'll contact her tomorrow 15 am to see how they have got on." 16 So there you're essentially describing, in 17 a single paragraph, the operation of the 18 Callendar Square bug, as I have summarised it, 19 in existence on 29 January 2003. 20 A. Yes. 21 Q. Then if we can look at your entry a couple of 22 days on, 31 January at 16.09, you say: 23 "[Postmaster] balanced okay. She has 24 reversed the transactions which she had had to 25 reenter (the original ones were included in the 132 1 new CAP). This was all caused by counter 4 2 being unable to see messages recently written on 3 counter 3 when the stock unit was being 4 balanced. 5 "There is no accounting discrepancy here, 6 but there is a problem in that the [postmaster] 7 was allowed to balance with no warning that the 8 counters weren't communicating. MSU informed 9 that I'm sending this to [Development] for 10 further investigation." 11 A. Yes. 12 Q. So would you have informed the MSU here? 13 A. Yes, I would have informed MSU because this call 14 was not raised by the postmaster but because of 15 an entry on one of the reconciliation reports. 16 At this point in time, the cash account was -- 17 so when the branch balanced, they did their 18 balance reports and then they produced a cash 19 account on the counter, but the cash account was 20 also reconstructed at the data centre from the 21 transactions that had been harvested -- had 22 reached the data centre, and a comparison was 23 done, line by line, to make sure the two were in 24 step. And in this case they weren't because the 25 data centre, when it did its recalculation, knew 133 1 about these counter 3 transactions, which hadn't 2 been picked up when the branch cash account was 3 produced. 4 So there was a mismatch, and 5 a reconciliation call was raised to investigate 6 why that had happened. 7 So in this case, that's how we knew about 8 the problem. I did phone the branch to see what 9 had happened, you know, whether they had 10 realised there was a problem. They had put 11 these transactions in again, because they didn't 12 think they'd had them once, but then, because 13 the original transactions hadn't been included 14 in the accounts for the period that they'd just 15 balanced, they automatically got carried forward 16 and then were picked up in the new period. 17 So, in order to avoid them going through 18 twice, she was then able to do a reversal on 19 them, which sorted out her branch accounts. But 20 obviously, yes, there was still a system 21 problem. 22 Q. That's why you sent it to Development for 23 further investigation? 24 A. So I sent it to Development for further 25 investigation. 134 1 Q. If we just go forward to complete the story on 2 this PEAK to page 4. Two years on, there's 3 a record at the top of the page, for 5 October 4 2005: 5 "This call is one of a set approved by ... 6 (Mik Peach) for closure without further action." 7 Was that because the fix was then thought to 8 be the S90 release? 9 A. I doubt it. I don't know. I'm not very happy 10 with that. 11 Q. Do you know -- 12 A. Was there anything above that at all to say if 13 anybody had looked at it? 14 Q. You can look back, please, at page 2. I don't 15 think there's anything relevant in your entry on 16 3 February. 17 A. No. 18 Q. If you scroll to the bottom half of the page, 19 you'll see two customer calls. If you just read 20 those. 21 A. Oh, I think that just happened automatically 22 when -- was that when we moved from PinICL to 23 PEAK, or something changed and all the calls had 24 to be closed and everyone reopened and then 25 everything got written in again. 135 1 Q. So nothing of substance there? 2 A. No nothing of substance. 3 Q. Then on to page 3? 4 A. No -- 5 Q. Nothing there? 6 A. Yes, nothing there. 7 Q. Then bottom half of the page. 8 A. No. 9 Q. Nothing there? 10 A. No. 11 Q. It's just closed off, isn't it? 12 A. It was just closed off. 13 Q. Why are you unhappy or not very happy? 14 A. I'm unhappy with myself because I should have 15 made it something more than a C priority. 16 I left it at the priority that it had come from 17 MSU in the first place, and I should have 18 shouted a lot louder about the fact that this 19 needed looking at. As time went by, I got 20 better at shouting louder. 21 Q. Who would you shout to? 22 A. Oh, anybody who'd listen. 23 Q. Meaning who: Mr Peach? 24 A. Yeah, Mr Peach, Development team. 25 Q. So, in essence, this call is closed off without 136 1 anything having been done on it for two years? 2 A. Yes. 3 Q. One way of describing that is suboptimal, isn't 4 it? 5 A. You could say that. 6 Q. Can we look at a different PEAK, please. 7 POL00000996, this is PEAK 0103864, opened on 8 3 June 2004. The "Summary": 9 "[Postmaster] reports that he had a problem 10 with some transfers." 11 A. Mm-hm. 12 Q. I think, without going into the detail, the PEAK 13 describes the problem where multiple transfers 14 in occurred to a stock unit? 15 A. Yes. 16 Q. Can we go forward to page 6, please. In the 17 middle of the page, 6 July at 11.47.27, you've 18 written: 19 "I've checked with Mike King; the BIMS 20 report for this problem was sent to POL on 21 [22 June] and should have result in an error 22 notice being sent to the branch. Mike says he 23 will send a note to POL saying that the 24 [postmaster] has been changing this issue; I've 25 asked [the helpdesk] to inform the [postmaster] 137 1 that they should have received an error notice 2 and to check with the department that issues 3 them. 4 "The corrected cash account that was sent 5 still had a [receipts and payments] mismatch. 6 The double Transfer In causes a mismatch both 7 because of the transfer and because of the 8 discrepancy which has been erroneously 9 generated. The host-calculated [cash account] 10 ignores the transfer but is still affected by 11 the accepted discrepancy which should not have 12 been generated. It is not really possible to 13 provide a fully balanced [cash account] ..." 14 There's an email on this subject. 15 A. I think we discussed this one yesterday. 16 Q. Yes. You're recorded as dealing with the BIMS 17 here on 6 July 2004. Can you explain, please, 18 again, exactly what you're doing there? 19 A. Yes. I wasn't personally involved with the 20 BIMS. I checked with Mike, who was the person 21 who had sent the report to Post Office about the 22 discrepancy that shouldn't have happened at this 23 branch. 24 Q. So are you just identifying that there's been 25 a delay here? 138 1 A. Yes, I think so. I think further up the call, 2 Catherine had been dealing with it. She had 3 sent the information to Mike in MSU informing 4 Post Office of this discrepancy that the branch 5 should not be held liable for. The branch had 6 not held anything. They were chasing it back, 7 so the call ended up with me, so I followed it 8 up, as best I could, with Mike. 9 Q. Again, could we go forwards in the tale, please, 10 to 2005 -- sorry, 2006, and look at POL00030241. 11 This is a chain of email correspondence on 12 Callendar Square itself -- 13 A. Yes. 14 Q. -- once the bug had seemingly been identified 15 and discovered within the Post Office. We 16 should just set the context by starting at the 17 foot of page 9 and on to page 10, with an email 18 from Sandra MacKay of POL, to Brian Potter of 19 POL. This is just to set the context for what 20 happens later in the chain. Can you see -- I've 21 said that's to Brian Potter. If we just go up, 22 it's Shaun Turner copied to Brian Potter. Just 23 trying to work out -- ah, yes. 24 Shaun says: 25 "Gary, 139 1 "Need your advice on this branch. There 2 appears to be an ongoing problem at this branch 3 with transfers between [stock units] causing 4 a receipts and payments mismatch. This first 5 came to my attention some 3 or 4 months ago, 6 when the branch was chasing up an error notice 7 to account for a loss that they had that was 8 related to the [receipts and payments] mismatch. 9 I believe in that case, that FS [Fujitsu, 10 I think] had taken it on board and were 11 investigating it as a problem (I seem to recall 12 it had a PinICL number). I had to do some 13 chasing around with P&BA [Products and Branch 14 Accounting] to ensure that the error notice got 15 issued, as there was a breakdown in processes 16 between them and FS relating to the BIM report. 17 "Since then it appears to have happened 18 again, although Fujitsu are saying no issue 19 could be detected. I am concerned that there is 20 a fundamental flaw with the branches 21 configuration, and would be interested to know 22 how [Fujitsu] put the ... issue to bed." 23 If we go further up on page 9 and on to 24 page 8, and if we just scroll gently up, we can 25 see that this gets passed around essentially 140 1 within POL, yes? 2 A. Liz Evans-Jones was Fujitsu. 3 Q. So it goes over to Liz Evans-Jones from 4 15 February 2006, from Gary Blackburn: 5 "Liz 6 "I have had the incident below forwarded to 7 myself by our Service Line ... could you please 8 update me on the corrective action plan as this 9 still appears to be occurring within the 10 branch." 11 Then if we go further up to page 8 12 Ms Evans-Jones replies: 13 "I have checked the call and this issue 14 appears to have been resolved in S90." 15 Could you, in a word or two, explain what 16 S90 was? 17 A. It was a fairly major release of updated 18 software. I can't remember what functionality 19 it included, but there were new areas of 20 functionality coming in fairly frequently and 21 so, as a part of that, there would be some bug 22 fixes and this was scheduled to be one of them. 23 Q. So a scheduled software release? 24 A. A scheduled major software release, which would 25 have been through a very thorough test cycle. 141 1 Q. It was proposed to include this fix within S90? 2 A. Yes. 3 Q. "S90 has already been deployed to the Datacentre 4 and counter release is scheduled to start", and 5 then there are some dates: 6 "3rd line support has been discussing with 7 the [postmaster], and the last contact with the 8 branch (according to PowerHelp) was on 1st Feb. 9 The call has been left open for 3rd line to 10 check to see if the issue reoccurs S90. 11 Please let me know if I can provide any 12 other assistance ..." 13 Then continue scrolling up, please, and 14 again, please. We can then see some passing, 15 essentially back down the chain, of Liz 16 Evans-Jones' reply within POL, yes? 17 A. Yes. 18 Q. On page 7, keep scrolling up. Thank you. If we 19 then just scroll down a bit so we can see it is 20 Mr Turner asking these questions. Thank you. 21 Shaun Turner, yes, and then just scroll back up. 22 Thank you. 23 He says, back to Mr Blackburn: 24 "Gary, 25 "Thanks for looking into this for us. 142 1 Couple of questions occur: 2 "Do we understand why this particular branch 3 has been having problems? Or are there other 4 branches in the network that have been having 5 this problem? 6 "Could this branch be front ended on the 7 counted release of S90 such that it gets the fix 8 as soon as possible? 9 "The email from Liz suggests that there may 10 be a reoccurrence following S90. What degree of 11 certainty do we have that it will definitely be 12 fixed?" 13 So some pretty direct and pertinent 14 questions there from Mr Turner, yes? 15 A. Yes. 16 "Sandra/Brian -- Appreciate this is 17 frustrating for the branch but from the email 18 below you can see that the branch's issue should 19 be fixed for the release of the S90 software. 20 I have asked Gary above to see if we can put 21 this branch to the front of the queue ... In the 22 meantime it is important that the branch 23 continues to report any issues into [the 24 Helpdesk]." 25 So the four or five rather pertinent 143 1 questions that Mr Turner asks, if we scroll up, 2 please. We then see these getting passed around 3 within Fujitsu. Keep scrolling, please. Keep 4 scrolling -- keep scrolling. 5 We can see an email from Mike Stewart to 6 Mr Simpkins, copied to you. Why were you copied 7 in? 8 A. I think because Mike says below, "As Anne is 9 away could I have your comments as you were 10 involved as well." 11 Q. That's saying why John is asked the questions, 12 but why is that addressed to you? 13 A. Well, before that, it's "Anne, you're always 14 a good place to start", so it was me being 15 a good place to start. 16 Q. I'm so sorry, why were you the good place to 17 start? 18 A. Because I knew what it was going on and because 19 I had to put an update on that call there that's 20 at the very bottom of the screen, so I'd 21 obviously had some involvement -- 22 Q. With this bug? 23 A. -- with this bug, as far as I can see. 24 Q. If we scroll down we can see what Mr Stewart 25 asked. He cuts in his explanation of the 144 1 position, yes? 2 A. He's pasted in my update from PinICL there. And 3 then, where it says, "I notice", that, 4 I believe, is his words from that point. 5 Q. Yes: 6 "I notice that in the early guise of this 7 problem in the call it states the PM as female." 8 Yes? 9 A. Yes, that's what Mike's saying. 10 Q. Then some more cutting and pasting. 11 A. Yes, then -- 12 Q. Then back to him -- 13 A. -- that's what the helpdesk had put on the call, 14 yes. 15 Q. "At the bottom of this email re a magical 16 [£43,000] appearing and disappearing the PM is 17 Male He reports", et cetera. 18 A. Mm-hm. 19 Q. Then scroll down, please. He says: 20 "Clearly the [subpostmaster] is concerned, 21 as we have just spent a number of months trying 22 to sort out the first instance and he doesn't 23 want a repeat performance. He is convinced that 24 there is something wrong with his Horizon kit. 25 I would be grateful if you could investigate 145 1 this and give him any support you can. I'm due 2 to visit the office tomorrow to look at his 3 paperwork and discuss the situation ..." 4 A. Again, isn't that again a cut and paste from 5 something that somebody in the Post Office had 6 said earlier in the chain? 7 Q. I'm not sure that it is, I thought that was -- 8 A. I don't think Mike would have been visiting 9 offices. 10 Q. No. I think you're probably right, then. Can 11 we go to your answer then, please, on page 3 of 12 this email chain. At the foot of the page. You 13 respond: 14 "Mike ... 15 "Haven't looked at the recent evidence, but 16 I know in the past this site had hit this 17 Riposte lock problem 2 or 3 times within a few 18 weeks. This problem has been around for years 19 and affects a number of sites most weeks ..." 20 Is it right that this Riposte problem had 21 been around for years? 22 A. Yes, because we had been seeing it since at 23 least the end of 2000. 24 Q. So five and a half, six years? 25 A. Yeah. 146 1 Q. Your witness statement -- there's no need to 2 turn it up, I don't want to disrupt the 3 narrative -- says at paragraph 76, "[You] 4 personally had known about this Riposte lock 5 problem since soon after I arrived at the SSC in 6 2000", and that's a reference back to those 7 November 2000 PEAKs we looked at; is that right? 8 A. Yes, although I think that when I wrote the 9 witness statement, I'm not sure that I'd 10 actually seen the dates on those but, yes, it 11 was a known problem. 12 Q. So that reflects the early PEAKs that we saw of 13 November 2000 -- 14 A. I believe so. 15 Q. -- and an early KEL that we saw of November 16 2000? 17 A. Yes. 18 Q. In relation to your part of the sentence which 19 says, "This affects a number of sites most 20 weeks", how did you know that it affects most 21 sites -- sorry, it affects number of sites most 22 weeks? 23 A. Because of the event storms that we could see. 24 I would say that I think there -- it was 25 something that did seem to have increased, as 147 1 time went by. I don't believe we were seeing 2 all these event storms several times affecting 3 many branches all the way through, although, 4 actually, since I wasn't necessarily checking, 5 I don't know. But when say it affects them, 6 I mean that we could see, if we looked, that 7 event storms were happening. It does not mean 8 that it necessarily had any impact on their 9 branch accounts. I'm not saying that every week 10 a number of sites were having incorrect 11 discrepancies because of this problem. 12 Q. But you didn't know one way or the other? 13 A. We would have known -- okay, it's slightly 14 peripheral. Some aspects, including the 15 transfer problem, and the rolling over without 16 the transactions included, would have caused 17 entries on the reconciliation reports, certainly 18 up to 2005. So that would -- those would all 19 have been investigated as they happened. And 20 I'm certainly not aware -- I don't remember that 21 every week we were having a branch or two with 22 those reconciliation report entries -- 23 Q. But, Mrs Chambers, this is a problem that's been 24 going on for five or six years. 25 A. Yes, and if we had been getting all those 148 1 reconciliation report entries at that sort of 2 level for five years, it would absolutely 3 definitely have been picked up and seen as being 4 a big ongoing problem. 5 Q. But isn't that relying on subpostmasters -- to 6 an extent, relying on subpostmasters calling in? 7 A. Not the reconciliation report entry reports. 8 Q. No, the other part of the answer that you 9 gave -- 10 A. The other part of them phoning and saying, "Oh, 11 I'm doing my balancing and it's all gone 12 haywire"? Yes, for us to know about it, they 13 would have had to phone in and that call would 14 have had to be passed to SSC and it's quite 15 possible/probable, that the majority of those 16 calls were not passed to SSC because they were 17 just being told to reboot. 18 Q. Yes. Did you know how many sites were affected 19 every week? 20 A. No, I could have known if somebody had asked me 21 to monitor that and, obviously, at the point 22 that I was doing some analysis, then I did know. 23 And I think I found -- I can't remember what 24 period it was, in that previous KEL we looked 25 at -- previous PEAK we looked at, where I did do 149 1 some monitoring and I'd found two branches with 2 the event storms. But, as I say, I can't now 3 remember what length of time that was over. 4 Q. You continue in that sentence: 5 "... and finally Escher say they have done 6 something about it." 7 A. Yes. 8 Q. In your witness statement, you say that: 9 "I and others in the SSC understood the 10 cause of the problem was to be a problem in the 11 Riposte software, which we thought was being 12 investigated by Escher." 13 Of the five or six-year period that we're 14 looking at, for how long had you thought that 15 Escher had been investigating the issue? 16 A. I thought they'd known about it all the time. 17 I now think -- well, I now know, since putting 18 all the calls together, and so on, that that's 19 extremely unlikely. 20 Q. What had gone wrong? 21 A. Um, it was -- nobody was managing it as 22 a problem. It was almost impossible for SSC 23 staff to see which calls were with Escher and 24 who was progressing those, because they sort of 25 went on to a separate PEAK stack, which I now 150 1 know, yes, we could see but I don't think at the 2 early days I knew quite where it was. And it 3 wasn't SSC's job, really, to be monitoring 4 those, but I'm not sure whose job it was. 5 Yeah, I think if we'd appreciated that 6 nobody effectively was looking at this for all 7 that time, we would have flagged it up and 8 jumped up and down. But that realisation just 9 didn't come until late in the day when, 10 finally -- you know, we did send a call over. 11 It did get picked up, eventually, and sent to 12 Escher and they did produce a fix. 13 Q. You say you were interested in whether they have 14 really fixed it. 15 A. Oh, I never believe anything anybody tells me. 16 Q. Was that -- 17 A. You check; you double check; you triple check. 18 Q. Was that more directed to what you knew about 19 Escher, rather than being cynical about the 20 world in general? 21 A. No, I was cynical about the world in general. 22 Q. You therefore left the call open, the PEAK open? 23 A. Yeah. 24 Q. If we turn up paragraph 81 of your witness 25 statement, please, on page 26. Paragraph 81 at 151 1 the top. You say: 2 "I am asked whether Post Office or 3 subpostmasters were told about the problem. It 4 was not raised as a wider problem with Post 5 Office; each instance was treated individually." 6 Does that mean, until that email chain that 7 we picked up a moment ago, that Post Office was 8 kept in the dark for the best part of six years. 9 A. I'm not sure they were intentionally kept in the 10 dark but I think they were in the dark, yes. 11 Q. You say each instance was treated individually. 12 Why was each case treated individually? 13 A. Because we would look at the calls that did come 14 through, where they came through to us, and if 15 there was an effect on the branch accounts, then 16 we would pass the information via MSU to Post 17 Office on a BIMS report or it was passed that 18 way. 19 Q. This had been a problem that had been around for 20 five or six years. Would you accept that 21 Fujitsu had failed properly to investigate and 22 address the bug? 23 A. Yes. 24 Q. And failed to tell -- I'm not saying that you 25 personally failed to tell -- the Post Office 152 1 about its existence? 2 A. Yes, I think because it was always treated as 3 individual instances, it wasn't raised as 4 a problem and flagged through to Post Office. 5 Q. You said in your email that you wanted to wait 6 and see to see whether the S90 release was 7 an effective fix. Was that a case of waiting to 8 see whether any more calls came in from 9 subpostmasters and, if so, whether any of those 10 calls could be linked to the Riposte lock 11 problem? 12 A. No, I monitored the events coming through the 13 system to see if there were any more of these 14 event storms occurring anywhere. 15 Q. To complete the loop, if we may look at a last 16 couple of documents. FUJ00083667. This is 17 a PEAK, 0127246, opened on 12 October 2005. If 18 we look at page 3 and the entry for 11.14.22 at 19 the top of the page. You say: 20 "Since [the release] S90 was distributed, 21 the number of these timeout events over the 22 whole estate has gone right down, with no storms 23 from an individual counter. So it looks as if 24 the Riposte change has been effective." 25 A. Yes. 153 1 Q. Was that your measure of working out whether the 2 fix had been effective? 3 A. Yes. 4 Q. Was there any other means of working out whether 5 the fix had been effective? 6 A. I think that was a reasonable check of whether 7 the fix had done what was expected of it. 8 Obviously if, after this, we got more calls 9 about balancing issues, failure to replicate 10 across counters, and so on, then they would be 11 investigated from scratch again. 12 Q. Lastly, can we look at an email exchange, 13 please, between you and Mr Jenkins from 2010. 14 That's FUJ00083722. 15 If we look at the email at the foot of the 16 page, please. Can you see your email, from you 17 to Mr Jenkins? 18 A. Yes. 19 Q. You forward, I think, if you look underneath it, 20 the chain from back in 2006 that we see? 21 A. Yes. 22 Q. Why were you doing this? 23 A. I presume he had contacted me and asked me what 24 I could remember about the Callendar Square 25 problem. 154 1 Q. Can you remember why, in February 2010, 2 Mr Jenkins would be contacting you about what 3 had happened to the Callendar Square bug back in 4 2006? 5 A. I -- no, I don't remember specifically why. It 6 might have been because he was involved in 7 a prosecution and wanted to know some of the 8 details in case this was raised. I did my best 9 to provide some information for him. 10 Q. So let's see what you told him, please. If we 11 just scroll up, please. You say, "Gareth", and 12 you give a reference to a KEL: 13 "I'd forgotten -- this did give 14 a discrepancy, but also a receipts and payments 15 mismatch, if they persisted and rolled over 16 (though it was usually obvious that something 17 was wrong). 18 "And a flood of NT events (not 'Riposte 19 events'!) which SMC should have noticed at the 20 time. 21 "Since we are now checking for these 22 particular events, and did a catch-up for old 23 retrievals, can you say that the current branch 24 did not have a problem??" 25 What were you referring to when you said 155 1 "the current branch" -- 2 A. That suggests he was asking me about some 3 specific branch and could we see if they had had 4 the Callendar Square problem or not. But I now 5 don't know which the "current branch" would have 6 been. 7 Q. Would you interpret that as meaning "Can you 8 say, in the evidence that you are to give, that 9 the current branch did not have this problem?" 10 A. I cannot remember precisely why he was asking me 11 but that is a possibility. 12 Q. You say: 13 "Anyway it stopped happening once S90 was 14 installed ... 15 "This particular problem would only affect 16 branches with more than one stock unit. It 17 happened several times at Callendar Square, 18 though we never found why they were so badly 19 affect. 20 "Is this sufficient?" 21 The line, "This particular problem would 22 only affect branches with more than one stock 23 unit"; was that accurate? 24 A. That was accurate if we were talking about the 25 Callendar Square part of the Riposte lock 156 1 problem. Callendar Square had had the very 2 specific problem of being able to do the 3 transfer ins multiple times, so that's what 4 I was referring to there. You could only do 5 transfers out and in if you had more than one 6 stock unit. There may well have been a separate 7 email or two or discussion behind this specific 8 email but I cannot remember now. 9 Q. Then scrolling up to see what Mr Jenkins' reply 10 was: 11 "Thanks Anne, 12 "Penny ..." 13 Is that Ms Thomas? 14 A. I would think so, yes. 15 Q. "... pointed out on Friday that [the Post 16 Office] have not asked us to retrieve any data 17 for this branch yet! Therefore we have no 18 message stores to check against Event Logs. 19 "This will probably do me more now." 20 Does the case of Seema Misra ring any bells 21 with you? 22 A. I've obviously heard the name, yes. 23 Q. Is that recently seen the name or something you 24 recall being asked about back in 2010 and 25 Mr Jenkins' preparation to give evidence in the 157 1 Seema Misra case? 2 A. I think it's likely I had some involvement back 3 then but I cannot remember definitely which case 4 that was. 5 Q. Can you remember the context of this exchange, 6 ie he was asking you for what he should say or 7 things to say by reference to your work and your 8 understanding of the issue in the Seema Misra 9 case? 10 A. I thought he was just, sort of, checking with me 11 just to see what I knew in general about the 12 problem, and so on, what my recollection of it 13 was. 14 MR BEER: Thank you. Mrs Chambers, they're the only 15 questions I ask you at the moment. I'm going to 16 have to draw a line under Bug 2 and ask you 17 about the remainder of things when you come back 18 in the future for your Phase 4 evidence. I need 19 to allow reasonable amount of time for other 20 people to ask their questions. Thank you very 21 much for the evidence you've given. 22 Sir, might that be an appropriate moment to 23 take a ten-minute break for the transcriber -- 24 15-minute break, she's mouthing to me -- until 25 3.35? 158 1 SIR WYN WILLIAMS: Well, on the basis of the 2 guesstimates that you were provided by your 3 colleagues yesterday, that should provide 4 sufficient time to finish by around about 4.15? 5 MR BEER: Yes, I think so. 6 SIR WYN WILLIAMS: All right. 7 MR BEER: Thank you very much. 8 (3.20 pm) 9 (A short break) 10 (3.34 pm) 11 MR BEER: Sir, thank you. Can you see and hear us? 12 SIR WYN WILLIAMS: Yes, I can. 13 MR BEER: Thank you very much. I think Mr Stein is 14 going to ask questions first. 15 SIR WYN WILLIAMS: Right. 16 Questioned by MR STEIN 17 MR STEIN: Mrs Chambers, I ask questions on behalf 18 of a large number of subpostmasters and 19 mistresses on behalf of a firm of solicitors 20 called Howe+Co. 21 You've given evidence and been asked a large 22 number of questions by Mr Beer, so I don't need 23 to cover those areas. 24 You said, as part of your evidence that on 25 some KELs you would see associated issues but 159 1 not all. You also said, as part of your 2 evidence today, that sometimes you would expect 3 to your colleagues and then you would discover 4 that they'd been dealing with another matter in 5 the last few weeks. Now, help us a little bit 6 with the question of duplicates. There was 7 a system that we know that the first tier of the 8 helpline were discouraged from sending through 9 duplicates to the third, fourth lines. Now, 10 would you see on your KELs that there were 11 duplicates? 12 A. You mean if the Helpdesk had used the same KEL 13 for various tickets that they had received? 14 Q. If they had recognised what they thought or 15 believed was a KEL that was the same as one that 16 was already being dealt with, would you see 17 that? 18 A. Not automatically, no. But, obviously, some 19 KELs would say, "Send the call to SSC". Even 20 though it's logged as a known problem, we would 21 still make it clear on the KEL that we still 22 needed to see the call, if that was thought to 23 be appropriate. 24 Q. So if you didn't see some of the issues that 25 were coming through, because they were thought 160 1 to be duplicates and therefore shouldn't go to 2 third and fourth line support, does that mean 3 that you were not always aware of how many calls 4 were being made through to the helpline on the 5 same issues? 6 A. That's true, if it was something that there was 7 a resolution or possibly a workaround, which 8 they could tell the postmaster about themselves. 9 Q. Who provided the training to the first line 10 support, the first line Helpdesk call answerers 11 so they could recognise that this was the same 12 or a different KEL; who provided that training? 13 A. I don't know who provided their training. 14 Q. Were you ever asked to provide such training? 15 A. No. 16 Q. During the evidence in this Inquiry, many of our 17 clients, who are ex-subpostmasters and 18 mistresses said that their accounts, branch 19 accounts, never seemed to balance. So that was 20 Janice Adams; Mujahid Faisal Aziz explained that 21 there were very any shortfalls that they had to 22 balance by paying in cash; Edward Brown said 23 similar matters occurred to him and that it 24 wasn't always a large shortfall but sometimes it 25 could be into the thousands; and Gary Brown 161 1 reported that the shortfalls happened so often 2 that it was hard to keep track. 3 Can you help us understand how it was that 4 the subpostmasters and mistresses experienced so 5 many shortfalls? 6 A. No, I can't, and it obviously concerns me if 7 this was happening and they weren't being given 8 assistance by anybody to get to the bottom of 9 these problems. 10 Q. You've mentioned at the closing part of your 11 statement, paragraph 212, I'll read it out, 12 that: 13 "A point of frustration with the system was 14 that the users, namely the subpostmasters, were 15 not our clients, and there was a practical limit 16 as to the extent to which we could work together 17 with them to investigate problems." 18 Was that true, this difficulty, having that 19 separation? 20 A. Yes. I mean, we had no ability to find out what 21 was actually taking place at the branch. I'm 22 not necessarily saying it was their -- they were 23 making errors that were causing these problems 24 but where we were checking, by one means or 25 another, that the system was correctly 162 1 calculating discrepancies based on the 2 transactions that had been recorded on the 3 system for the branch, then if that calculation 4 and those checks show no problems, then you've 5 got to try to find out what has been recorded in 6 the branch accounts or what is missing in the 7 branch transaction list which is causing the 8 discrepancy. 9 And unless you've got some way of going to 10 a branch and actually finding out what should 11 have gone onto the system, then you cannot 12 identify the cause of the discrepancy. 13 Q. When you say at paragraph 212 that the 14 subpostmasters were not your clients, your 15 client was the Post Office? 16 A. Yes. 17 MR STEIN: Excuse me one moment. 18 Thank you, Mrs Chambers. 19 Questioned by MR MOLONEY 20 MR MOLONEY: Mrs Chambers, one matter and it's just 21 one document, FUJ00138385. 22 A. I'm sorry but I can't hear you very clearly. 23 Q. I'll speak up and get closer to the microphone. 24 Sorry, Mrs Chambers. 25 A. Thank you. 163 1 Q. It's just one document, FUJ00138385, and it 2 should go onto the screen -- and it's on screen 3 now. Thank you. 4 The subject line, the title, is "Requesting 5 journal data from Audit", the author is you, 6 Mrs Chambers, and it's created on 25 August 7 2011. 8 A. Yes. 9 Q. So that's after the migration to HNG-X or 10 Horizon Online -- 11 A. Yes. 12 Q. -- however one terms it. We see from this 13 document that: 14 "All journal messages arriving at the data 15 centre are retained for audit. Occasionally SSC 16 may need to ask for data to be retrieved to 17 enable issues which happened more than six 18 months ago to be investigated." 19 Then in brackets: 20 "(Less than six months, there may still be 21 sufficient and more accessible information in 22 BRSS). 23 "When asking for transactions for a FAD ..." 24 That's a branch, isn't it? 25 A. That's a branch. 164 1 Q. "... and date range, ask for the 2 QUERY_AT\FINAL\Filteredhx.xml file which will 3 contain all transactions for a given date range 4 and FAD code in XML format. These will however 5 lack the JSN and ReqMessageID. 6 "The alternative is raw files containing 7 data for all 80 or so branches with the FAD 8 hash, which is far harder to read but does 9 contain JSN and ReqMessageID. 10 "Route the PEAK with the request to Security 11 Ops. 12 "If you think the call may be part of a Post 13 Office investigation into a branch that might 14 lead to litigation, then this should not be 15 handled by SSC unless already authorised by the 16 SSC manager." 17 So it follows from this document, as we are 18 already aware, that different types of audit 19 data provided different information to analysts? 20 A. Yes. 21 Q. So here, it refers to raw files of data or raw 22 data? 23 A. Err ... 24 Q. As one type of data available? 25 A. One type of data is the raw files, yes. 165 1 Q. And then there's the XML files, as well? 2 A. Yes. 3 Q. The raw files of data, the raw data, would 4 contain information not contained in the XML 5 file? 6 A. It -- yes, I cannot now remember precisely what 7 the details are but, obviously, there's two 8 fields there that might have been of use. 9 Q. Yes. How frequently would raw data be requested 10 in your experience? 11 A. I can't remember ever personally actually 12 requesting journal data from Audit. I almost 13 certainly did on occasion but I've got no memory 14 of doing it. It certainly wasn't a frequent 15 thing. 16 Q. Why would the PEAK and request have to go to 17 Security Ops, as is said in the third line from 18 the bottom? 19 A. Because they were the ones who could access this 20 data and they had to extract it from the audit 21 servers. Only they could do that. 22 Q. Right. How would an SSC analyst be aware that 23 a request might lead to litigation? 24 A. Not might lead to but might already be part of 25 a Post Office -- sorry, yes. Um, it would just 166 1 depend, you know, where the call had come from, 2 if anybody had mentioned this. I don't know. 3 I mean, how we would tell that now, I'm not 4 100 per cent sure. 5 Q. Yeah. Because this does say that: 6 "If you think the call may be part of a Post 7 Office investigation into a branch that might 8 lead to litigation, then this should not be 9 handled by SSC unless already authorised by the 10 SSC manager." 11 Why would the request not be handled by SSC 12 unless already authorised by the SSC manager? 13 A. I presume I was told that. I'm not sure. 14 I mean, this is an unusual situation if we are 15 asked to investigate something that happened 16 more than six months ago. Normally, we're 17 investigating -- would have been investigating 18 things that had happened recently. So I presume 19 that was what I was told, and that is why 20 I added that into the work instruction. But 21 I cannot remember of any conversation about 22 that. 23 Q. So this is a work instruction, is the term 24 you've just used; is that right? 25 A. Sorry yes, a "WI" is a work instruction. 167 1 Q. Work instruction. You would have been told that 2 if it was part of a Post Office investigation 3 into a branch that might lead to litigation, 4 then it shouldn't be handled by SSC, unless 5 already authorised by the SSC manager? 6 A. I appear to have included that in the work 7 instruction, so I assume that was what I was 8 told. 9 Q. So did you understand why it was that you were 10 told that this should be included in the work 11 instruction? 12 A. Because presumably, in that case, Post Office 13 would be putting in their own request for the 14 data. 15 Q. Why would that need to be authorised by the SSC 16 manager? 17 A. Just because I was told that. I'm sorry, I have 18 no real recollection of this. I don't recall 19 being told it. I don't actually recall writing 20 the work instruction but I obviously did. 21 Q. Just to float one possible reason, could it be 22 a reason of payment for this type of request? 23 A. That may have come into it because I think Post 24 Office were -- they were charged for data 25 retrievals but, yeah, I think this would be 168 1 better directed at my -- one of my managers, 2 probably. 3 Q. Of course. I only ask you because you're the 4 author of this document, Mrs Chambers. 5 A. Yes. I expect Steve or somebody said, "Oh, this 6 ought to be in a work instruction, can you 7 create one?" 8 Q. That was the final question I had for you, which 9 is: in what capacity were you providing 10 direction of this nature to your colleagues? 11 A. Obviously, I would have a good understanding of 12 the technical messages and the content of the 13 messages, and so on, or, you know, a reasonable 14 understanding of that. And, yeah, I'm -- 15 I wrote a work instruction. I'm sorry, I don't 16 really remember any more about it than that. 17 Q. Given that you didn't know really why that final 18 three lines, or rather final two lines, were 19 included within this work instruction or you 20 can't remember why, would this have been better 21 being a work instruction emanating from your 22 manager rather than from you? 23 A. Um, yes, it probably would have been, but if 24 a work instruction was felt to be needed for 25 something, then somebody might well be asked to 169 1 just "Oh, could you write that work 2 instruction?" But that didn't mean that all the 3 content necessarily came from me. We could all 4 write work instructions and it was the sort of 5 job that got shared out amongst us. 6 MR MOLONEY: Thank you very much, Mrs Chambers. 7 Questioned by MS PAGE 8 MS PAGE: I'm so sorry, can I just ask those in 9 front of me to just sit as they normally do, so 10 I can see the witness. Thank you very much. 11 A. I still can't hear you terribly clearly. 12 Q. Is that any better? 13 A. That's better, thank you. 14 Q. It's Flora Page, also acting for a number of 15 subpostmaster Core Participants. I'd like to 16 take you to two documents. They're both PEAKs 17 or possibly one might be a PinICL and they both 18 deal with the process that you went through in 19 order to insert transactions to ensure that the 20 accounts were balancing properly. 21 So the first one is FUJ00152239. What we 22 can see is that the summary at the top shows us 23 that the office can't balance as there are 24 "incorrect fees on POs", and I think that stands 25 for Postal Orders? 170 1 A. Yes, it does. 2 Q. We can see that this dates from July 2001. If 3 we then -- I don't think we need to see anything 4 in the first few dealings but if we go down to 5 page 4 and we can see that it starts off in your 6 department with Barbara Longley, who's the 7 administrator; that's right, isn't it? 8 A. Yes. 9 Q. Then we see that John Simpkins initially picks 10 this up, and we can see that he has -- or 11 somehow, about halfway down the page, he has 12 noticed -- he says: 13 "PRESCAN: Check date/time runs in message 14 store for time BU was swapped." 15 So we can see a base unit has been swapped 16 out; is that fair? 17 A. Yes, that would be fair. 18 Q. Then if we go a bit further down, John Simpkins 19 has assigned this to you, "Team Member: Anne 20 Chambers", or somehow it has been assigned to 21 you; do you see that? 22 A. Yes, I do. 23 Q. Then he says that it might be: 24 "... a problem due to the corrupt storage 25 unit, check the message store for any corrupt 171 1 entries then insert a REM OUT for PO Fees ..." 2 A. I see that. 3 Q. Yes, but if we then go over the page, you seem 4 to take a slightly different view at one point, 5 but we'll go through it logically. At 15.41, 6 you say: 7 "It looks as if Adjusts Stock on 4th Jul was 8 showing incorrect figures ..." 9 Then you've referenced a KEL. 10 "As a result, the PM did couple of sets of 11 unnecessary SAPs ..." 12 After base units swapped, it seems. 13 Then if we go down almost to the bottom of 14 the -- I'm so sorry, to about four lines below 15 that, you say: 16 "I've raised OCR AChambers ... to allow us 17 to correct the messagestore." 18 So that's an instance, is it, of you saying 19 that you need to go through the change control 20 process to insert transactions; is that right? 21 A. That appears to be what I did. I have to say 22 I have no recollection of this at all. 23 Q. I wouldn't be surprised, it's obviously going 24 back a very long way, isn't it? 25 A. Yeah. 172 1 Q. Then it says, below that: 2 "Incident Under Investigation." 3 Then, if we go further down, we can see "New 4 evidence added", and I'd just like to try to 5 understand what these evidence types are. 6 We've got: 7 "New evidence added -- Full message store." 8 Then we've got: 9 "New evidence added -- audit logs." 10 Then we've got: 11 "New evidence added -- PSStandard logs." 12 Is "full message store" the equivalent of 13 what became the ARQ data? 14 A. No, it's not. It would include all the ARQ data 15 but it's the -- all the messages for the branch 16 that were in existence on the day and time that 17 I did the retrieval. I'd have retrieved it from 18 the copy of the message store that was held on 19 the correspondence server centrally and so it's 20 all the transaction messages and a lot of other 21 messages that have been written in the last 42 22 days, it would have been at this point, plus all 23 the reference data relating to the branch. But 24 the ARQ data is the same data but it's captured 25 in a different way. 173 1 Q. That last bit you said is the bit I wanted to 2 get at. It is the same data, is it? 3 A. It's the same data which I'd retrieved from the 4 correspondence server but the messages, as they 5 came in from -- I think it was happening all the 6 time -- as they came in from the branch to be 7 fed into the correspondence server message 8 store, they were also -- there was a stream of 9 them going out all the time into the audit 10 files. So the ordering, in particular reference 11 data, and so on, would be rather different but, 12 overall, it's the same data. 13 Q. So that's the evidence that could be captured 14 for a significant period of time afterwards and 15 that was stored -- 16 A. The ARQ data files were kept for a significant 17 amount of time. The message store, some of the 18 messages persisted, but others would be archived 19 or deleted after 42 days. 20 Q. All right, so it's not identical. All right. 21 Well, then, audit logs. What is that? Is that 22 identical with ARQ or not? 23 A. No, this is the files that I had totally 24 forgotten about until Mr Beer reminded me of 25 their existence yesterday, I think it was, which 174 1 were kept on the counters, were written on the 2 counters, and contained a certain amount of 3 diagnostic information written to the file by 4 the counter application as things were done. 5 I can't -- I've got very little recollection of 6 precisely what that looked like. 7 Q. Then PS standard logs, what are they? 8 A. That's another counter log file in which you 9 could see messages to and from the counter 10 peripherals. Things like the printer and the 11 barcode reader, and so on. 12 Q. None of that was kept for any significant period 13 of time; is that fair? 14 A. No, that wasn't kept and it normally wasn't 15 retrieved from the counter. It was only if we 16 were investigating something we would get the 17 file from the counter. 18 Q. Yes. All right. Then further down, we can see 19 that you're asking for Development to look at 20 this and then, if we go over the page, you say 21 you haven't been able to reproduce it: 22 "This counter had a box swap an hour before 23 the problem occurred but I don't think that is 24 relevant." 25 You've then said that you've got another 175 1 report of the same problem elsewhere and you 2 give the forward number for the branch: 3 "... so please can this be looked at quickly 4 (especially as if it is not reported before 5 rollover, we have to get POCL authorisation for 6 the fix and so it is very visible)." 7 That means, does it, that if you had to fix 8 it after rollover by inserting transactions, 9 then it would be visible in some way? 10 A. I'm struggling to remember any details of this. 11 I think if it was reported, but not until after 12 the rollover, we probably possibly couldn't have 13 fixed it at all. There would have been 14 a receipts and payments mismatch, and then we 15 would just have reported it to Post Office 16 through the MSU BIMS route. But I can't be 17 certain about this, having no memory of it, and 18 I don't think I've seen this document until this 19 moment. 20 Q. The word there, "visible", visible to whom, do 21 you think? 22 A. I don't know why I used that word. Yeah, 23 visible to Post Office, I suppose. But, as 24 I say, I have no memory of this. 25 Q. Or visible to the subpostmaster? 176 1 A. Um, I think the subpostmaster knew about the 2 problem already because they had reported the 3 problem. 4 Q. True, but you've identified that there is 5 another problem elsewhere. 6 A. Yes. That had been reported to us as well, so 7 that was another postmaster who had noticed that 8 it had happened. I mean, the summary is that 9 they can't balance. I can't remember if it 10 actually totally stopped the process because of 11 this inconsistency, or if they could push on. 12 I mean, obviously Postal Orders and the fees 13 associated with them should always be in step 14 with each other. You shouldn't be able to have 15 one without the other, and something had gone 16 wrong here, and they were out of step. 17 Q. A little further down, it says that there may 18 be -- the counter is M1 and M1R. You may not be 19 able to recall what that means. But you do 20 say -- sorry, Les Ong says: 21 "There are two fixes that I know of ... 22 relating to Postal Orders that could have 23 a bearing on this ..." 24 Then, if we go over the page, what we see 25 when it comes back to you -- and this is on 177 1 10 July, so it's a subsequent day -- at 15.31: 2 "Authorisation for messagestore amendment 3 now received from ..." 4 Then we seem to get an email address, 5 "mick.theobald", and it has been edited out so 6 we don't have the full email address. Is that 7 a name that rings a bell? 8 A. I think he was a Post Office person but I can't 9 be a 100 per cent sure. 10 Q. Following that and a little further down, we see 11 that: 12 "Applied fix to message store ..." 13 Then there's the reference to the OCR again, 14 and: 15 "Balance snapshot now shows 19 POs and fees. 16 Leaving call open until balancing/cash account 17 done." 18 So it looks as if it has been possible to 19 apply this before rollover. 20 I suppose the question that I'd like to ask 21 is how would it be possible to see, on the ARQ 22 data, what had happened here? 23 A. Right. In the ARQ data, you would see, when the 24 initial problems were happening, the SAP -- 25 stock adjust positive and stock adjust 178 1 negative -- lines that related to where they 2 were trying to adjust their stock of Postal 3 Orders which they'd documented, and it was in 4 there that the amounts got out of step in that 5 one, for the Postal Order itself it was for 6 a certain quantity, and for the fee it was for 7 a different quantity. So that's where the 8 problem arose. 9 And then, in the ARQ data for 11 July, you 10 would see the transaction also affecting the 11 postal -- presumably affecting the Postal Order 12 fees product, which I'm guessing now, but 13 I imagine was another stock amount -- stock 14 adjust transaction for the difference that was 15 wrong. Whether there is anything on that 16 individual message which, in the ARQ data, 17 enables you to know that it was me who did it 18 and not somebody at the branch, I do not now 19 know, because I have got no record in this PEAK 20 here of exactly what it was that I inserted. 21 It's possible, as I said before, sometimes 22 we used a dummy counter number. Sometimes we 23 inserted a comment, but that is not necessarily 24 going to be visible in the ARQ data as 25 retrieved. Sometimes we used a username to try 179 1 and make it obvious that it was SSC who had made 2 the change but it's not recorded on the PEAK 3 here precisely what was done. Those messages 4 would have been captured somewhere and recorded 5 for posterity but I don't know where. 6 Q. The transactions would be asynchronous, would 7 they, in the sense that the balancing 8 transactions that you've inserted would show the 9 date that you inserted them, not backdated? 10 A. Yes. 11 Q. All right. Thank you. If I may just then 12 briefly -- the next one is a bit quicker. If we 13 look at FUJ00152240. We can see this is 14 summarised as "Cannot put transfer through", and 15 this dates from 2004. The last entry that we 16 can currently see says: 17 "PM reports that he cannot put a transaction 18 through it keeps coming up with an error 19 message." 20 If we then pick this up on page 2 at 12.30, 21 if we scroll down a little bit, this is where 22 Barbara Longley has assigned it to you and 23 you've stated that: 24 "The transfer causing the problem was 25 started while the user was attached to the 180 1 SU BDC." 2 Can you just remind us: SU BDC? 3 A. Stock unit called BDC, which was most likely 4 Bureau de Change. 5 Q. That certainly seems to be an issue with this 6 one. There does seem to be problems with it 7 being foreign transfer. Then you've said, in 8 the last paragraph of this entry: 9 "I've spoken to the PM and asked him not to 10 balance stock units BDC, MM or MC until we have 11 sorted out the problem. I'm loading up the 12 messagestore on a test counter and hope that by 13 amending the EPOSSTransfers object it will then 14 be possible to reverse the transfer." 15 If we scroll down, just going over the page 16 line: 17 "I've made a messagestore correction ..." 18 So you've then given an OCR reference so it 19 looks as if, again, we've got this process where 20 you're formally seeking the approval to insert 21 a transaction; is that right? 22 A. That's what it looks like. Again, I've not seen 23 this and I've got no recollection of it, but it 24 looks like it. 25 Q. It says: 181 1 "Before and after messages attached." 2 Is that a practice that rings a bell? 3 A. Well, we always -- yes, we'd always make 4 a record of what we were changing or adding in. 5 I don't know precisely now what they look like. 6 Q. When you say "attached", is that attached to the 7 PEAK? 8 A. Yeah. 9 Q. So the PEAK would have had the messages before 10 and after the message you inserted attached to 11 it; is that right? 12 A. That's what it sounds like, yes. 13 Q. You've then recorded: 14 "Have spoken to PM and informed him he 15 should be able to continue with the balance 16 now." 17 A. Yes. 18 Q. What you don't say is "I've informed him I've 19 inserted a transaction into your account", do 20 you? 21 A. I don't explicitly say that, but I imagine 22 I would have explained to him what I had done, 23 and that I had removed the transfer, which I -- 24 I mean there's -- it's -- I'm not clear from 25 this whether I wrote a pair of corrective 182 1 messages that actually removed -- how would it 2 have done it? Um, I'm not clear. Further up in 3 the call there was mention of an EPOSS transfers 4 object and suggesting that that needed to be 5 rewritten in order to let this progress. So 6 I don't know if that's what I actually did, or 7 whether I did insert a pair of opposite messages 8 to undo the transfer that was outstanding. 9 I certainly wouldn't have hidden from him 10 the fact that I was changing something on his 11 system which would remove this transfer that was 12 preventing him from balancing his office and 13 continuing to trade, to do his normal business. 14 Q. Not hidden, but maybe not mentioned in the sense 15 that it's not recorded? 16 A. It's not written down but that doesn't mean 17 I would have said it because I usually did. 18 I didn't make a secret of the fact that system 19 problems happened, and I think it was 20 perfectly -- I'm sure it was perfectly clear to 21 him that there had been a system problem to do 22 with a transfer which I then -- and then, once 23 I had done whatever it was I did, the transfer 24 that he didn't -- that was stopping him had been 25 removed in some way in order for him to 183 1 progress. 2 The fact I didn't write it down does not 3 mean I said absolutely nothing to him. 4 I certainly wouldn't have phoned him back and 5 said, "Oh look, it's miraculously all okay now, 6 you don't need to bother any more". I wouldn't 7 have approached it in those terms, but I do not 8 know precisely what I said to him. 9 Q. Well, a system problem is one thing, but 10 actually inserting transactions into the data 11 that's stored on his counter is a different 12 thing, isn't it? If it's not written and 13 recorded here, how would he -- how would 14 posterity ever know that you'd ever said that to 15 him and told him that's what you were doing? 16 A. If I'd known posterity was going to be asking 17 I would have written it down. But I don't know 18 if there's any more information in the OCR. 19 That certainly would make it clear exactly what 20 it was that I did. 21 Whether I explicitly said, "I have removed 22 the transfer, I have accessed your counter 23 transactions and removed the one that was 24 causing the problem", which is effectively what 25 I'd done, I'm not sure it even needed saying. 184 1 I would have thought he would have realised that 2 that was what I had done. But I cannot -- I do 3 not know what I said. 4 MS PAGE: Thank you. Those are my questions. 5 SIR WYN WILLIAMS: Is that it? Anyone else have any 6 questions? 7 MR BEER: No, they don't, sir. Thank you very much. 8 SIR WYN WILLIAMS: Right. Well, thank you, 9 Mrs Chambers, for giving detailed answers to 10 detailed questions over two days. As you know, 11 you will be asked to return at some future date. 12 I don't think we can yet tell you what that date 13 is. If you haven't already received it, the 14 probability is that you will get another Rule 9 15 Request so that the general questions will be 16 provided to you in advance and although, in 17 a sense, you're in the middle of giving your 18 evidence, it's unreasonable for me to expect 19 that you don't have access to your lawyers if 20 you want to have access to your lawyers. 21 So unless anybody immediately shouts out and 22 says to me "You can't do that", I'm now going to 23 tell you that if you want to speak to your 24 lawyers, you can. All right? 25 A. Thank you. 185 1 MR BEER: Thank you very much, sir. 2 For reasons that you know, we return on 3 Tuesday next week, 9 May, to hear evidence from 4 Barbara Longley at 10.00 am. 5 SIR WYN WILLIAMS: All right. Then the Inquiry is 6 adjourned until then. Thank you all very much. 7 MR BEER: Thank you, sir. 8 (4.12 pm) 9 (The hearing adjourned until 10 Tuesday, 9 May 2023 at 10.00 am) 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 186 1 I N D E X 2 ANNE OLIVIA CHAMBERS (continued) .....................1 3 4 Questioned by MR BEER (continued) .............1 5 6 Questioned by MR STEIN ......................159 7 8 Questioned by MR MOLONEY ....................163 9 10 Questioned by MS PAGE .......................170 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 187