1 Thursday, 27th June 2024. 2 (9.45 am) 3 GARETH IDRIS JENKINS (continued) 4 Questioned by MR BEER (continued) 5 SIR WYN WILLIAMS: Morning, my colleague, 6 Ms Eliasson-Norris, can't be here in person today. 7 MR BEER: Thank you, sir. 8 Good morning Mr Jenkins. 9 A. Good morning. 10 Q. Can we start with the Seema Misra case study, please. 11 Just for reference, no need to display it, you set out 12 in your third witness statement, paragraphs 326 to 521, 13 your involvement in Seema Misra's case. Just by way of 14 reference, before you ever became involved in it, back 15 on 1 April 2008, Mr Singh, Jarnail Singh, had advised 16 there was sufficient evidence for a prosecution of 17 Mrs Misra. The reference for that, again, no need to be 18 displayed, is POL00049658. 19 Can I turn, please, to what might be your first 20 involvement or the genesis of your first involvement in 21 the case. 22 A. Okay. 23 Q. FUJ00152866, please. If we can turn to page 6, please. 24 An email -- if we just scroll down, and a bit more. We 25 can't in fact see it. We'll have to scroll up to see it 1 1 was sent from Mr Dunks. There. Can you see an email of 2 26 November 2009 from Mr Dunks to Jane Owen? 3 A. I think it's copied to Mr Dunks, isn't it from Jon 4 Longman? 5 Q. You're quite right. Mr Longman to Ms Owen, copied to 6 Mr Dunks. You're quite right. If we scroll down it 7 says: 8 "I attach a report from the defence expert where he 9 has highlighted a number of problems with the Horizon 10 system. Our barrister, Warwick Tatford, has asked that 11 the problems with Horizon that he has raised in his 12 report are replied to in a witness statement form. 13 I presume that an employee of Fujitsu would have to 14 produce the witness statement ... 15 "In addition to this the defence have also requested 16 some information", which I'm not going to read. 17 Okay? 18 A. Okay. 19 Q. If we go to page 5, please, at the top of the page. We 20 can see that Jane Owen sends that to Penny Thomas in 21 Fujitsu on 1 December: 22 "This is the email and attachments that we chatted 23 about. Please let me know if there is anything else you 24 need from me and if this kind of request needs to be 25 raised in a more official way." 2 1 Okay? 2 A. Yes. 3 Q. You're not included at this stage? 4 A. Correct. 5 Q. If we go to page 3, please. If we just scroll up a bit 6 more -- thank you -- you'll see an email from Penny 7 Thomas to David Hinde. Do you remember David Hinde? 8 A. Oh, yes, he was one of the Programme Managers within the 9 Post Office Account and I worked quite closely with him 10 on a number of occasions. 11 Q. So he was a Post Office -- 12 A. No, he was a Fujitsu person. 13 Q. Working on the POA? 14 A. Yes. 15 Q. Thank you: 16 "David 17 "We had a conversation last week and you advised me 18 I should no longer ask Gareth to support our prosecution 19 ... activity. I now have a request from [the Post 20 Office], which they would like addressed prior to 21 Christmas, and I need an expert to respond. Could you 22 help me to identify a suitable candidate?" 23 Then she attaches the expert's report. 24 Do you know anything about why you were not to be 25 approached, "We should no longer ask Gareth to support 3 1 our prosecution" -- 2 A. Possibly because I was quite business working on Horizon 3 Online at the time. 4 Q. All right so it was about capacity, was it, as far as 5 you knew? 6 A. As far as I know, yes. 7 Q. Then if we scroll up, please, at page 2. Mr Hinde 8 replies: 9 "Penny 10 "We have not identified anyone else who can take 11 this on at present so this will need to be handled by 12 Gareth." 13 Then scroll up. An email that we've looked at 14 before, I think we looked at it yesterday or the day 15 before, which is really about generally what you do to 16 support prosecution activity. 17 A. Yes. 18 Q. So this chain doesn't actually -- even though there's 19 a request for the provision of an expert report in the 20 form of a witness statement, or an expert's reply to the 21 expert's report in the form of a witness statement -- 22 ask you directly to perform that function, does it, when 23 it's sent on to you? 24 A. No not at this stage, no. 25 Q. Can you help us, do you know why you're replying to this 4 1 chain with this series of questions and answers? 2 A. The questions are from this Dave Jennings and the 3 answers are mine. So, again, I'm interspersing things 4 so I think the darker font is his questions and the 5 lighter font, more to the left, are my responses, is how 6 I'm interpreting that. 7 Q. My question was more: do you know why you're saying all 8 this in answer to the series of emails below, which is 9 about, "Can we have a reply from somebody within Fujitsu 10 in the form of a witness statement", to the defence 11 expert report in the Seema Misra case? 12 A. I don't have a clear answer. I'm assuming, from what 13 this is saying, that Dave Jennings wanted to have some 14 sort of background as to the sort of things I'd done in 15 the past but I'm -- I'm guessing from the sort of 16 questions he's asking. 17 Q. Can you recall when you were first asked to provide 18 assistance to the Post Office in relation to Mrs Misra's 19 case? 20 A. I think it was from -- as a consequence of this email. 21 There may be -- I think there may be a later one, either 22 on this chain or a separate chain, that actually passes 23 me the report to have a look at but it was certainly 24 some time in December 2009. 25 Q. You tell us in your witness statement that you do not 5 1 think you received any kind of briefing or explanation 2 about the background to Mrs Misra's case; is that right? 3 A. At that stage, I don't believe I did. 4 Q. And that you weren't in fact provided with Professor 5 McLachlan's first interim technical report? 6 A. I don't believe I've ever seen that. 7 Q. Without a briefing or explanation about the background 8 to Mrs Misra's case or the prosecution case against her 9 and without a copy of Professor McLachlan's first 10 interim report, how did that impact on your ability to 11 respond to the Post Office's requests? 12 A. I saw the Post Office's requests as being purely to 13 comment on the second report that I was provided with, 14 either at this time or shortly afterwards. 15 Q. So, again, performing a narrow function? 16 A. Yes, because that's what I thought I was being asked to 17 do. 18 Q. Did you ever think, "Hold on, I don't know anything 19 about this case"? 20 A. No, I was just being asked "Please can you comment on 21 this report", so I commented on the report in the way 22 I normally did by interspersing my comments within the 23 electronic version of the report. 24 Q. You also, in your witness statement, describe the Post 25 Office's management of Seema Misra's case -- this is 6 1 something you mentioned previously in your oral 2 evidence -- as "fairly chaotic". 3 A. Yes, I would agree with that. 4 Q. What did you mean by your reference to the management by 5 the Post Office of Mrs Misra's case as being "fairly 6 chaotic"? 7 A. I think this came a bit more later. There were a number 8 of cases where I would respond to an email and then get 9 asked exactly the same question again by the same person 10 and, therefore, ended up sort of saying, "Well, I told 11 you this yesterday, here's the answer I gave you when 12 you asked me the question before". And there are number 13 of examples of that in the exchanges I had with Jarnail 14 Singh. 15 Q. Other than receiving repeated requests for the same 16 thing, was there anything else that made their 17 management of the case chaotic or fairly chaotic? 18 A. I wasn't being asked very clear questions. I was having 19 to make all sorts of assumptions of what they wanted me 20 to do, I did try and set out what I thought I was being 21 asked to do, but it was very much a case of I definitely 22 didn't have clear instructions as to "Please do exactly 23 this", let alone formal instructions, as I now realise 24 I should have had. 25 Q. Did you realise at the time that you weren't receiving 7 1 clear instructions? 2 A. I just thought that was the way things worked. So -- 3 Q. That's a slightly different point. Did you realise at 4 the time that you were not receiving clear instructions? 5 A. No, because I didn't know that I should have been. 6 Q. Irrespective of the formality of a letter of 7 instruction, which you say you now realise that you 8 ought to have received, did you at the time realise that 9 the requests being made of you were unclear? 10 A. They were certainly unclear and I did the best I could 11 to interpret what I thought I was being asked. And 12 sometimes I did actually seek clarification to say, 13 "I think what you mean is this, isn't it?" 14 Q. You also describe in your witness statement that: 15 "At times I struggled to deal with the Post Office's 16 demands." 17 Is that because you were heavily involved in the 18 rollout of Horizon Online? 19 A. Yes. I mean, there was clearly a scheduling problem 20 I had, which I think I managed to fulfil, but, yes. 21 Q. Did you feel that there was pressure applied to you by 22 the Post Office to provide evidence in Mrs Misra's case 23 that conformed to the Post Office's expectations? 24 A. Yes, they clearly wanted me to say that everything was 25 perfect and I don't think I actually said that. 8 1 Q. In what way did that pressure manifest itself? 2 A. I kept on getting emails -- it was in spurts, it wasn't 3 solid throughout the time, but I kept on being asked 4 please can I do this, please can I do that, and I kept 5 getting occasional phone calls as well. 6 Q. Did you feel uncomfortable about the pressure you were 7 being placed -- that was being placed upon you? 8 A. Uncomfortable -- I just got on with things, which is 9 what -- my normal reaction to try and handle things, 10 rather than worry about things. 11 Q. So you felt uncomfortable but got on with it? 12 A. Yes. 13 Q. Did you raise the fact that you felt uncomfortable with 14 any of your managers? 15 A. I -- yes, for example, there was the discussion with 16 David Jones at the beginning of February that we had and 17 that was part of my uncomfortableness, if you like, and 18 there were occasions when I'd raised with my line 19 manager, "Look, I'm being asked to do all this but you 20 want me to do all my day job stuff, how does this all 21 tie together?" 22 Q. I was not thinking so much about the capacity issue but 23 rather the "we want evidence", the substance of the 24 evidence. 25 In your words, I think you said to say that the 9 1 system was all okay or all alright, and you felt 2 uncomfortable about doing that. 3 A. I think that was the reason why we got in touch with 4 David Jones, for example, at the beginning of February. 5 Q. Other than that, did you escalate it in any way within 6 your organisation? 7 A. I can't remember properly. I think there may be some 8 examples within the emails but I can't remember off the 9 top of my head exactly which ones they were. 10 Q. Okay. Can we look, please, at FUJ00122794. 11 By this time, this is February 2010, for context, 12 you had received, by this time, a request to produce 13 a witness statement and, if we go down, please, to 14 page 2 -- thank you -- we'll see an email from Jarnail 15 Singh, in fact it's from his secretary or PA on his 16 behalf, of 5 February at 3.50 to Mr Jones, and copied to 17 you and Penny Thomas. 18 A. Yes. 19 Q. The email body is addressed to you: 20 "Dear David and Penny ..." 21 He says: 22 "On first glance points 2-4 have not been answered 23 and I reproduce below." 24 Item 3: 25 "When Gareth completes his statement ..." 10 1 As I said already, by this time, you'd been asked to 2 provide a witness statement? 3 A. Yes. 4 Q. "When Gareth completes his statement could he also 5 mention whether there are any known problems with the 6 Horizon system that Fujitsu are aware of. If none could 7 this be clarified in the statement." 8 This is an email from the Post Office to three of 9 you in Fujitsu, correct? 10 A. Yes. 11 Q. You're getting the request directly here from the 12 lawyer, the prosecution lawyer, Jarnail Singh? 13 A. Yes. 14 Q. Just looking at of the request in number 3, would you 15 agree that that's a broad request? 16 A. Yes, it is. 17 Q. It does not ask whether there were known problems which 18 affected Seema Misra's branch at West Byfleet, does it? 19 A. No, not at that stage. 20 Q. It doesn't ask you whether you personally know of 21 problems within Horizon; it asks whether you can say 22 whether Fujitsu knows of problems with Horizon, doesn't 23 it? 24 A. Yes, and I responded to that saying that -- identifying 25 a problem that I was aware of that would need to be 11 1 checked out. 2 Q. I'm going to come to your answers in a moment. I'm just 3 looking at what the request means at the moment. 4 I think you've agreed that it was broad. It's not 5 limited to Seema Misra's branch and it's not limited to 6 your own knowledge; it's Fujitsu's knowledge? 7 A. I see now that is what it said. I'm not sure that 8 I would have analysed it in that way at the time. 9 Q. Why might you not have analysed it or why -- put aside 10 analysis of it: why might you not have read it that way 11 at the time? 12 A. I thought that all I could be really commenting on was 13 what was happening in the particular branch when we 14 actually had some data to look at for the specific 15 branch. 16 Q. That plainly doesn't say that though, does it? It 17 doesn't ask that? 18 A. I realise that now. 19 Q. We now know that request originated from the prosecution 20 barrister, Warwick Tatford. If we go up, please, to the 21 bottom of page 1, if we just look at your reply, 16.06. 22 Can you see that you replied to David Jones and Penny 23 Thomas? 24 A. Yes. 25 Q. So you have cut the Post Office person, Jarnail Singh, 12 1 out of the chain? 2 A. Yes. 3 Q. Why did you cut Jarnail Singh out of the chain? 4 A. I wanted advice from David Jones in particular as to 5 exactly what I should be saying, particularly since 6 I was reluctant to make a clear statement and I was 7 aware of this problem so I wanted guidance as to what 8 I should be saying about it. 9 Q. So you were really looking for a steer from Fujitsu 10 Legal on how to deal with this request? 11 A. Yes. 12 Q. You say: 13 "Brief responses as [below], but not sure that 14 I should put them in a witness statement ..." 15 A. Again, I was seeking guidance. 16 Q. Just help us with why you weren't sure you should put 17 them in a witness statement? 18 A. I was just trying to get some clarity as to whether I -- 19 exactly what I should be saying. I was looking for 20 guidance particularly from David Jones as to what 21 I should be saying. 22 Q. But why were you not sure you should put them in 23 a witness statement? 24 A. I just didn't know. I'd not really had such much to do 25 with this sort of thing, so I was just seeking some sort 13 1 of guidance. 2 Q. But what were the reasons that would prevent you from 3 putting them in a witness statement? 4 A. I didn't know. I was just ignorant. 5 Q. What, that, "Is this something that we do, we provide 6 witness statements?" I mean, you were providing one 7 already? 8 A. Yeah, yeah. I was just -- I was just asking the 9 question. 10 Q. Wasn't it more about the content; you were worried about 11 putting the content in a witness statement? 12 A. I was just looking for guidance as to what I should be 13 saying. 14 Q. Weren't you really not focusing on "Should I provide 15 a witness statement or not", you were saying by that, 16 "I'm not really sure that I should put this content in 17 a witness statement"? 18 A. I don't know. 19 Q. Look at number 3. You'll remember this is the answer to 20 the question: 21 "Could he also mention whether there are any known 22 problems with the Horizon system that Fujitsu are aware 23 of? If none, could this be clarified in the statement?" 24 You say: 25 "This is where I'm reluctant to make a clear 14 1 statement. I am aware of one problem where transactions 2 have been lost in particular circumstances due to 3 locking issues. When this happens we have events in the 4 eventing logs to indicate that there was an issue and 5 whenever we provide transaction logs to [the Post 6 Office] we check for any such events. In the case of 7 West Byfleet we have not provided any transaction logs 8 and so have not made these checks." 9 Would you agree that the implication of that was 10 that you may be able to respond to question 3 when you 11 had the ARQ data? 12 A. Yes. I think that's basically what I was trying to say 13 there. 14 Q. Now, you refer to one problem in that answer: you were 15 referring to the Callendar Square bug, right? 16 A. No, I was referring to the Craigpark issue. I wasn't 17 aware of the Callendar Square issue at that time. 18 Q. Could be the Craigpark be checked for by reference to 19 ARQ data? 20 A. It could be -- no, but it could be checked in reference 21 to the NT event logs. 22 Q. Is that what you're referring to where you say 23 "transaction logs"? 24 A. Well, the transaction logs I'm referring is to ARQ data, 25 but I was saying that, when we provide transaction logs, 15 1 we also check against the corresponding NT events as 2 part of that process. 3 Q. That's what you refer to when you say "eventing"? 4 A. Correct. 5 Q. So that means NT event logs? So the third line -- 6 A. Yes, yes. 7 Q. At this time, did you know of other problems which would 8 require more than the ARQ data to investigate whether 9 they were in play at this branch? 10 A. I wasn't aware of any other such problems. 11 Q. What about accounting discrepancies due to hardware 12 failures? 13 A. I thought that those would have been picked up as -- 14 would have been picked up at the time because they would 15 have been fairly obvious at the time when there were 16 such hardware failures. 17 Q. You'd need to look, wouldn't you? Whether they'd been 18 picked up at the time or not was the very issue that 19 you'd be looking for, wouldn't you? 20 A. I'm not sure whether I was thinking of that at the time. 21 This was -- 22 Q. You agree now? 23 A. This was this issue that I thought about, so therefore 24 I raised this and asked for guidance as to what should 25 be done about it. 16 1 Q. What about bad blocks; do you remember what they were? 2 A. I don't think bad blocks actually caused problems. What 3 that effectively meant was that the system came to 4 a halt and then needed to be restarted to sort things 5 out. 6 Q. Would that be shown on an NT event log? 7 A. Yes. 8 Q. You'd need the Horizon Helpdesk records to look for 9 calls suggesting that hardware had been at fault, 10 wouldn't you? 11 A. I was aware that someone else had done an analysis of 12 Helpdesk calls or -- 13 Q. Were you aware -- 14 A. -- maybe not at this point but I became aware of that, 15 as time went on. 16 Q. At this point, in order to answer that question, even if 17 it was restricted in your mind to West Byfleet, you 18 would need to look at the Horizon Helpdesk records to 19 look for calls made by Mrs Misra or other people at 20 branch level, for example suggesting hardware issues, 21 wouldn't you? 22 A. I wasn't aware that I would have needed to have done 23 that because, as I say, I thought that was something 24 that someone else would be covering. 25 Q. Why did you think that it would be something that 17 1 somebody else would be covering? 2 A. Because my expertise was in the software rather than the 3 hardware. 4 Q. Wouldn't you need the full message store to look for 5 unusual restarts? 6 A. You could see those, you should be able to see those 7 from the ARQ data, I think. 8 Q. You think or you know? 9 A. I'm trying to remember now. It's a long time ago since 10 I looked thorough the detailed ARQ data. 11 Q. If you were going to give a complete answer to the 12 question whether there were any known problems with the 13 Horizon system that Fujitsu are aware of, you would need 14 to investigate PinICLs and PEAKs, wouldn't you? 15 A. Yes, but, as I say, at this stage, I was just saying 16 I couldn't make a clear statement and I wasn't thinking 17 about what else I would need to do. I was just saying, 18 "This is my immediate response to the question I've been 19 asked". 20 Q. But do you agree that, if you were going to give 21 a complete answer to the question of whether there were 22 any known problems with the Horizon system that Fujitsu 23 are aware of, you would need to look at PinICLs and 24 PEAKs? 25 A. There would have been some need to do that and I believe 18 1 I did so at some point. 2 Q. You would need to look at KELs to see which problems 3 were in play for at least the period when Mrs Misra was 4 alleged to have stolen money from the West Byfleet 5 branch, wouldn't you? 6 A. I'm not sure I would have thought of looking at KELs 7 because my normal approaches was to look at the PEAKs 8 and PinICLs which are sat behind them, rather than the 9 actual KELs. 10 Q. But you agree that, if you were to answer the question 11 completely of whether there were known problems with 12 Horizon that Fujitsu are aware of, you would at least 13 have to look at any PinICLs or PEAKs to see whether any 14 known problems were in play for at least the period when 15 Mrs Misra was alleged to have stolen money from West 16 Byfleet? 17 A. I wasn't actually answering the question at this point; 18 I was just saying that I didn't think I could answer the 19 question -- 20 Q. I'm asking you: if you were to answer it accurately and 21 completely, that's what you would need to do? 22 A. Possibly. As I say, I don't think I ever did actually 23 answer that question, as such. 24 Q. We know that you didn't. I'm looking at what you would 25 need to do in order to have done it properly, 19 1 accurately, completely. The implication of the answer 2 that you gave, namely "I could answer the question if 3 I was given event logs and ARQ data", that wasn't right, 4 was it? You needed much more information to answer it 5 accurately and completely? 6 A. I had not thought of that at the time. 7 Q. Do you agree now? 8 A. Possibly. I don't know. 9 Q. Why don't you know? 10 A. Yeah, I probably would have had to do some further 11 research. 12 Q. Thank you. 13 Now, I think you and your colleagues within Fujitsu 14 made a number of suggestions to the Post Office in the 15 following days to obtain relevant transaction data in 16 relation to Mrs Misra's tenure at the West Byfleet 17 branch, didn't you? 18 A. Yes. 19 Q. Can we look at those repeated requests to the Post 20 Office. FUJ00122713, and page 1. Thank you. 21 5 February, 2.46; Mr Jones, the lawyer, to Mr Singh, the 22 lawyer; you and Penny Thomas copied in. It's just in 23 relation to the part -- if we scroll down, please, the 24 penultimate paragraph, Mr Jones says to the Post Office: 25 "One concern is that [the Post Office] have not 20 1 apparently requested transaction data for West Byfleet 2 for the period and transactions in question. This would 3 normally be provided in previous cases and would include 4 Fujitsu extracting log files from the system to enable 5 us to provide details of transactions. Surprisingly 6 this has not been requested in this case. Perhaps you 7 would consider the need for this." 8 That's a request or sentiment I think you would 9 agree with? 10 A. Oh, definitely. 11 Q. Thank you. 12 If we look, please, at FUJ00152930. Same day at 13 5.10; same distribution, Jones to Singh; you and Penny 14 Thomas copied in. If we scroll down, please -- thank 15 you -- he effectively passes your reply, that we've just 16 looked at, into numbered paragraph 3 -- 17 A. Yes. 18 Q. -- and, in the last part of the paragraph, he says to 19 Mr Singh: 20 "In the case of West Byfleet we have not been asked 21 to provide any transaction logs and so have not made 22 these checks." 23 A. Yes. 24 Q. Correct? So, essentially, making a similar demand or 25 request or point to the Post Office? 21 1 A. Yes. 2 Q. Agreed? 3 A. Agreed. 4 Q. So twice in one day, essentially? 5 A. Yes. 6 Q. Then can we look at POL00093961, page 19, please. This 7 is a draft witness statement you provided on 8 February 8 2010 in Seema Misra's case. If we just go forwards for 9 present purposes to page 21. We can see in the top 10 paragraph, I think you're performing the familiar task 11 of cutting into a document of yours what somebody else 12 says, here it's Professor McLachlan; is that right? 13 A. Yes, I think this is his third report, I believe? 14 Q. Yes. 15 A. And so the italics is my response to the normal font of 16 his actual statements in the report -- 17 Q. So in his report he'd said: 18 "The implication is that EPOSS transactions can be 19 lost due to equipment failures. Without access to the 20 sub post office data records and the intermediate data 21 records in the end-to-end process it will not be 22 possible to identify the extent to which this may 23 explain the accounting discrepancies." 24 You say in your draft witness statement: 25 "No request has been made to Fujitsu for any data 22 1 relating to this branch. The logs would show any 2 equipment failures and replacement which might possibly 3 relate to lost transactions." 4 So, again, you're making the point, this time a bit 5 more formally in a witness statement, that there has 6 been no request for any data relating to the branch. 7 A. Correct. 8 Q. Was that unusual? 9 A. My understanding was that it was normal when there was 10 a prosecution to request the data to have the basis of 11 what was alleged to have taken place in the branch, but 12 I wasn't very close to things but I thought that was the 13 whole point of the prosecution support team within the 14 Fujitsu, was to provide that data in support of 15 prosecutions. And I assumed it was done in all cases 16 but I didn't really know. 17 Q. And normal to obtain the data relating to the period 18 covered by the charge? 19 A. Whatever period Post Office requested. As I say, 20 I wasn't that close to know exactly what period we were 21 talking about. 22 Q. Did you know whether they normally requested the period 23 covered by the charge? 24 A. No, I didn't know that. 25 Q. You didn't know one way or the other? 23 1 A. Correct, yes. 2 Q. Would it make sense to you that they would request data 3 for the period covered by the charge? 4 A. Yes, I can understand that. 5 Q. Can we go forwards, please, in fact on the same day. 6 FUJ00083722. If we scroll down, please, to the top of 7 page 2, we see a chain there beginning -- I'm not going 8 to go through it all -- from Anne Chambers to others 9 about Callendar Square in 2006; do you see that? 10 A. Yes, I see that. 11 Q. If we just go back up to the bottom of page 1, we see on 12 8 February, so the same date we were looking at that 13 draft witness statement, Anne Chambers essentially 14 forwarding you that chain; is that right? 15 A. Yes, that's right. 16 Q. Do you remember the circumstances in which she forwarded 17 you that chain? 18 A. I think I'd asked her to do so because one of the 19 requests I had -- and if we go back to the witness 20 statement we were just looking at, there's a comment 21 there saying I've asked about Callendar Square and 22 I don't know anything about it but I can do some 23 research, and this was me doing the research. 24 Q. Okay, so you had been asked to comment in a witness 25 statement about the Callendar Square bug, you had no 24 1 personal knowledge and you were trying to find out some 2 information -- 3 A. Yes. 4 Q. -- from Anne Chambers? 5 A. Yes. 6 Q. One of the things she did was forward you this chain? 7 A. Yes. 8 Q. She highlights a KEL and asks you, "Can you see KELs?", 9 and says: 10 "I'd forgotten -- this did give a discrepancy, but 11 also a receipts and payments mismatch, if they persisted 12 and rolled over (though it was usually obvious that 13 something was wrong). 14 "And a flood of NT events (not 'Riposte events'!) 15 which SMC should have noticed at the time. 16 "Since we are now checking for these particular 17 events, and did a catch up for old retrievals, can you 18 say that the current branch did not have this problem?? 19 "Anyway it stopped happening once S90 was installed 20 (around 4 March 2006, according to the info below). 21 "This particular problem would only affect branches 22 with more than one stock unit. It happened several 23 times at Callendar Square, though we never found why 24 they were so badly affected. 25 "Is this sufficient?" 25 1 She said, "I'd forgotten -- this did give 2 a discrepancy", had you asked her whether the bug would 3 produce discrepancies? 4 A. I can't remember the conversation I had at the time. 5 I was just asking for information. 6 Q. That rather implies there had been an exchange or 7 a conversation when she had forgotten that this bug had 8 produced discrepancies? 9 A. Yes, I could well have asked about that. As I say, 10 I just can't remember that conversation. 11 Q. Would you be interested to know whether the bug did or 12 did not produce discrepancies? 13 A. Clearly, because, if there had been a discrepancy, then 14 that would be something important to have reported on. 15 Q. Also, whether the bug was or the effects of the bug were 16 noticeable to the user, the branch user? 17 A. Yes. 18 Q. That would be something that would be important to 19 discover? 20 A. Yes. 21 Q. She asked the question, or she says: 22 "Since we are now checking for these particular 23 events, and did a catch up for old retrievals ..." 24 She then asked the question: 25 "... can you say that the current branch did not 26 1 have this problem??" 2 A. Well, at that stage I couldn't but that was a check that 3 we did later do when we checked the NT events for the 4 West Byfleet branch. 5 Q. Do you agree that there's an implication there that 6 you're looking for ways to find or to say that the bug 7 had no application in Seema Misra's case? 8 A. Yes. 9 Q. Was there a bigger point arising from this email, namely 10 the fact that the Callendar Square bug existed, that it 11 produced discrepancies, meant that before testifying as 12 to the reliability of Horizon accounts, you needed to 13 carry out a thorough review of PinICLs, PEAKs and KELs? 14 A. I thought that what I needed to do was a thorough review 15 of the NT events, which is what I did. 16 Q. I think you reflected earlier in your evidence that 17 you've now realise that something more broad was needed 18 or a broader -- 19 A. Looking back now but, at the time, I thought it was 20 sufficient to just look at the NT events. 21 Q. And Helpdesk records as well, would you include that in 22 the list of things that you ought to have looked at? 23 A. Now, yes; at the time, I felt that that had -- or at 24 least maybe not at this exact time but I was aware at 25 some stage that Andy Dunks had actually done a survey of 27 1 Helpdesk calls and -- 2 Q. We're going to come to those a bit later. 3 Do you agree that, in order to answer the question 4 that had been asked of you, you would need to look at 5 the full message store? 6 A. I would certainly need to look at the ARQ data, whether 7 I needed to look at the full message store or not, I'm 8 not 100 per cent sure. 9 Q. Why aren't you 100 per cent sure? 10 A. I'm not sure what I would have found in the full message 11 store that was not obvious in the ARQ data. It's 12 certainly useful to have the full message store there to 13 look at if I would need to do so, and I made sure I did 14 get it. 15 Q. Winding forwards, you say eventually in your witness 16 statement that you had little personal knowledge of the 17 Callendar Square bug until 2010, when you were asked to 18 find out about it and comment on it in the Misra case? 19 A. And that's what this email exchange is all about. 20 Q. If you didn't have any personal knowledge at all, do you 21 know why you were asked to comment on it? 22 A. No, is the simple answer. 23 Q. Did you question why somebody with no knowledge about 24 an issue was asked to provide a witness statement and 25 then go to court about the issue? 28 1 A. No. I don't know why that was done. 2 Q. Well, did you question why somebody with actual or real 3 experience of the bug in question was not being asked to 4 answer the questions? 5 A. No, I don't know why that was done. 6 Q. Do you agree that, if you had no firsthand knowledge of 7 the bug, you weren't truly in a position to give 8 evidence about it, its effects, and whether or not it 9 had afflicted Mrs Misra's branch? 10 A. I think in the research that I did, I learnt sufficient 11 about the bug to be able to say that it hadn't affected 12 that particular branch. 13 Q. What about the broader issues as to the bug itself? 14 Were you in a sufficiently knowledgeable position to 15 give evidence about that? 16 A. I knew enough about it, I don't think anyone knew what 17 the real underlying cause was because that was something 18 in the Escher software, as we discussed the other day. 19 Q. Can we move forwards, please. POL00054220, and look at 20 page 2, please. If we just look at the foot of page 1, 21 we can see that this is an email from you, I think. 22 There we are: you to Mr Singh, copied to Penny Thomas, 23 and we're on 25 February now. 24 A. Okay. 25 Q. Scroll down. You apologise for a delay in replying and 29 1 say things were hectic and you had a day of leave. Then 2 you're essentially reporting back your communication or 3 conversation with Professor McLachlan, yes? 4 A. Yes. 5 Q. In the fourth paragraph, you say: 6 "I also explained to him how some of Horizon works 7 and why this means that some of his hypotheses were 8 invalid. I also pointed out that in order to identify 9 exactly what was happening, then it would be necessary 10 to go through the detailed logs of the relevant times 11 and that as far as I was aware, no request had been made 12 for any such logs (though I think they may now have been 13 requested)." 14 So it was still the case that you were, by 15 25 February, conducting a conversation with the defence 16 expert without an analysis having occurred of what you 17 call the detailed logs of the relevant times? 18 A. Correct. 19 Q. Was that problematic? 20 A. It really depended what the questions were. So, I mean, 21 I think we had quite a useful discussion in terms of, at 22 a high level, how Horizon was operating but, clearly, in 23 terms of proving some of the detail and some of his 24 hypotheses, we would need to examine the logs. 25 Q. Can we go to the next day, please, 26 February 2010, by 30 1 looking at FUJ00152992. Can we see you send a further 2 email the next day to Mr Singh, copied again to Penny 3 Thomas, and you say: 4 "Jarnail 5 "I've no idea what it is that the defence is looking 6 for in this case or exactly what is alleged to have 7 happened. It is normal to identify a specific time 8 period of about a month to look for some specific 9 fraudulent transactions. As I've no idea exactly what 10 is being alleged I can't really advise as to what 11 evidence might be required either to support the 12 prosecution or the defence. [The Post Office's] 13 Prosecution Support Team have a formal mechanism to 14 request logs for specific periods and there is a process 15 to do that. Although I have suggested for some time 16 that these logs are requested, I understand that no such 17 request has been made to Fujitsu. Trying to analyse 18 transactions over a period of 2 or 3 years ..." 19 There you're referring essentially to the period of 20 the allegations against Mrs Misra; is that right? 21 A. I'm not sure that I was fully aware of the period of the 22 allegation at that stage, but I think I was probably 23 generally aware that we were talking about a long 24 period, not a short period. 25 Q. "... is likely to take several weeks or months of 31 1 effort -- especially if it is not clear what is being 2 looked for -- and I certainly cannot commit that amount 3 of time to it." 4 So, again, you're pressing the point that you've 5 been asking for some time now -- or Fujitsu have been 6 asking for some time now -- that the logs be requested? 7 A. Yes. 8 Q. Can we look, please, at POL00169122, and look at page 3, 9 please -- thank you. This is an email from Jon Longman 10 to Mark Dinsdale but copied to Jarnail Singh, but it 11 records the fact that: 12 "Jarnail Singh has just telephoned me ... to get 13 transaction log data for the above [post office] from 14 1 December 2006 to 31 December 2007." 15 Had you advised or suggested that that limited time 16 period be the extent of the request for transaction log 17 data? 18 A. No. 19 Q. But, in any event, it seems that, at this time or by 20 this time, on the 26th, there had been a request for 21 transaction log data for that limited period? 22 A. So I understand, and I know that's the data that 23 I eventually got. 24 Q. When you eventually got the data for that limited 25 period, did you think, "Why am I getting the data for 32 1 this period"? 2 A. I took the fact that I was being given data for that 3 period, that that was the limit of what I needed to 4 investigate. 5 Q. That's circular. Did you know why you were given data 6 between those two time periods? 7 A. I think I may have been told that that was the period 8 during which false accounting had occurred and there'd 9 been some question about theft. I think I've set that 10 out in my witness statement but I can't remember the 11 exact details. 12 Q. Can we go forwards, please, to what you say about the 13 analysis of that transaction data, FUJ00156128. We're 14 on 3 March now, email from you to Warwick Tatford, 15 copied to Penny Thomas: 16 "Warwick, 17 "... I've added my comments to [the] fifth report." 18 The paragraph below, you say: 19 "I've now also got hold of the transaction logs for 20 December 2006 to December 2007. These amount to a total 21 of nearly half a million transactions ..." 22 Can you see that? 23 A. Yes. 24 Q. So by this time you had the ARQ data, is that right, for 25 the limited period that had been requested? 33 1 A. Yes, and I think I also had the raw logs, but the 2 analysis I did was restricted to what I had in the ARQ 3 data. 4 Q. Why was the analysis restricted to what you had in the 5 ARQ data? 6 A. Because I was looking for specific things that had been 7 mentioned in Professor McLachlan's second report, as 8 I lay out in this email. And this email is saying, 9 "This is the investigation that I have done on the data 10 that I've got", and I also, I believe, suggested other 11 things that could be looked at and I don't think anyone 12 ever took me up on it and said, "Can you please go and 13 have a look at that as well?" 14 Q. So, again, it was a restricted function that you were 15 performing, ie responding to that which Professor 16 McLachlan had alleged? 17 A. Yes. I believe I'd been told that this was all needed 18 very quickly because I believe there was a court date 19 about a week or so later, so there wasn't really time to 20 do anything more thorough than that. 21 Q. Indeed. Can we look at your witness statement, please. 22 It's your third witness statement in the Misra case. 23 POL00001643. Can you see that's your witness 24 statement -- 25 A. Yes. 34 1 Q. -- of 9 March, which is your third witness statement in 2 the Misra case. By this time, you'd got the ARQ data 3 for the limited period that had been allowed and you'd 4 also had the exchange with Anne Chambers about Callendar 5 Square? 6 A. Yes. 7 Q. I think you addressed both of those in this witness 8 statement? 9 A. Yes, and I think I also addressed Professor McLachlan's 10 first report as well. 11 Q. Do you agree in the witness statement it's not disclosed 12 on the face of the witness statement what work you had 13 done yourself and what work had been done by others? 14 A. Yes, because, at that point, I wasn't aware that there 15 was a need to make that sort of distinction, though 16 I realise now that I should have done. 17 Q. If we go forward to page 12, please. You're here 18 dealing with request to comment on Callendar Square. 19 A. Yes. 20 Q. You say: 21 "I have been requested to comment on the issue 22 raised by the defence in relation to a post office 23 called Callendar Square, Falkirk that was mentioned at 24 the Castleton trial. I have examined our records and 25 can confirm the following: 35 1 "The problem occurred when transferring cash or 2 stock between stock units. Note that West Byfleet ... 3 does operate multiple stock units so the issue could 4 have occurred. It manifests itself by the receiving 5 stock unit not being able to 'see' the transfer being 6 made by the 'sending' stock unit and is compounded by 7 attempting to make a further transfer. Please note that 8 such transactions usually reappear the next day. It is 9 clearly visible to the user as the 'Receipts and 10 payments mismatch' at the time that one of the stock 11 units is balanced. This usually results in the branch 12 raising a call. There are no such calls in Andy Dunks' 13 witness statement of 29 January which summarises the 14 calls raised by West Byfleet. Also this can be checked 15 on any balance reports or trading statements that are 16 available from the branch which should show that 17 receipts and payments do [not] match and that the 18 trading position is zero. The problem is also visible 19 when looking at the system events associated with the 20 branch. The system events from 30/06/2005 to 31/12/2009 21 for West Byfleet have been checked and no such events 22 have been found. The problem was fixed in the S90 23 release which went live in March 2006 and so would not 24 have been relevant at the time of the detailed 25 transaction logs obtained for West Byfleet between 36 1 December 2006 and December 2007." 2 Do you agree much of what you said here was 3 assertions made by you, based on what Anne Chambers had 4 told you? 5 A. I had also looked at the underlying PEAK, which had got 6 the analysis that she had done at the time. 7 Q. Is that the reference, if we go back up the page, to 8 "I have examined our records"? 9 A. I think so, yes. 10 Q. So, for example, you say the problem was fixed. Could 11 you say that the problem was fixed? 12 A. I -- the PEAKs said that -- the PEAK was closed with the 13 comment saying that this is -- this fix appears to have 14 done it and the problem has not reoccurred since the fix 15 has gone in. 16 Q. That's slightly different from asserting in a witness 17 statement that the problem was fixed, wasn't it? 18 A. I didn't think so at the time. 19 Q. Looking back now, do you appreciate the difference? 20 A. Okay, yes, I probably appreciate the difference now but, 21 at the time, I felt that it was the same thing. 22 Q. If you carry on reading, over the page, you say: 23 "Therefore I can conclude that the problems 24 identified in Callendar Square ... are not relevant to 25 West Byfleet ..." 37 1 Then you go on to say: 2 "On 2 October 2009 I produced a report about Horizon 3 data integrity. Within this report are details about 4 transactions (sometimes known as EPOSS transactions) and 5 various scenarios that could occur following system 6 failures. In rare circumstances it is possible for 7 transactions to not be recorded on the local system but 8 in all such cases the user would be aware of this. 9 I produce this report as [your] exhibit GJ/01." 10 That's your Horizon data integrity report, the first 11 one, that we looked at yesterday -- 12 A. It is. 13 Q. -- is that right? 14 A. Yes. 15 Q. The cross reference -- no need to go there -- is 16 FUJ00080526. 17 Why were you producing that report in the context of 18 the Misra case? 19 A. Because I thought it gave a useful summary in section 3 20 of the sort of hardware errors that could occur that 21 could possibly cause loss of data. 22 Q. You continue: 23 "As with any large system, there will be occasional 24 failures, such as the one found in Callendar Square, 25 Falkirk. Any such faults, whether during testing or 38 1 from live user feedback would be investigated and 2 resolved appropriately. I am not aware of any such 3 faults that have been raised by West Byfleet. If 4 specific transactions can be identified where the user 5 feels the system has caused losses then further 6 investigation can be made." 7 Was that last paragraph as close as you ever came to 8 answering the broad question that originated from 9 Mr Tatford, namely whether there were any known problems 10 with the Horizon system that Fujitsu are aware of? 11 A. That was probably addressing that request, yes. 12 Q. You make your answer conditional, don't you: any such 13 faults, whether found during testing, et cetera, would 14 be investigated and resolved appropriately, don't you? 15 A. That was my belief. 16 Q. But it's conditional, isn't it? It suggests to the 17 reader that there weren't any which Fujitsu actually 18 knew about, other than Callendar Square? 19 A. And I believed I was confident of that because, by this 20 time, I had actually looked at the NT event logs at the 21 time and I would have expected, if there had been 22 faults, for there to be evidence in the NT event logs to 23 reflect those. 24 Q. The first part of the paragraph is dealing with Horizon 25 more generally though, isn't it? 39 1 A. Yes. 2 Q. Not the branch in question: West Byfleet? 3 A. Yes. 4 Q. Would you agree that a complete answer would say: There 5 are many known problems with Horizon. Fujitsu keeps 6 records of them in documents called PinICLs, PEAKs and 7 KELs? 8 A. I can understand that now. I don't think I would have 9 thought of it that way at the time. 10 Q. By answering the question in this way, you did not 11 disclose the existence of PinICLs, PEAKs and KELs, did 12 you? 13 A. I didn't know that I needed to. 14 Q. By answering the question in this way, conditionally, 15 "Any such faults would be investigated", rather suggests 16 that the only one is Callendar Square, doesn't it? 17 A. But certainly later on in the process, I did make it 18 clear that there were PinICLs, PEAKs and KELs, and no 19 one asked me to say anything about those in any sort of 20 statement. 21 Q. In any case, you say that any faults would be 22 investigated and resolved appropriately? 23 A. That was my belief. 24 Q. You believed in the life of Horizon all faults had been 25 resolved appropriately? 40 1 A. That was my understanding. 2 Q. What investigation did you carry out in order to be able 3 to confidently assert that? 4 A. I'm not sure that I did any specific investigation at 5 the time. 6 Q. Thank you. 7 Can we turn to a different topic, please, whether 8 you mentioned in your witness statements your knowledge 9 of other bugs -- 10 A. Okay. 11 Q. -- by looking to start with, please, at FUJ00117478. 12 This a document I don't think we've seen before or 13 much before. It's prepared by you, correct, on 14 29 January? 15 A. Yes, but this relates to Horizon Online, not Legacy 16 Horizon. 17 Q. I'm going to look at a series of Horizon Online -- just 18 so you know where I'm going -- issues? 19 A. Okay. 20 Q. The heading tells us what it's about, the "Issue of 21 Duplicate Statements found at Derby": 22 "The purpose of this note is to describe the issues 23 found at Derby this week with duplicate settlements and 24 present options for fixing the defect[s]." 25 So it sets out the problem and this had led to 41 1 duplicate transactions; is that right? 2 A. Yes, I need to put this a bit of context. 3 Q. Oh, please do. 4 A. We're talking now about the pilot of Horizon Online. At 5 that stage it was operating in about a dozen/couple of 6 dozen branches, so we were expecting to get some initial 7 problems, and we did, as expected, get a number of 8 problems during the pilot phase. But, as far as I'm 9 aware, the problems were all resolved before we actually 10 rolled out Horizon Online later on that summer. 11 Q. Okay, that's important context which may -- 12 A. Very important. That's why I wanted to bring it out. 13 Q. That may apply to a number of the issues that I'm going 14 to ask you about. 15 A. I suspect it will apply to all of them but we'll see. 16 Q. You, at the foot of the page -- if we just go over the 17 page -- suggest some solutions/fixes which are either 18 a tactical solution or a strategic solution, yes? 19 A. Yes. 20 Q. If we look at the foot of the page, second paragraph 21 under "Strategic", you say: 22 "The change is a straightforward code change ... 23 However, full regression testing of the counter is 24 recommended." 25 A. Yes. 42 1 Q. What does that mean, "full regression testing of the 2 counter"? 3 A. That before we would put that change into live, then we 4 would need to go through a regression testing cycle, 5 which would take a few days, so it was a case of how 6 quickly were we going to get the fix in place, and so, 7 therefore, there may be some benefit in doing the 8 tactical solution first and then the strategic solution 9 later, so as to actually make sure that other branches 10 were -- had the minimum impact of the problem. 11 Q. Got it. Can we go forward, please, to 8 February. 12 FUJ00092922. 13 These are some notes of a Horizon Next Generation 14 implementation meeting on 8 February, and we can see 15 that you are present. 16 A. Yes. 17 Q. This is to discuss the Derby issue, or the Derby bug 18 issue. Yes? 19 A. I think there were -- it's not just the Derby bug. 20 I think other things were being discussed as well. 21 Q. Are they the issues at Coton, Warwick? 22 A. Yes, I can't remember what those issues were now but we 23 had a number of issues during the pilot period. 24 Q. I think the note records that what was called the Derby 25 bug had manifested itself at Coton and Warwick? 43 1 A. Ah, okay. 2 Q. Does the fact that this group of people were drawn 3 together reflect the fact that the Derby bug was 4 a serious issue? 5 A. Yes, because it affected the accounts and, obviously, we 6 didn't want any problems that affected the accounts in 7 Horizon Online moving forward. 8 Q. Can we go forwards, please, to FUJ00093031. This is the 9 next day, 9 February. I don't think you're on this 10 circulation list. 11 A. No, but I think it was forwarded to me later. 12 Q. Did the Derby bug issue lead to a full Fujitsu internal 13 audit, essentially? 14 A. Yes. It took place the previous week when I was off 15 sick, which is probably why I wasn't actually involved 16 in the circulation of it, because the auditors wouldn't 17 have talked to me. 18 Q. If we just look at the foot of page 1, please. The last 19 paragraph. It says: 20 "The net effect [I think that's of the bug] would be 21 that the Post Office and the branch records would not 22 match. Where this happens, the Post Office investigates 23 the branch and postmaster, with a view to retraining or 24 even uncovering fraud. It would seriously undermine 25 Post Office credibility and possibly historic cases if 44 1 it could be shown that a discrepancy could be caused by 2 a system error rather than postmaster/clerk action. 3 More importantly, the central database as the system of 4 record would be called into question." 5 Is what is recorded there reflective of the views 6 that you held at the time about the bug? 7 A. I was certainly -- my view was that these bugs all 8 needed to be fixed before we could put any reliability 9 into the way that Horizon Online operated. I didn't see 10 that it had been particularly important in terms of 11 prosecution. What we needed was a system that correctly 12 handled accounts and, until we had a system that 13 correctly handled the accounts, then we needed to do 14 that moving forward, and I believe that that's what 15 happened by the time the rollout of Horizon Online 16 happened. 17 Q. Why would you not have seen it through the complexion or 18 the lens of it affecting prosecutions? 19 A. Because I didn't think we should be going anywhere near 20 prosecutions, in terms of the state of that system, 21 until we actually had the system that was stable and 22 operating correctly. Prosecutions was -- the important 23 thing was to have the system operating correctly. 24 Q. Do you agree the existence of the bug, if revealed, 25 would seriously undermine the Post Office's credibility? 45 1 A. If the bug was left unfixed yes, but my approach was 2 much more to do with fixing the bug, rather than any 3 suggestion of not fixing it and that was where I was 4 coming from: that the important thing to do was to get 5 these bugs fixed and so that we actually had a properly 6 operating system. 7 Q. Do you agree that the existence of the bug would 8 possibly undermine historic cases? 9 A. Don't see that it was anything to do with historic cases 10 because, as I say, the important thing was to get the 11 thing fixed, rather than leave the thing unfixed, and 12 that was what I was focusing on. 13 Q. Do you know how it was said, then, in this document, 14 that the existence of this bug in Horizon Online could 15 possibly undermine historic cases? 16 A. I didn't see that it could affect the historic cases 17 because Horizon Online was a totally different system 18 from the Legacy Horizon system, as far as the accounting 19 was concerned. So I didn't see that it would have that 20 sort of impact. 21 Q. The paper explains it in terms of that that is because 22 it could be shown that a discrepancy could be caused by 23 a system error, rather than the postmaster or clerk 24 action? 25 A. I see that's what it says. I wouldn't have necessarily 46 1 agreed with that, but my focus, as I say, was on 2 actually getting the bugs fixed not on what the impact 3 they would have had if they hadn't been fixed. Because 4 it was clearly very important that these bugs were 5 fixed. 6 Q. You didn't reveal the existence of this bug in your 7 March 2010 witness statement or any of the subsequent 8 witness statements, did you? 9 A. No, because it was totally irrelevant. 10 Q. Was that the view: because this was about Horizon 11 Online -- 12 A. Yes. 13 Q. -- you needn't reveal anything about Horizon Online 14 in -- 15 A. It was a totally separate -- oh, it -- it was a totally 16 separate system. 17 Q. Was that your mindset at the time? 18 A. Yeah. 19 Q. "I needn't tell the court or the defence anything about 20 the existence of bugs that caused financial 21 discrepancies, which aren't revealed to a subpostmaster 22 in the system we're currently operating, because those 23 facts cannot be relevant to the court?" 24 A. Yes, and that is still my technical belief. 25 I understand now that the legal position is different 47 1 from that but, as a technician, which is basically what 2 I am, not a lawyer, then I still believe that that's the 3 case. 4 Q. But you were stepping out of the computer lab and 5 walking into a court, weren't you? 6 A. I didn't realise that there was that difference that 7 I needed to worry about and no one advised me as to 8 that. 9 Q. Did you think, "I need to take a bit of advice here. 10 There are problems with the current system where, as I'm 11 making these witness statements in the Seema Misra case, 12 we're discovering problems which don't reveal themselves 13 to subpostmasters, which cause financial discrepancies, 14 and they're caused by bugs. Have I got to reveal that; 15 should I reveal that in a candid way to the court"? 16 A. It never occurred to me that anything to do with Horizon 17 Online was relevant to Legacy Horizon. 18 Q. You said that -- 19 A. Yeah -- 20 Q. -- that it never occurred to you, and so it didn't occur 21 to you to reveal it but it didn't occur to you to ask 22 whether you should reveal it? 23 A. No, it just seemed so totally illogical to me. 24 Q. You didn't know how the legal system worked -- 25 A. I -- 48 1 Q. -- you tell us -- 2 A. I assume the legal system had some sort of logic to it. 3 I still don't understand why the legal system would 4 think that I should be revealing problems to do with 5 Horizon Online in Legacy Horizon. Okay, I understand 6 now, having been told that I should have considered it 7 but it just doesn't actually make an awful lot of sense 8 to me as a technician. 9 Q. Again, I'll ask the question: do you not realise that's 10 the very reason why you should seek advice? 11 A. I didn't ... well, I had sought advice from -- 12 Q. Not about this issue? 13 A. Well, not this specific issue but I had sought advice in 14 terms of what I should be saying from David Jones. 15 Q. Yes, but not about the broader issue of "I've got this 16 knowledge of many bugs, including bugs that are 17 revealing themselves as we roll out or as we develop 18 Horizon Online, have I got to reveal those too?" 19 A. It just never occurred to me that it was at all 20 relevant. 21 Q. Just remember that last paragraph, there, in particular 22 the sentence, "It would seriously undermine Post 23 Office's credibility and possibly historic cases", and 24 can we look at FUJ00094392. Look at the foot of the 25 page, please. This is an email exchange, if we just go 49 1 up -- and a bit more, thank you. 2 This is a discussion, you're not party to this 3 discussion, but I just want to know whether you knew 4 about it. This is a discussion about amendments being 5 made to that document we just looked at. You understand 6 Mr Jenkins? 7 A. Yes, I understand that. I can't remember if I was 8 copied in on this email or not. I think -- I may 9 possibly have been passed it before but I certainly 10 wouldn't have taken much notice of it. 11 Q. If we look at the bottom of the page, thank you, just 12 a bit further, I think we see it's signed off by "JP"? 13 A. Yes. JP gave me some advice and was in contact with me 14 at the time that I was at the Seema Misra trial. I also 15 think he phoned me up a couple of times to see how I was 16 getting on, and so on, and he certainly never mentioned 17 to me anything about why I needed to think about the 18 problems such as this, which he was well aware of. 19 Q. Just focusing on this for a moment, he says to 20 Mr D'Alvarez: 21 "... please find below the two comments I had raised 22 ..." 23 A paragraph: 24 "First paragraph, page 2 -- the following section is 25 potentially problematic: 'It would seriously undermine 50 1 Post Office credibility and possibly historic cases if 2 it could be shown that a discrepancy could be caused by 3 a system error rather than a postmaster/clerk action. 4 Most importantly the central database as the system of 5 record would be called into question'. As discussed, 6 there is no need to paint this in the worst possible 7 light. I would suggest the following as being accurate 8 without being unduly alarmist ..." 9 Then the new drafting is: 10 "If it could be shown that a discrepancy could be 11 caused by a system error rather than a postmaster/clerk 12 action, it could potentially call into question the 13 effectiveness of the central database as a system of 14 record." 15 If we scroll up, please: 16 "After review with Legal ... two amendments have 17 been made to the documents and ... the sections ... have 18 been removed as they do not materially add to the 19 primary purpose of the review which is to determine 20 whether the solution, as designed, would protect data 21 integrity. 22 "Version 3 has been sent to Dave Smith, Post 23 Office." 24 So this seems to record that the sentence about data 25 integrity affecting possibly past historic cases appears 51 1 to have been removed before it was sent to the Post 2 Office; do you agree? 3 A. I agree that's what this says, yes. 4 Q. If we just go to the top of the page. I think we see 5 that exchange was sent on to you? 6 A. Yes. 7 Q. Do you know why you were being told that the report had 8 been edited to remove the section that said that this 9 discovery of the bug in Horizon Online might affect past 10 historic cases -- 11 A. I don't think -- sorry -- 12 Q. -- was going to be removed in the version that was being 13 sent to the Post Office, why that chain was being sent 14 to you? 15 A. I would have taken that as just being sent the latest 16 version of the document, rather than taking any notice 17 of the -- of that particular discussion. 18 Q. Not because they knew that you were giving evidence 19 about past historic cases? 20 A. Not that I thought of at the time. 21 Q. No. 22 SIR WYN WILLIAMS: Mr Beer, have I got this right: this 23 seems to have been a period of, what is it, seven or 24 eight months, between the exchange about the contents of 25 the document and it actually being forwarded to Mr -- 52 1 MR BEER: No, this is one of those dates where it has been 2 Americanised. That's actually -- 3 SIR WYN WILLIAMS: I see. It's actually the other way 4 round -- 5 MR BEER: -- 11 March. 6 SIR WYN WILLIAMS: -- 11 March, okay, fine. 7 MR BEER: Can we go forwards, please, to FUJ00142152. 8 This is a further report of yours, a few days later, 9 12 February 2010, ie a few days later than the Derby bug 10 issue report, okay? 11 A. The original one. I mean, the one you just showed me 12 was March but the original probably yes, yes. 13 Q. You wrote this report and, in summary, is this about 14 another bug which resulted in final balances for branch 15 trading statements being inaccurate? 16 A. Yes. 17 Q. This was first reported at the Post Office in Warwick; 18 is that right? 19 A. I believe so, yes. 20 Q. If we look at FUJ00094268, and the third page of that, 21 we'll see that you sign off an email there and, if we 22 scroll up, please, is this is an email discussion 23 between you and others about the issue that we've just 24 looked at in that report, ie a bug resulting in trading 25 statements being inaccurate? 53 1 A. I believe so, yes. 2 Q. If we scroll down, can you see a passage that's 3 indented, which says: 4 "Given the legal status of these reports (they 'the 5 final' are often used in court proceedings when we are 6 trying to recovery monies from dismissed subpostmasters) 7 and the potential 'integrity' challenges that could be 8 levelled against the [Horizon] system as a result of any 9 differences; POL need assurance from Fujitsu that they 10 could/will explain the cause of the problem and prove 11 the system has integrity should we be challenged." 12 That's not you speaking; is that right? 13 A. No, that's Phil Norton, I think it was, who was sent the 14 email. So, again, I've got his questions indented and 15 my responses not indented. 16 Q. You say: 17 "I understand this, though I would have thought most 18 legal proceedings would be based on the first part of 19 the report (which covers cash levels) rather than the 20 second part." 21 Irrespective of that answer, you were being told by 22 Mr Norton and being left in no doubt as to the 23 importance of the accuracy of the Branch Trading 24 Statements for legal proceedings. 25 A. I was aware of that, and my approach was: we've got to 54 1 fix this problem. We certainly could not leave that 2 problem unfixed for Horizon Online to move forward and 3 be rolled out, and so, therefore, it was important that 4 the problem did get fixed, which it did. 5 Q. Did you know at this time that, at least since the Lee 6 Castleton case, the Post Office relied on the implicit 7 accuracy of Branch Trading Statements when bringing 8 civil claims against subpostmasters? 9 A. I'm not sure that I was really -- had any distinction 10 between civil or non-civil claims but I knew that the 11 Branch Trading Statements, once that came about as part 12 of IMPACT, was the -- effectively the legal document 13 that was used to say that the accounts had been signed 14 off. 15 Q. When this bug arose, did you consider the possibility 16 that there may have been other bugs which affected the 17 accuracy of Branch Trading Statements? 18 A. I thought -- I could see that this bug was causing 19 a problem in the accuracy of that part of the Branch 20 Trading Statement report and felt that it was important 21 to actually get the bug fixed. 22 Q. Did you consider the possibility of other bugs which 23 affected the accuracy of Branch Trading Statements? 24 A. We would look for other bugs there and I'm not aware 25 that we actually ever found any other -- 55 1 Q. Did you look for other bugs there? 2 A. Yes, we were looking for all sorts of bugs that could be 3 happening during the pilot. That was the whole point of 4 the pilot. 5 Q. Can we move forwards, please, to FUJ00094235, and in 6 sheet 1, and row 20, please, can you see -- first of 7 all, do you know what this document is; do you remember 8 what the document is? 9 A. I think this is some sort of tracker of issues during 10 the pilot of Horizon Online. 11 Q. Number 20, "BTS reports incorrect", the narrative, if 12 you just read that to yourself. Can you see that, in -- 13 A. Yes. 14 Q. -- column I. 15 I'm so sorry, it's the one above. Thank you. So 16 you should really read column B and column I. (Pause) 17 A. Right. I can't remember that problem. 18 Q. Is this another problem, a separate one from the two 19 that we've just looked at? 20 A. Yes, this was a problem -- as part of the migration of 21 a branch from Legacy Horizon to Horizon Online, a report 22 was produced on the last day it was operating on Legacy 23 Horizon, which was giving effectively a snapshot of the 24 balance, and then another report was generated the 25 following morning showing, again, the snapshot of the 56 1 balance, and these two reports were compared. And this 2 looks likely to be some sort of issue in that comparison 3 but I can't remember the details of that particular 4 problem. 5 Q. It looks like it's a discrepancy bug where stock 6 movements weren't included in the receipts part of the 7 stock units balance report? 8 A. I -- sorry, I just can't remember the detail. I can see 9 that's what the words say but I cannot remember this 10 particular problem. 11 Q. The words say that the data that was "included in 12 Horizon was incorrect", whereas the data shown in 13 Horizon Online is incorrect -- 14 A. That -- 15 Q. -- sorry is correct. 16 A. That suggests to me that there was a problem in the 17 report that was produced on the Horizon side, on the 18 date before we migrated, but the data that was on the 19 Horizon Online version was actually accurate and, 20 therefore, there would be a mismatch between the two 21 reports, and the reason for the mismatch was because the 22 data had been calculated inaccurately on the Horizon 23 side, but it was correct on the Horizon Online side, is 24 what I think I'm reading from this. 25 Q. I see. 57 1 A. But, as I say, I think I would need to do further 2 research to say anything more than that. 3 Q. So it seems like it's a problem with Legacy Horizon, 4 rather than Horizon Online? 5 A. It's a problem with the reporting of the situation on 6 Legacy Horizon, so I don't think it's actually a problem 7 in Legacy Horizon, as such, it's purely in the way that 8 that report is being generated. There was a special 9 report that was produced as part of the migration 10 process, I believe. 11 Q. You'll see that the recommendation on the right-hand 12 side, under "17 February 2010", is not to fix the 13 error -- 14 A. Yes. 15 Q. -- and, instead, to raise a KEL for coverage. What does 16 that mean, "KEL for coverage"? 17 A. I don't know, is the simple answer. 18 Q. Was that an expression used, "raise a KEL for coverage"? 19 A. I've not heard of that expression before. 20 Q. Then a KEL is issued and then it's closed. 21 A. Yes. As part of the migration process, there was 22 an automated check made of the electronic versions of 23 these reports because they're printed in the branch, but 24 there was a -- it was stored away. So there was 25 a comparison, a before and after report. So I'm 58 1 assuming that, as part of the migration process, some 2 sort of alert would have been raised if the reports 3 didn't match and so the purpose of the KEL was that if 4 those reports didn't match and it was down to this 5 particular reason, then it would then be known about, 6 and therefore understood. But -- 7 Q. You're assuming a lot there? 8 A. I am assuming a lot, but to say more than that I would 9 need to go back to the individual PEAKs. 10 Q. Do you recall making any checks to see whether this bug 11 had affected the West Byfleet branch? 12 A. That wouldn't have had any impact. I wouldn't have made 13 any checks because this would have had no impact on the 14 West Byfleet branch because that was still running on 15 Legacy Horizon. This was purely a migration issue, this 16 one. 17 Q. Did it affect the data that was moved over to Horizon 18 Online? 19 A. I don't believe so, no. 20 Q. What was, on your understanding, the impact of the bug, 21 then? 22 A. It was purely to do with the contents of the report 23 produced before and after the migration. 24 Q. A Branch Trading Statement report? 25 A. No, this was a migration report, so it wasn't a Branch 59 1 Trading Statement -- you didn't -- it was decided that 2 it was not sensible to actually get branches to actually 3 produce a Branch Trading Statement on the day they 4 migrated, because that would restrict very much the rate 5 at which migration could take place. So there was just 6 a snapshot that was produced at the time that the 7 branches migrated from Legacy Horizon to Horizon Online. 8 Q. Thank you. That can come down. 9 So when, in early March 2010, you signed your third 10 witness statement in Seema Misra's case, you knew about 11 a bug in Horizon Online that had affected system 12 created, system generated discrepancies. You knew about 13 the so-called Warwick bug, which resulted in inaccurate 14 Branch Trading Statements in Horizon Online, and that 15 there was a stock unit discrepancy bug in which affected 16 migration reports. You didn't reveal any of those three 17 things in your witness statement, did you? 18 A. But as I said before because I thought they were totally 19 irrelevant. 20 Q. But you, in that paragraph we looked at, were you trying 21 to answer Mr Tatford's general question about known bugs 22 in Horizon? 23 A. I was trying to address problems -- I assumed he was 24 referring to Legacy Horizon, as it operated at the time 25 that Mrs Misra was operating the branch in West Byfleet. 60 1 I didn't see that a problem in Horizon Online in 2010 2 was relevant to something that had taken place in 3 2006/2007, which is what I had been looking at. 4 Q. Why did you assume that his question, which I think you 5 agreed first thing this morning, was broadly drawn and 6 did not have the caveats that you've introduced into it? 7 A. I didn't see any relevance of the -- Horizon Online to 8 the situation we had there. 9 Q. So you inserted the limitations to the question in the 10 way that you chose to answer it? 11 A. I think that's putting it slightly strongly but 12 I just -- it never occurred to me that there was any 13 relevance of that whatsoever. 14 MR BEER: Thank you. 15 Sir, that's a good time for the morning break. Can 16 we take ten minutes, please, until 11.25. 17 SIR WYN WILLIAMS: All right. 18 MR BEER: Thank you. 19 (11.15 am) 20 (A short break) 21 (11.26 am) 22 MR BEER: Mr Jenkins, can we finish off the topic we were 23 just looking at by examining an email that you wrote in 24 September 2010 concerning, essentially, these three 25 Horizon Online bugs that emerged in testing in the early 61 1 part of 2010, and look at FUJ00156217. 2 If we go to the last page, please, and then if we 3 just scroll up. You're talking here about a PEAK and is 4 it SYSMAN or SYSMAN3 Events? 5 A. SYSMAN is how I'd call it, it's short for "system 6 management". 7 Q. What was this, the System management 3 events? 8 A. Basically NT events. So Legacy Horizon used a subsystem 9 called SYSMAN2, whilst Horizon Online had a slightly 10 different version of the technology which was called 11 SYSMAN3. 12 Q. Thank you. If we scroll up, please, and if we keep 13 going -- and keep going. Thank you, just stop there. 14 You're emailing Penny Thomas and others, saying: 15 "Steve Porter has had a look at these events ... and 16 his responses are below." 17 I'm not going to go to those responses. You say: 18 "Unfortunately they are a bit inconclusive, but 19 there is nothing that is obviously suspicious. 20 "Given that this relates to [Horizon Online] in 21 March, [then you say this] then we do know that there 22 were a number of potentially serious issues around at 23 that time, so we would need to be wary about making any 24 witness statements associated with such data. However 25 it should be fine to pass the data to [Post Office] with 62 1 an ARQ. 2 "We're going to need to tread fairly carefully for 3 ARQs in this area." 4 Can you explain what you meant, please? 5 A. Part of the process of extracting ARQ data was to check 6 any NT events that were associated with it, and what 7 I believe had happened here is that there had been 8 a request for ARQ data for this particular branch, and 9 some NT events had come back in the search that looked 10 a bit odd. And so, therefore, what I was suggesting is, 11 yes, we could pass the information to Post Office, but 12 we needed to actually put on a caveat saying that, in 13 this particular case, there was potentially going to be 14 some sort of problem with that. 15 Q. You don't say, "We're going to need to put on a caveat 16 when we pass the ARQ data to the Post Office"? 17 A. I think I was implying that with what I was saying, and 18 I believe that is what actually happened. 19 Q. Is it, in fact, what happened: you put caveats on when 20 the ARQ data -- 21 A. I think there was a subsequent email exchange where 22 I believe Tom Lillywhite told Post Office, "Sorry, we 23 can't guarantee the ARQ data for this particular case 24 because we've come up with these events". 25 Q. What about the passage that says, "We need to be wary 63 1 about making witness statements associated with such 2 data", the potentially serious issues around that time, 3 do those include the three bugs that I've drawn to your 4 attention this morning? 5 A. Yes, I would certainly not have been happy about any of 6 that data from that pilot period being used in any 7 sort -- anywhere near a prosecution because the system 8 was not stable at that time. 9 Q. So "wary about making witness statements", means what: 10 "we should not make a witness statement relying on such 11 data"? 12 A. Yes, basically. 13 Q. "We should tell the Post Office that you should not 14 confidently prosecute anyone in reliance on such data"? 15 A. That was my feeling at the time, yes. 16 Q. Thank you. Did you consider whether that kind of 17 sentiment or view ought to have been revealed when you 18 were making your witness statements in the Misra case, 19 "look, it can happen in the operation of the Horizon 20 system that data is produced upon which reliance by 21 a court should not be placed"? 22 A. No, because this was a totally different circumstance. 23 Here we were talking about a pilot of Horizon Online 24 whilst, by the time -- the time period we're talking 25 about for Mrs Misra's case, then we actually had what 64 1 I felt was a stable system. 2 Q. Thank you. Can we turn to a separate issue: the 3 duplication of transaction records contained in ARQ 4 data? 5 A. Yes. 6 Q. Can we start by looking at FUJ00097058. This is 7 a document, I think you know, authored by Penny Thomas? 8 A. Yes. 9 Q. Its date is 22 June 2010 and it records duplicate record 10 issues in ARQ returns. Can you explain, in layman's 11 terms, what the problem was? 12 A. Yes, when ARQ data was being collected from Legacy 13 Horizon, there were certain circumstances in which we 14 might record the same transaction more than once into 15 the audit server. So there could be multiple copies of 16 the same message recorded in the audit server. And the 17 way in which ARQ data was extracted from that, in terms 18 of producing the spreadsheets, on Legacy Horizon 19 discarded these duplicates. 20 Q. It was supposed to discard it? 21 A. On Legacy Horizon, it did. 22 Q. Yes. 23 A. What we had here was, when we moved to Horizon Online, 24 from the same raw data, we had a different mechanism for 25 extracting the ARQ spreadsheets and it was found that 65 1 the mechanism that was used for generating the ARQ 2 spreadsheets that was now in use with Horizon Online did 3 not detect and discard these duplicates that were in the 4 raw audit trail. 5 Q. Thank you. If we just scroll down, please, numbered 6 point 1 identifies the issue as urgent in "CS prayers"; 7 what were they? 8 A. It was a regular meeting. So "CS" stands for Customer 9 Services. 10 Q. Okay, so a morning type meeting? 11 A. Yes. They used to be referred to as "prayers", I'm not 12 quite sure why, but it was really a case of 13 a get-together first thing in the morning to work out 14 what the agenda of the urgent issues for the day was. 15 Q. I think we see in numbered paragraph 4, a separate issue 16 was identified where a seemingly duplicated transaction 17 had a different NUM. That's some form of identifier; is 18 that right? 19 A. Yes, each record should have a unique identifier which 20 was actually in three parts. So there was the FAD code 21 for the branch, the counter position, and then a serial 22 number, which started at 1 when the branch was first 23 installed and, by this time, was up into the millions 24 probably, relating to the number of records that had 25 been produced for each counter. 66 1 Q. So that's an additional problem on top of the problem? 2 A. No, this isn't for a problem at all. 3 Q. Right. 4 A. What this is all to do with is if you actually look at 5 the data without taking account of the NUM, you may have 6 things that look like they're identical data but they 7 aren't actually genuine duplicates -- sorry, they are 8 not duplicates, they are actually two real records that 9 look almost identical. 10 Q. If we go to point 6, please -- if we just go back up: 11 "We need to identify which cases provided with ARQ 12 returns since the [Horizon Online] application has been 13 live have progressed to prosecution and identify whether 14 duplicate records were included. We will need [Post 15 Office] involvement to ensure all instances are covered. 16 A very quick review identifies that both West Byfleet 17 and Porters Avenue are included here, to what extent is 18 not yet known." 19 Do you know why West Byfleet was analysed? 20 A. I think Penny had been going through all the ARQ returns 21 and clearly we had produced ARQ data for West Byfleet, 22 using the Horizon Online extraction mechanism back in 23 March, as we discussed earlier this morning. 24 Q. So was it -- it was relevant, this problem, to the 25 prosecution of Mrs Misra? 67 1 A. Yes, it was, which is why I produced a separate witness 2 statement explaining the problem. 3 Q. Why was the Post Office involvement necessary to ensure 4 that all instances were covered? 5 A. I think it was a case of doing a double check that we 6 knew exactly which ARQ returns were involved in 7 prosecutions because sometimes Post Office would ask for 8 ARQ data that wasn't relevant to prosecutions. So 9 I think it was -- Penny had done an analysis of all the 10 ARC returns that she'd done using this mechanism but she 11 wanted to get Post Office involved to make sure that she 12 hadn't missed any and understood which ones were the 13 high priority ones to sort out, so as to get some sort 14 of workload scheduling. 15 Q. So it wasn't the case Fujitsu could not identify all 16 incidents? 17 A. No. Penny had, as I understand it -- as I say, I don't 18 have personal knowledge of that, but my understanding 19 was that she had good records of every ARQ that she had 20 ever extracted and when she produced it, and so on, and 21 she could go through those records. I think there is 22 an email a day or two later where she actually produces 23 some statistics. 24 Q. Can we look at FUJ00097047, and start at page 5, please. 25 We see that she sends her report we've just looked at -- 68 1 yes -- 2 A. Yes. 3 Q. -- through to you and others on 23 June. Then if we 4 scroll up, please, Penny says, "Here's some analysis". 5 This, I think, is a day later. 6 A. Yes, so these are the statistics that I mentioned a few 7 minutes ago. 8 Q. Yes: number of ARQs affected; number of ARQs where one 9 or two instances highlighted which indicates bona fide 10 activity -- do you know what that means? 11 A. Sorry, I don't. 12 Q. Number of ARQs work in progress, presumably? 13 A. Yes. 14 Q. 12 ARQs where court action is known, number of cases is 15 two. Did that include West Byfleet? 16 A. I'm assuming that's Porters Avenue and West Byfleet that 17 were mentioned in the previous note but I'm not 100 per 18 cent of that. 19 Q. 8 ARQs returned where witness statement requested but 20 not yet provided, number of cases was three; and ARQs 21 where no court activity is known is 76. 22 Then scroll down: 23 "Audit development are currently working on a fix 24 which is expected to be available by [the 29th]." 25 Then you're said to have suggested "the following 69 1 explanation" for the Post Office. 2 A. Yes. 3 Q. Did you suggest that explanation? 4 A. I believe so and, hopefully, that is pretty well what 5 I just said a few minutes ago. 6 Q. If we scroll on. She says she wanted to speak with her 7 counterpart in the Post Office, presumably -- 8 A. Yes. 9 Q. -- comments please. Then if we scroll back up, please, 10 to the next page, Guy Wilkerson asks: 11 "Would the additional transactions make any 12 difference to the charges for a subpostmaster?" 13 Then keep scrolling. She replies: 14 "These are original records which have been 15 duplicated when copying to audit server. We are not 16 suggesting that original records have been duplicated." 17 That, of course, is correct, isn't it? 18 A. Correct, yes. 19 Q. "If analysis was undertaken on the audit data some 20 transactions would be duplicated; both plus and minus 21 (we hope!). Analysis on stock units could be out as TI 22 would show duplicated transfers and equally would TO." 23 What do both of those mean? 24 A. Transfer in and transfer out, so it's to do with 25 transferring cash between stock units. 70 1 Q. "Cash on hand analysis would also be out. 2 "Gareth -- is there anything else I need to add?" 3 Scroll up. You say that covers it: 4 "... there is no guarantee that the duplicates are 5 even complete sessions in which case the sum of all 6 transactions may even be out. 7 "In summary, any detailed analysis of the finances 8 of a branch which is done with duplicate transactions 9 without realising that there are duplicates (and so 10 removing them) will give incorrect results." 11 That would be a serious issue, wouldn't it? 12 A. Oh, yes. 13 Q. Then scroll up -- and then keep going, and keep going. 14 Geoff Butts says: 15 "... do not make any communication with [Post 16 Office]. We've been looking ... and are waiting [for 17 a workaround]." 18 Scroll up, and scroll up, Penny says that she's not 19 going to communicate with Post Office: 20 "I do have questions ... 21 "My assumption that any records which have been 22 duplicated and presented to court will need to be 23 replaced. The 2 forthcoming, high profile cases, West 24 Byfleet and Porters Avenue, immediately spring to mind. 25 "We have only been presenting duplicate records 71 1 since the beginning of the year and, although not 2 an impossibility, it is unlikely that we would have 3 provided records for any cases heard in such a short 4 time frame. I can find no record of any witness 5 statement provided, apart from those listed ... we need 6 confirmation ... from POL; and guidance from Legal. 7 "All of the returns listed may not end up in court, 8 but we need to be aware that a significant amount may 9 need replacing." 10 If we go back down to your explanation, I don't 11 think that includes the point that was made originally, 12 that you had spotted, the problem with a duplicate with 13 a different NUM. Instead, it refers to the NUM as being 14 a reliable way to identify a duplicate, doesn't it? 15 A. I think you may have misunderstood what the other issue 16 was. 17 Q. Okay. 18 A. The other issue was that you should get -- 19 Q. That's completely possible, by the way. 20 A. Sorry? 21 Q. That's completely possible. 22 A. Sorry. What the other issue was, where you would 23 actually get two transactions that were identical, apart 24 from their NUM, and therefore if you tried filtering 25 things out without looking at the NUM, you might be 72 1 filtering out things that weren't actually genuine 2 duplicates. 3 Q. I see. So is that why that is not mentioned in this 4 section? 5 A. Yes. So, therefore, what it's saying there: the 6 reliable way of identifying the duplicate is to use the 7 NUM for the filtering, rather than by looking at the 8 text of what was presented in the ARQ. 9 Q. Thank you. So, overall, would you agree this was 10 a serious problem? 11 A. Oh, yes. 12 Q. Can we turn to things that you were saying at this time, 13 as a separate issue -- that can come down -- as perhaps 14 a way of understanding what was operative upon you at 15 the time. FUJ00152888. 16 Look at page 2, please. Can you see that you have 17 been reading through a report; can you see that? 18 A. Yes, I can see that. 19 Q. Is that about Shoeburyness and Leigh-on-Sea? 20 A. I can't -- I thought this was to do with Porters Avenue 21 but I'm not -- I -- I don't recall this particularly. 22 All I can go by is the email that you showed me. 23 Q. If we just scroll down, and again, please -- thank you: 24 "I have received a copy of an accountant's report 25 and attach." 73 1 A. Yes. 2 Q. "... and a statement from Phil Budd ... 3 "Please could you read the report and decide how to 4 continue; the request from the accountant, Charles 5 McLachlan ... suggests a telephone call ..." 6 Does that help or not? 7 A. Not really. As I say, I do remember looking at 8 something to do with Porters Avenue and I do remember 9 that Phil Budd had something to do with it. So that's 10 why I'm assuming this is to do with that but I can't 11 remember the context and I've certainly no recollection 12 of what the report said and why I was concerned about 13 it. 14 Q. If we go up to page 2, then, where we were -- thank 15 you -- you say: 16 "Penny, 17 "I've now read through this report. I agree that 18 there probably needs to be some further investigation 19 here. However I'm also aware that this is probably 20 highly political. Therefore I'm not sure how best to 21 address this. We don't really want to be seen to be 22 undermining a [Post Office] prosecution! 23 "I think the report does raise a number of questions 24 without a detailed analysis of the various transactions 25 over the period described (it's 6 to 9 months) it is 74 1 difficult to progress. The cost of such an analysis is 2 likely to be high. We did some similar analysis over 3 a two-month period for a defence accountant about a year 4 ago and were paid for it (Pete Sewell set it up -- Anne 5 did most of the hard work and I presented it to the 6 accountant). I think we need some management guidance 7 on this." 8 You say that this is "highly political". 9 A. Basically, what I meant it's not a technical issue; this 10 is something that management need to think about. 11 Q. Then you would say this is a management issue? 12 A. Yes, so that's basically saying that -- repeating my 13 point. 14 Q. In what respect was it political? 15 A. It was to do in the way in which -- I'm not using 16 "political" in terms of party politics, I just saw 17 that -- when you -- 18 Q. It means -- 19 A. -- "commercial" might have been a better term to use. 20 It's to do with how Fujitsu interacts with Post Office. 21 Q. Okay, so it obviously doesn't mean high politics? 22 A. Oh, no, it's to do with the higher level communication 23 between the companies, which was not the sort of thing 24 I liked to get involved in, if I could help it. 25 Q. Why was it political in that sense? 75 1 A. Because I think we needed guidance as to how to approach 2 this. As I say, I can't remember any of the details of 3 what the details were in this particular case but 4 I think this was another thing that caused Penny to get 5 David Thomas involved because I think -- not David 6 Thomas, David Jones -- because that occurred about 7 a week after this and, certainly, he was asked about 8 Porters Avenue as well as West Byfleet, and that's why 9 I think this is to do with Porters Avenue. 10 Q. Political in the sense that we've been asked to give 11 evidence about the reliability of the system that we're 12 supplying to the Post Office? 13 A. I'm not sure what the issue was to do in this case. 14 Q. You continue, and I want to ask whether this gives us 15 an understanding of what you meant: 16 "We don't really want to be seen to be undermining 17 a [Post Office] prosecution!" 18 Firstly, did that represent your state of mind, that 19 Fujitsu should not be seen to be undermining Post Office 20 prosecutions? 21 A. Well, I'm not sure that that's -- I think I needed to 22 get some guidance on that but there was certainly 23 an example email I've been shown, as part of this 24 Inquiry, where I have actually said, in response to 25 a report that I'd been asked, that I agreed with the 76 1 defence expert and Post Office decided to ignore what 2 I'd said. 3 Q. Can we just focus on the question, please, Mr Jenkins? 4 A. Yeah. 5 Q. Was it your state of mind at the time that you believed 6 that Fujitsu should not be seen to be undermining a Post 7 Office prosecution? 8 A. Well, I was wanting guidance as to what exactly we 9 should be doing. 10 Q. Why did you need guidance on whether or not Fujitsu 11 should just present the facts, irrespective of whether 12 they supported or undermined a Post Office prosecution? 13 A. I just wanted to get some sort of guidance from senior 14 management as to any communication I had in this area 15 with Post Office and the guidance I got was to just tell 16 the truth, which is what I would like to have done 17 anyway. 18 Q. Did you need guidance from management -- 19 A. I wanted to make sure that I didn't undermine anything. 20 Q. Did you need guidance from management on whether you 21 should just tell the truth or not? 22 A. No, I didn't need guidance. I would just tell the truth 23 but I just wanted to make sure that I had -- I was being 24 covered by my management for any actions that I took. 25 Q. You needed top cover for telling the truth? 77 1 A. I don't know what to say to that. 2 Q. Why did you need management guidance on whether or not 3 Fujitsu should be seen to be undermining a Post Office 4 prosecution? 5 A. I just wanted to get some sort of feeling as to how to 6 approach the response to the report. 7 Q. When you subsequently came to give evidence in the Misra 8 trial, very limited questions were asked about your -- 9 the status in which you were giving evidence and the 10 duties of an expert witness, agreed? 11 A. Agreed. 12 Q. One of the only things that was asked was, in fact, by 13 the judge -- I'm not going to turn it up, I'm going to 14 read it. 15 A. No, I know what you -- 16 Q. You know the question. 17 A. Yes. 18 Q. "Is there any question of your evidence having been 19 influenced by the fact that you're a Fujitsu man?" 20 You said no. 21 A. And I stand by that. 22 Q. If that's the case, why were you asking for management 23 guidance on whether or not the evidence that should be 24 given in response to a defence expert report -- 25 A. I don't think we were talking about evidence at this 78 1 stage. It for just a case of comments on a report. 2 Q. Okay, so comments on a report -- 3 A. Yeah. 4 Q. -- a defence expert report. 5 A. Yeah. 6 Q. If your evidence wasn't influenced by the fact that you 7 were a Fujitsu man, why were you seeking management 8 guidance on what should be done if your answers might 9 undermine a Post Office prosecution? 10 A. It was a case of whether my answers should be sent -- my 11 comments on the report should be sent through or not, 12 I suppose. As I say, I just cannot remember what the 13 issue was that I was concerned about in this case. 14 Q. When you gave witness statements and then subsequently 15 were asked questions and answered them in court, was 16 a factor operative on your mind the fact that you, on 17 behalf of Fujitsu, should not be seen to undermine 18 a Post Office prosecution? 19 A. No. 20 Q. Because that's what the politics of the situation 21 demanded? 22 A. No, certainly when it came to court, then that didn't -- 23 that didn't come into it. I just wanted to understand 24 what I needed to do about my thoughts on this particular 25 report, which was not at that stage anywhere near 79 1 a court, as I understood it. 2 Q. So are you saying that there's a material difference 3 between providing comments on a report in an email 4 chain, on the one hand, where you wouldn't wish to be 5 seen to undermining a prosecution, and then ultimately 6 providing a witness statement, when such impure motives 7 would have evaporated? 8 A. Yeah, I was just looking for guidance as to what 9 I should be doing in this particular circumstance. 10 Q. Did you get guidance as to whether or not you should 11 simply set out the facts, irrespective of whether or not 12 they supported or undermined a Post Office prosecution? 13 A. Yeah, I think that would have affected the guidance 14 I got from David Jones the following week. 15 Q. Can we go forwards, please, to FUJ00156122 and look at 16 page 12, please. We're dealing with a different branch 17 here, "Alresford large debt outstanding", and there's 18 a reference to an email trail below. I'm not going to 19 turn to that. You say: 20 "This I another example of postmasters trying to get 21 away with 'Horizon has taken my money'. Dave Smith 22 seems to have put me forward as the expert to help on 23 this. 24 "How should I respond to his request?" 25 Does that reveal what your mindset was in February 80 1 2010 when you were conducting investigations into Seema 2 Misra's case and providing witness statements? 3 A. No, that is me very poorly trying to summarise what 4 was -- what I thought was being laid out in the email 5 trail below and I apologise for the wording that I used 6 there. But it was me trying to do a crude summarisation 7 of the email trail for management. I don't think I had 8 anything further to do with that particular case. 9 Q. A lot of people have said that they've used poor words 10 in their emails when they say that what is recorded does 11 not reflect truly what they believed. Did you believe 12 that there were many examples of postmasters trying to 13 get away with "Horizon has taken my money"? 14 A. I can't remember what I believed. 15 Q. Presumably -- 16 A. What I was trying to do was summarise the email trail 17 below. 18 Q. Yes, but you say it's another example of it. 19 Presumably, it does reflect your mind at the time that 20 postmasters were trying to get away with "Horizon has 21 taken my money"? 22 A. Well, I didn't believe that there were any problems with 23 Horizon that was causing Horizon to lose money. 24 Q. So you did believe that postmasters were blaming Horizon 25 by saying, "it's taken my money" falsely? 81 1 A. I just wasn't thinking things through properly. 2 Q. Hold on. I think you said that you believed that 3 Horizon was operating correctly, so this would 4 accurately reflect your belief. Postmasters were trying 5 to get away with it by blaming on it Horizon. 6 A. I don't think -- I think the issues are actually much 7 broader than that. I think the issues are as much to do 8 with the way that they were -- the Post Office was 9 treating the postmasters and the support and the 10 training they were getting, rather than actually 11 problems with Horizon, as such. 12 Q. Can we move on to POL00175839, and look -- if we scroll 13 down, please -- an email from you to Jarnail Singh on 14 1 March in the Seema Misra case: 15 "Thanks for the information, which I've now read 16 [it] through. 17 "... I can summarise ... 18 "Ms Misra ... initially identified that the sub post 19 office was short of cash by about £80-£90,000. 20 "She tried to cover this up while she repaid the 21 losses which she thought was due to staff theft. 22 "By the time of the audit in January 2008 she was 23 £73,000 short suggesting she had made good about £10,000 24 of the losses. 25 "When she went to court, she saw an article in 82 1 Computer Weekly indicating that Horizon was unreliable 2 and decided to jump on the bandwagon." 3 A. Again, that's totally inappropriate wording on my part 4 and for which I apologise. 5 Q. It seems that, rather than Mr Singh being the originator 6 of the "bandwagon" phrase, it appears to be you, doesn't 7 it? 8 A. No, I was reflecting that from conversations I'd had 9 with other people. It's not a term I would have used 10 myself. The only reason I would have used it here was 11 because I'd heard it being used by others. 12 Q. Who had -- 13 A. I can't remember now who it would have been, possibly 14 Mr Singh but I can't say definitively. 15 Q. So back to Mr Singh, perhaps? 16 A. Possibly. But, like I say, it's not a term I would -- 17 I would have used off the top of my head. I was clearly 18 reflecting it, that I'd heard it from someone else. 19 Q. Were you able to approach the task of conducting 20 investigations for the purposes of giving evidence in 21 court, compiling witness statements and then giving 22 evidence dispassionately, or -- 23 A. Yes. 24 Q. -- were you afflicted by this belief that these were 25 subpostmasters jumping on a bandwagon which blamed 83 1 Horizon? 2 A. No, I did the analysis conscientiously all the way 3 throughout. 4 Q. Did you see your role, as a Fujitsu man, to be to help 5 prove that Mrs Misra had jumped on the bandwagon and 6 that, to use the phrase from the other email, was trying 7 to get away with "Horizon has taken my money"? 8 A. As I say, I didn't think the problem was within Horizon; 9 I thought there were other reasons as to what had 10 happened and so my focus was looking at seeing whether 11 there were any problems with Horizon at that point, 12 rather than the broader focus of what the prosecutions 13 were about. 14 Q. Can we look lastly on this topic, just to see whether we 15 find another tell as to what your state of mind was by 16 looking at FUJ00153115. This is an email of 15 June 17 about the Misra case, saying that you had had a call 18 from Professor McLachlan. I'm going to skip over the 19 first three or four paragraphs. In the penultimate 20 paragraph, you say to Penny Thomas: 21 "The good news is that he [Professor McLachlan] is 22 looking for problems in people's use of Horizon and not 23 bugs in Horizon as such." 24 Why was it good news that the defence expert was 25 looking at people's use of Horizon and not looking at 84 1 bugs. 2 A. Because that's where I thought the genuine problems 3 were. 4 Q. Were you pleased that his attention was not on bugs? 5 A. Because I didn't think that there were relevant bugs at 6 the time. 7 Q. Were you worried that, if he did divert his attention to 8 the existence of bugs, he may have found some? 9 A. No. I was concerned that I would be wasting time 10 looking at other bugs that didn't actually apply, in the 11 same way as we'd already discussed with Callendar 12 Square. 13 Q. You thought that it was a waste of time looking for 14 other bugs? 15 A. Because I didn't think there were relevant bugs to be 16 looked at because, as far as I was aware, the bugs that 17 had occurred within Horizon had been fixed. 18 Q. That was all based, I think we established yesterday, on 19 assumption and chat? 20 A. Yes, but that was still my true belief at the time and 21 still is. 22 Q. I think that's because you reject some of the judge's 23 findings in the Horizon Issues trial? 24 A. Yes. 25 Q. Can we turn to the receipts and payments mismatch bug 85 1 and your treatment of it, by looking at POL00029084. If 2 we scroll down, please, there. I'm afraid it's cut off 3 on the left-hand side of the page. I think we can 4 establish from other evidence that this is 28 September 5 2010. You say to Mark Wright -- what role did Mark 6 Wright perform? 7 A. He was in the SSC. 8 Q. "The attached note summarises the problem ..." 9 The problem is in the title, "Receipts payments 10 mismatch issue". 11 A. Yes. 12 Q. "... and describes what we need to do to identify the 13 full scope of the problem and discuss with POL if/how we 14 fix the problem." 15 So you're attaching your lost discrepancies note -- 16 A. Yes. 17 Q. -- which we're going to look at in a moment? 18 A. Yes. 19 Q. You ask for feedback. You say: 20 "We probably need to formally raise this as 21 a problem with [the Post Office]. I'm not sure how this 22 is done ... We should then plan to do the initial 23 analysis [with Post Office] with a view as to scope and 24 then agree how to progress it." 25 Skipping a paragraph: 86 1 "Jon is easily able to reproduce the problem in 2 a development environment ... it is probably worth 3 starting on the data extraction to ascertain the full 4 scope of the issue ASAP since it has probably been 5 around since day one ..." 6 A. By that, I meant day 1 of Horizon Online. 7 Q. Yes, ie since January 2010? 8 A. Yes. 9 Q. "... and data more than 6 months old is being dropped 10 from BRSS ..." 11 Remind us, BRSS? 12 A. I can't remember exactly what the acronym stands for but 13 it was basically a copy of the live transaction database 14 that was used for support purposes. 15 Q. So this was a relatively time critical task because the 16 relevant data was being dropped off? 17 A. It was being dropped off there, it wasn't being lost 18 altogether. It could still be retrieved using ARQ data 19 or that, but it was available online through the BRSS. 20 Q. You say: 21 "... so the sooner we run the queries, the better." 22 A. Yes. 23 Q. If data more than six months old was being dropped off, 24 would that mean that there would be no way of knowing 25 how many branches were affected? 87 1 A. No. And, in fact, the way that we eventually decided 2 was the simplest way of identifying it was through 3 events that were being picked up from SYSMAN and from 4 the databases, and events were being held longer. 5 Q. So there was still a facility? 6 A. Yes. 7 Q. But, in any event, you thought this had probably been 8 around since January 2010? 9 A. Yes. 10 Q. Can we turn please to POL00028838, a document which 11 I think you'll be familiar with -- 12 A. Yes. 13 Q. -- and with which the Inquiry has been familiar for 14 a number of months now. There appears to have been some 15 meetings between Post Office and Fujitsu about the 16 receipts and payments mismatch bug and this is about one 17 of them, this note was compiled. You, in your witness 18 statements, try to date this meeting, don't you? 19 A. Yes, I can see there were a number of meetings during 20 the week and I'm not sure exactly which or whether this 21 was just a summary of the problem based on all of the 22 meetings. I don't think I saw this note until 2018. 23 Q. You were present at this meeting; can you see that? 24 A. I was present at some of the calls that we had during 25 that week, whether this is a note of a meeting or not, 88 1 is not 100 per cent clear. 2 Q. You, if we look at page 3, please, top paragraph: 3 "The Receipts and Payments mismatch will result in 4 an error code being generated which will allow Fujitsu 5 to isolate branches affected by this problem, although 6 this is not seen by the branches. We have asked Fujitsu 7 why it has taken so long to react to and escalate 8 an issue which began in May." 9 Did you tell them that the problem began in May? 10 A. I can't remember what exactly what I said in the 11 meeting. I doubt I would have scoped it in that way. 12 Q. Because you thought that the problem had been around 13 probably from day 1, ie January 2010. 14 A. Yes, there was a fairly obscure set of circumstances 15 which caused this problem, which is why it only occurred 16 about 60-odd times across what was by then a whole 17 estate of 12,000 branches. So it is quite possible that 18 the first occurrence of the problem was in May, in that 19 none of the branches had actually gone through that 20 particular set of interactions earlier. But, if they 21 had gone through that set of interactions before May, 22 then they could have hit the problem. As I say, I can't 23 remember now exactly when the first occurrence was but 24 that's maybe where May comes from, I don't know. 25 Q. It's said that they -- that's Fujitsu -- will provide 89 1 feedback on that issue in due course, ie "We are in 2 October now, why is it that we are being told about this 3 now?" 4 A. Yes, and I can't remember the circumstances as to why it 5 was only at this point that the problem had actually 6 been discovered. My first involvement with it, as far 7 as I can remember, was at the end of September, so about 8 a week before these calls with Post Office. 9 Q. The code fix is mentioned in the next paragraph. 10 A. Yes. 11 Q. Was it the case that, at this meeting, or as a result of 12 these communications, Post Office was told that Fujitsu 13 had data from which it was possible to ascertain all of 14 the branches that had been affected? 15 A. That is my understanding. We certainly did have that -- 16 we certainly worked out that information at some point. 17 Exactly how that fits in with this meeting or not, 18 I can't remember the exact chronology. 19 Q. Now, we've seen that you authored a document about 20 correcting accounts for loss discrepancies. If we can 21 look at that, please, it's part of this package of 22 material, it's at page 6. We'll see this is your 23 document, correct -- 24 A. Yes, I believe so. 25 Q. -- dated 29 September. 90 1 A. Ah, that's right. Most of the other versions that I've 2 seen have been dated 28 September, so whether it's -- 3 how that differs, I'm not sure. 4 Q. Okay. 5 A. Probably not materially. 6 Q. Under the heading "Introduction", you discuss some 7 PEAKs, 765 and 263, and also 864, which is a duplicate 8 of 263. You ask the question: 9 "Are these really duplicates? I'm a bit confused as 10 to which one to refer to. Can one be closed as 11 a duplicate of the other?" 12 Was that an issue that happened frequently, namely 13 multiple or duplicate PEAKs and some were closed in 14 order to keep only one open if possible? 15 A. I am not an expert on the process. You would need to 16 talk to the SSC people about that but my understanding 17 was that, if the same problem was raised for more than 18 one branch, then the approach was just to have one PEAK 19 that actually controlled the issue, and there was 20 a mechanism within that PEAK of listing all the branches 21 that were affected by that issue and then closing other 22 PEAKs that are duplicates, so there was just one PEAK 23 which could actually sort out the result of the problem. 24 Q. In the paragraph underneath, just two lines from the 25 bottom of it, you say: 91 1 "... PEAK [765] is a Master PEAK to record all 2 affected branches ..." 3 A. That is really what I was trying to say, yes. 4 Q. There should be a master -- 5 A. There should be a master PEAK, yes, because there is no 6 point in investigating the same problem more than once, 7 once it has been established that it is the same 8 problem. 9 Q. Was that something that was always done, ie when 10 a problem or issue or bug was reported, there was 11 a system in place to consolidate reports across Fujitsu 12 to collect together all reports relating to that problem 13 into a master PEAK? 14 A. That is what I understood was the way that things should 15 be working. I'm not sure that it always did work in 16 exactly that way but that was how the process was 17 supposed to work, as I understand it. This wasn't my 18 process. 19 Q. Over the page to page 7. If we just scroll down, 20 please. At the foot of the page, you say in the box 21 immediately at the bottom of the page: 22 "Mark Wright has produced a list of 16 occurrences 23 of event 903 in the last 30 days. This needs to be 24 extended." 25 What was event 903; can you remember? 92 1 A. It was the primary symptom of this particular problem. 2 I can't remember the details more than that, I'm afraid, 3 now. 4 Q. Mr Wright, had found 16 occurrences of that problem, or 5 symptom, in the last 30 days, and that needed to be 6 extended. Do you mean -- 7 A. So therefore I was saying you need to look back more 8 than 30 days ago, you need to look back at all 9 occurrences of this event going back to January 2010. 10 Q. Was that done? 11 A. Yes. 12 Q. How many were eventually detected? 13 A. I think at the end we found there were 64 occurrences in 14 62 branches but they hadn't necessarily all happened at 15 this point because some of them -- the problem continued 16 to occur until the code fix went in about two or three 17 weeks after this time. 18 Q. In the report overall there was an acknowledgement that 19 subpostmasters may be unaware of the problem because 20 there wasn't essentially a prompt to them to make them 21 aware of it. Is that right? 22 A. There wasn't an explicit prompt. If they had looked in 23 detail at the branch trading statement, they would have 24 seen that there was a problem there but there wasn't 25 a message that actually popped up and said, "Yes, you've 93 1 got a problem". 2 Q. So they would need to examine, very carefully, a branch 3 trading statement -- to -- 4 A. There was one figure in there which should always be 5 zero and, in this particular case, it wasn't. 6 Q. If we go over to page 8, please. At the bottom of the 7 page, please, under "Communication with the Post 8 Office", you say: 9 "Once we have the information [above] which will 10 enable us to identify the full scope ... we need to 11 communicate this to the Post Office Limited through the 12 problem management mechanisms. We will then need to get 13 Post Office to agree if and how we should be correcting 14 the data." 15 Why did you think there was a possibility that you 16 might not correct the data? 17 A. We didn't correct the data in the branches. What we 18 actually did was we corrected the data in the Post 19 Office's back-end systems. 20 Q. "Post Office should also be able to check up on POLSAP 21 to confirm these discrepancies are still visible even 22 though they have been lost in the branch. 23 "It should be noted that as discrepancies are 24 normally losses, then a lost discrepancy would [need to] 25 work in the branch's favour so there is no incentive for 94 1 the branch to report the problem. Also, if we do amend 2 the data to re-introduce the discrepancy, this will need 3 to be carefully communicated to the branches to avoid 4 questions about the system integrity." 5 A. We never did amend the data in the -- 6 Q. Yes, just before you get that answer in too quickly, why 7 would you, if you did amend the data, need carefully to 8 communicate that to the branches, in order to avoid 9 questions about system integrity? 10 A. Well, I can see that this was an important problem. 11 I didn't want to be amending the data in the branches, 12 and we didn't amend the data in the branches -- 13 Q. No, no, that's a different issue. 14 A. If it was -- 15 Q. It -- 16 A. If it was badly communicated, it could show that there 17 were serious problems within the ongoing operation of 18 Horizon Online, and we clearly didn't want to do that, 19 so it wasn't a case of covering up; we just didn't want 20 to have bad publicity for what was actually a very 21 isolated problem. 22 Q. That's not what this says. You wanted to avoid 23 questions about system integrity, didn't you? 24 A. That was behind what I was trying to say at the time. 25 Q. Well, on this document, that's all that was behind what 95 1 you wanted to say. You wanted to avoid questions being 2 raised about the system's integrity, didn't you? 3 A. I didn't think -- I didn't think there was anything 4 fundamentally wrong and I wanted to ensure that things 5 were actually communicated accurately. 6 Q. This isn't about the accurate or inaccurate 7 communication of information to branches, it's about 8 avoiding questions concerning the system's integrity. 9 That's your motive, isn't it? 10 A. No, my motive was to make sure the communications were 11 done accurately. 12 Q. Why didn't you say, "If we do amend the data, we must 13 communicate that to the branches accurately"? 14 A. Perhaps that's what I should have said at the time. 15 But, as I say, I wasn't exploring every word that 16 I wrote in an internal document expecting it to be 17 picked apart in the way that it's being picked apart 18 now. 19 Q. Apologies for picking it apart, I'm just trying to work 20 out what was operative on your mind at the time. 21 A. Operative on my mind was that we didn't want to blow 22 this out of proportion. This was a relatively small 23 problem that was affecting a very small number of 24 branches and we didn't want to actually undermine the 25 overall system, which was working perfectly correctly in 96 1 many, many branches, and this was a problem that was 2 going to get fixed, and it was easily scoped and 3 identifiable. 4 Q. But you didn't know any of that at this time. You told 5 us that there had only been a backward look for 30 days. 6 You thought this went back to January, rather than May. 7 You didn't know the extent of the problem at all -- 8 A. But I knew that -- 9 Q. -- a limited problem affecting a small number of 10 branches that was easily fixed -- when you were writing 11 this sentence, did you? 12 A. I was aware that -- I had a fairly good idea what the 13 sort of scope was going to be. At this point, we'd 14 identified 16 branches, I could see that we would 15 identify a few further branches and, as I say, the 16 eventual figure was about 60-something. 17 Q. I'm going to continue the picking apart, if I may. The 18 reason that you used these words was because that's what 19 you were thinking at the time, isn't it? It's not 20 a poor choice of words, it's not a choice of words that 21 was in error, it's not a choice of words that reflects 22 something that you didn't think. You wanted to avoid 23 questions about the Horizon system's integrity, didn't 24 you? 25 A. I don't think that's true but ... 97 1 MR BEER: Sir, it's just gone 12.20. 2 SIR WYN WILLIAMS: Yes. 3 MR BEER: Can we break until 12.35, please? 4 SIR WYN WILLIAMS: Yes. 5 MR BEER: Thank you. 6 (12.22 pm) 7 (A short break) 8 (12.35 pm) 9 MR BEER: Good afternoon, Mr Jenkins, can we continue your 10 evidence by finishing off the receipts and payments 11 mismatch issue, and go back to POL00028838, please. Can 12 we look, please, at page 2. 13 Can you see, in respect of this meeting, in relation 14 to which you're recorded as an attendee, under the 15 heading "Impact" at the bottom -- if that can be blown 16 up, please. Thank you. 17 "Impact 18 "The branch appeared to have balanced, whereas in 19 fact they could have had a loss or gain. 20 "Our accounting systems will be out of sync with 21 what is recorded at the branch. 22 "If widely known, could cause a loss of confidence 23 in the Horizon system by branches." 24 Just stopping there, at the third bullet point: 25 "If widely known it [the receipts and payments 98 1 mismatch bug] could cause a loss of confidence in the 2 Horizon system by branches." 3 You were at this meeting; did you say that? 4 A. I don't know, is the simple answer. I don't think so 5 but I don't know. 6 Q. Do you recall who did say that? 7 A. I have no recollection of -- whether this was one 8 meeting or a consolidated report of all the meetings, 9 have no idea. I remember being on conference calls 10 about the issue but what was said in detail on the 11 conference calls I've no recollection, I'm afraid. 12 SIR WYN WILLIAMS: Was there any face-to-face meeting, as 13 far as you can remember or were they all conference 14 calls? 15 A. There were face-to-face meetings internally within 16 Fujitsu but, in terms of communication with Post Office, 17 I think it was all conference calls. 18 SIR WYN WILLIAMS: So this list of people we see at the 19 start of this document, which is both Post Office and 20 Fujitsu, was there ever an occasion when you were all in 21 the same room together discussing it? 22 A. I don't believe so. 23 SIR WYN WILLIAMS: Right, fine. 24 MR BEER: Was there ever an occasion when you were all on 25 the same call? 99 1 A. I can't remember exact -- as I say, I thought it was 2 just one call. I've since seen suggestions that there 3 were two or three calls and I just can't remember the 4 details, I'm afraid. 5 Q. Was the outcome of the calls an agreement that, if the 6 problem was widely known, it could cause a loss of 7 confidence in the Horizon system by branches, as 8 recorded here? 9 A. I don't know. My role in these calls was explaining 10 what the issue was technically, not in terms of -- and 11 understanding what it was that was going to be decided 12 in terms of how we fixed it moving forward. That was my 13 main focus. 14 Q. If a view was expressed in the call or calls, whether by 15 Fujitsu, including you, or by the Post Office, that if 16 the receipts and payments mismatch bug was widely known, 17 it could cause loose of confidence by branches in the 18 Horizon system, did that fact, did that view, as 19 expressed, affect your willingness to mention it in 20 court in the following weeks when you gave evidence? 21 A. No, it never occurred to me that it was relevant 22 because, as I said earlier, this was an issue with 23 Horizon Online, and where I was giving evidence was to 24 do with Legacy Horizon. 25 Q. The next bullet point, therefore, becomes relevant: 100 1 "Potential impact upon ongoing legal cases where 2 branches are disputing the integrity of Horizon data." 3 By this time, I think only two branches had been 4 identified in respect of which there were ongoing court 5 cases, correct? 6 A. I've no idea. 7 Q. We looked at the report earlier, didn't we? One of them 8 was -- 9 A. That was in a different context. That was when we were 10 looking back at ARQ data back in June. 11 Q. By this time, had any court cases been identified? 12 A. I don't know. I don't recall any mention of legal cases 13 as part of this discussion. That wasn't where I was 14 focused on. What I was concerned with at this point was 15 the technical aspects of this issue. I have no 16 recollection of discussion of legal implications of 17 this. 18 Q. What, to your knowledge, ongoing legal cases were there? 19 A. I wasn't aware of any legal cases involved with Horizon 20 Online by that time. 21 Q. No, it doesn't say Horizon Online, does it? It simply 22 says, "ongoing legal cases". You knew about one, didn't 23 you, a big one: Seema Misra's case? 24 A. Yes, but I didn't see it as being relevant to that and 25 I would not have taken it as being relevant in -- that 101 1 that statement was referring to that at all. 2 Q. What ongoing legal cases were there concerning Horizon 3 Online? 4 A. I don't know. I wasn't aware of any. 5 Q. Were there, were there any at all? 6 A. I don't know. I don't understand the context of that. 7 I mean, I understand now what the context of that is 8 but, at the time, I wouldn't have known what that was 9 about and I'm not sure if I'd even have noticed if 10 anyone said anything along those lines. 11 Q. Did you say that: that the bug has a potential impact on 12 ongoing legal cases? 13 A. No. 14 Q. Who said that? 15 A. I've no idea. As I say, I don't even remember hearing 16 that being said. It's clearly recorded there in the 17 minutes. 18 Q. You agree that it's a record in this document of the 19 receipts and payments mismatch bug, having a potential 20 impact upon ongoing cases. Given the start date of 21 Horizon Online, you would have known at this time that 22 the likelihood of there being any ongoing legal cases 23 involving Horizon Online was minimal or even zero? 24 A. I've no recollection of there being a discussion about 25 legal cases as part of the -- these exchanges. I've 102 1 subsequently seen, as part of the work with the Inquiry, 2 that there was discussion along those lines but, at the 3 time, I am not sure that I'd have even noticed that sort 4 of thing being mentioned. I was concerned about the 5 technical aspects of this problem. 6 Q. So if it was discussed at a call or calls that the bug 7 had an impact or potential impact on ongoing legal 8 cases, that wouldn't have affected your willingness or 9 otherwise to mention it in the Seema Misra case? 10 A. No, because, so far as I was concerned, it was totally 11 irrelevant. Now, if someone had then come and said 12 I should mention it, I would probably have argued 13 against it. But if they said "No, you've got to do it 14 for legal reasons", then, obviously, I would have done. 15 Q. We know, in fairness to you, that you had disclosed it 16 in your note of 28/29 September 2010, and we know that 17 that was forwarded to Post Office Legal, Jarnail Singh? 18 A. Yes, I have learnt that as part of the work I've done 19 for the Inquiry, yes. 20 Q. The last bullet point: 21 "It could provide branches with ammunition to blame 22 Horizon for future discrepancies." 23 Do you recall who said that? 24 A. No, I don't, I'm afraid. 25 Q. Whether that was somebody from Post Office or Fujitsu? 103 1 A. I've no idea I'm afraid, sorry. 2 Q. Do you agree these three bullet points taken together 3 seem to reflect a concern about the disclosure of the 4 receipts and payments mismatch bug, in that it would 5 have undesirable consequences? 6 A. I can understand that now but, as I say, I don't think 7 I was conscious of that sort of discussion as part of 8 the meeting at the time. 9 Q. Can we go forward, please, to page 3. We can see the 10 solutions, and I think you'll be familiar with these? 11 A. Yes. 12 Q. If we scroll down -- thank you -- which one did you 13 support? 14 A. Where it was sorted out in the back end. So I think 15 that's Solution Two, isn't it -- 16 Q. Two. 17 A. -- where it gets journalled in POLSAP. And that was the 18 one, I believe, that was eventually done. 19 Q. The first and the third solution are said to have moral 20 implications, the first because it would involve 21 changing branch data without informing the branch and 22 the third, moral implications to the integrity of the 23 business; can you see that? 24 A. Yes, I don't remember the term "moral implications" 25 being used there but, certainly, it was -- I was always 104 1 favouring the second option, doing things in the back 2 end. I felt that was the right way of doing things and, 3 as I say, that was what was eventually done. 4 Q. The first solution was said to have significant data 5 integrity concerns and could lead to questions of 6 tampering with the branch system and could generate 7 questions about or around how the discrepancy was 8 caused. Is that because this would involve Fujitsu 9 manually writing values into branch accounts? 10 A. That is how that would have been done and I was against 11 doing that. 12 Q. Why were you against doing it? 13 A. Because I think it was the wrong thing to do. It was 14 much better to actually sort things out through 15 a business process that was there in the back end of 16 journalising the entries in POLSAP. 17 Q. That discussion revealed, of course, that the facility 18 was available to Fujitsu, a form of remote access? 19 A. I wouldn't call that a form of remote access, as such. 20 I think I -- because what we would have had to do is we 21 would have had to develop a specific bit of code to 22 actually make those sort of changes to those affected 23 branches and so it wouldn't have been using any of the 24 regular remote access type facilities that we had. So 25 it would have been a special bit of code that would have 105 1 been developed and tested specifically for that purpose. 2 But, again, that wasn't the way that we went. 3 Q. Was consensus reached on that in the meeting? 4 A. I can't remember the actual discussion but it was 5 certainly Solution Two that was done, going forward, in 6 terms of actually sorting things out at the back end. 7 I can't remember the discussions behind it and how long 8 those discussions were. 9 Q. Thank you. Can we move on, please, to the evidence that 10 you gave at trial, by looking at POL00029406. Can you 11 see this 14 October 2010 at the Crown Court at 12 Guildford. If we scroll down, it's a transcript of the 13 proceedings -- 14 A. Yes. 15 Q. -- and it's a transcript of your evidence. 16 If we go forward to page 123, please, and just go 17 down at H. At the foot of the page, thank you. 18 Re-examination, so that's questions at the end of your 19 evidence session, being asked by the prosecution 20 barrister, Mr Tatford. He says: 21 "Are there problems that the subpostmaster at the 22 Post Office is going to be unaware of?" 23 Can you see that? 24 A. Yes. 25 Q. You say: 106 1 "Most problems manifest themselves so they can be 2 visible. So, for example, if there was some problem 3 with balancing and so on, then that -- that I would 4 expect to be investigated to see whether there was 5 an underlying problem as a result of it." 6 The question that you were asked, "Are there 7 problems that a subpostmaster at the Post Office is 8 going to be unaware of?", you say, "Most problems 9 manifest themselves so they can be visible". 10 A. Yes. 11 Q. A month earlier, you'd been dealing with a problem that 12 didn't manifest itself in an obvious way to the 13 subpostmaster. 14 A. It did manifest itself, it was visible. If you actually 15 read thorough the BTS, there was a non-zero entry in the 16 BTS where there should be a zero entry. 17 Q. I think you had said already that that's if you examine 18 them carefully -- 19 A. Yes. 20 Q. -- or a similar phrase? 21 A. Yes. 22 Q. Was your reason for not mentioning that here because 23 that was to do with Horizon Online and you thought you 24 were answering questions only about Legacy Horizon? 25 A. I was certainly thinking I was only talking about Legacy 107 1 Horizon. 2 Q. Does that answer apply across the transcript -- 3 A. Yes. 4 Q. -- that, even if questions are asked of you in a broad 5 way, like that one was, which wasn't specifically about 6 Legacy Horizon, you thought you were talking and only 7 talking about Legacy Horizon? 8 A. Yes. 9 Q. Can I turn to the extent to which you were asked to 10 consider wider issues when you were providing your 11 witness statements and giving evidence, and just start 12 by looking at your witness statement, your third witness 13 statement, at page 125. This is just what you say at 14 the end of paragraph 369, the last four lines. You're 15 dealing here with the event timeout locking issues and 16 you say, in the last four lines: 17 "Nor have I seen any emails that suggest my raising 18 these event timeout/locking issues prompted anyone 19 within [Post Office] or Fujitsu to ask me to provide 20 a witness statement about any other past problems that 21 had affected Horizon." 22 A. Yes. 23 Q. By that, are you saying that you don't more generally 24 recall being asked to provide a witness statement about 25 any past problems that had affected Horizon? 108 1 A. Yes, I think that's what I was saying. What I thought 2 I was being asked to look at and what I think I was 3 looking at was specific problems at the specific branch, 4 and I found no evidence of any specific problems at the 5 branch. I had raised with people within Post Office the 6 fact that I was aware of problems that didn't affect 7 that branch, and no one said, "Oh, in that case, you 8 need to talk about it and put it in your witness 9 statement". 10 Q. You say, therefore, that you don't think you were being 11 asked to provide witness statements or a witness 12 statement about any past problems that had affected 13 Horizon. Can we look back, please, at FUJ00122794. 14 FUJ00122794, and page 2, please -- and scroll down. 15 We looked at this email first thing this morning. 16 Jarnail Singh, and we see that you're copied, and 17 Mr Singh has said that points 2 to 4 haven't been 18 answered, and 3: 19 "When Gareth completes his statement could he also 20 mention whether there are any known problems with the 21 Horizon system that Fujitsu are aware of." 22 So you had been asked that very question. 23 A. And I'd responded to it at the time saying that until 24 I looked at the logs, I wasn't able to make a clear 25 statement, and then nothing further was actually done 109 1 about that. So I thought that that had been addressed. 2 Q. Well, let's break that down. In your witness statement, 3 you said, "Nobody asked me to provide a statement about 4 any other past problems that had affected Horizon". 5 This is a request to do exactly that, isn't it? 6 A. I'd not considered it that way when I was putting 7 together the witness statement. 8 Q. No, but looking at it now, it's a request to do exactly 9 that, isn't it? 10 A. I'd not seen it that way, but possibly. 11 Q. So you'd been asked that question, the broad question of 12 known problems with Horizon, through the lawyer's route, 13 I'm going to call it, seen to you. Yes? 14 A. Yes. 15 Q. Can we look, please, at FUJ00152902, page 2, please. 16 Can you see an email, Jon Longman to Penny Thomas -- 17 yes -- 18 A. Yes. 19 Q. -- on 1 February. If we just scroll up, just the bottom 20 of the page there, Penny Thomas to you: 21 "Please see ... below; this is now extremely 22 urgent." 23 So you get forwarded this chain. 24 Go back to where we were, top of page 2, so Longman 25 to Thomas. I'm going to call this the investigator's 110 1 route, rather than the lawyer's route: 2 "At a pre-court hearing today the judge has ordered 3 that all defence requests [be answered]." 4 Then: 5 "Gareth's statement needs to cover the following 6 four points." 7 Then do you see (3): 8 "When Gareth completes his statement he should also 9 mention whether there are any known problems with 10 Horizon system that Fujitsu are aware of. If none could 11 this be clarified in the statement." 12 So, in fact, far from what you say in your witness 13 statement that "Nobody had asked me to provide 14 a statement about any other past problems that had 15 affected Horizon", you were asked that very question 16 twice, weren't you? 17 A. Well, I didn't respond to that either way. 18 Q. Well, let's just agree: you were asked the question 19 twice, weren't you, once through the lawyer's route and 20 once through the investigator's route? 21 A. I'd not considered it in that way. 22 Q. I know you say that you hadn't but the fact of the 23 matter is, rather than you not being asked a question 24 about broader issues or problems with Horizon at all, 25 you'd in fact been asked it twice: once by a lawyer and 111 1 once through an investigator's email, hadn't you? 2 A. It looks like it. 3 Q. Can we scroll up to what you replied: 4 "I've finally managed to go through the witness 5 statements [et cetera]. 6 "I don't know anything [presently] about Falkirk." 7 Then: 8 "I'm not aware of issues in Horizon other than the 9 event timeouts. Not sure how to cover that in the 10 witness statement." 11 A. Which is basically the same thing as I said to David 12 Jones later in the week. 13 Q. Well, firstly, do you agree that answer there in that 14 third paragraph is your response to the question, or the 15 request, to cover whether there were any known problems 16 with the Horizon system that Fujitsu are aware of? 17 A. I think it probably was, yes. 18 Q. So that's a direct response to the broad question. You 19 were aware of issues in Horizon, other than the event 20 timeouts, weren't you? 21 A. Not ones that were still outstanding. 22 Q. Is that what you say? 23 A. I had assumed that as part of the context but I accept 24 that I didn't spell it out. 25 Q. The context is framed by the question that you were 112 1 asked, "Are there any known problems with the Horizon 2 system that Fujitsu are aware of?" That's talking about 3 the operation of Legacy Horizon at the very least, isn't 4 it? 5 A. I wasn't understanding the question that way but I can 6 understand that that may have been what was meant. 7 Q. You say that this answer is supposed to read, "I'm not 8 aware of any outstanding issues in Horizon, other than 9 the event timeouts which have not been resolved". Why 10 would you not say that, if that's the meaning that you 11 now attribute to those words? 12 A. I don't know, is the simple answer. 13 Q. That statement there is not true, is it, "I am not aware 14 of issues in Horizon other than the event timeouts"? 15 A. I was thinking of outstanding issues but I can accept 16 that I have not qualified it correctly. 17 Q. So you were asked twice, once by Jon Longman and once by 18 Jarnail Singh, explicitly to mention any known problems 19 with the Horizon system. Why on both occasions have you 20 applied a restrictive approach, rather than simply 21 mentioning all of the problems that you were aware of? 22 A. That was not how I'd understood the question but 23 I accept now that I'd misunderstood the question. 24 Q. Or was it an unwillingness to reveal known problems with 25 Horizon system? 113 1 A. I think it was a case of me misunderstanding the 2 question. 3 Q. Can I turn to the topic of your knowledge of remote 4 access before Seema Misra's trial by looking at 5 FUJ00083737. 6 This is, I think, your fifth witness statement, 7 dated 8 October 2010. If we go forwards, please, to 8 page 8, and if we scroll down, please. Thank you. Just 9 back up again, so we get the whole of that paragraph. 10 Thank you. You say: 11 "... in section 1.2.3 [that's of Professor 12 McLachlan's report] there is the hypothesis that 13 'External systems across the wider Post Office operating 14 environment provide incorrect externally entered 15 information to the Horizon accounts through system or 16 operator error outside Horizon'." 17 You say: 18 "I was not quite clear what Professor McLachlan was 19 referring to here. In the updated version of the report 20 [he] has clarified this by adding 'For example, 21 incorrect transaction corrections are submitted from the 22 central systems for acceptance by the subpostmaster'." 23 You continue: 24 "However in my view this is not really relevant 25 since any transaction that is recorded on Horizon must 114 1 be authorised by a user of the Horizon system who is 2 taking responsibility for the impact that such 3 a transaction has on the branch's accounts." 4 Then this: 5 "There are no cases where external systems can 6 manipulate the branch's account without the users in 7 branch being aware of what is happening and authorising 8 the transactions." 9 That's not correct, is it? 10 A. I was referring to the transaction corrections and 11 external systems. 12 Q. You don't say that, do you? 13 A. I do say that external systems cannot manipulate the 14 branch accounts, and the context of the question that 15 I was referring to was to do with transaction 16 corrections, and that's how I was understanding the 17 scope of it. 18 Q. The month previously, you had been present on a call or 19 calls where, when discussing the receipts and payments 20 mismatch bug, a solution was alteration of figures at 21 accounts, at branch level, the local branch accounts. 22 That was rejected because it would be done without the 23 subpostmaster knowing about it and it therefore lacked 24 integrity. 25 A. I don't think that was the only reason it was rejected. 115 1 I think it was rejected because it was a complicated way 2 of doing things and it was much simpler to do things at 3 the back end system. 4 Q. But didn't that make it clear that external systems 5 could manipulate the branch's accounts without the user 6 being aware? 7 A. I wouldn't have called that external systems and 8 manipulating accounts. That would have been a specific 9 code change made to actually change the accounts, and it 10 was not something that I would have been in favour of, 11 and that didn't happen. 12 Q. So wouldn't a full answer here have been, "There are 13 facilities for remote access that I've known about since 14 [I think you said yesterday] 2000 which are, in my view, 15 theoretically available to alter branches' accounts. My 16 understanding is they are used infrequently but that 17 facility is there"? 18 A. I'd not understood that as being part of the question. 19 I thought we were talking about business as usual -- 20 I think, in my fourth witness statement, I talk about 21 business remote access, and that was the context that 22 I was taking this discussion to be in. 23 Q. Did you, in any of your witness statements, disclose the 24 other forms of remote access that were possible? 25 A. I didn't think they were relevant. 116 1 Q. Why did you think they were not relevant? 2 A. Because I did not think that those had taken place in 3 West Byfleet. 4 Q. Therefore, there was no necessity to tell the defence or 5 the court about the facility for Fujitsu remotely to 6 alter branch accounts without a subpostmaster's 7 knowledge? 8 A. That was where I was coming from. 9 MR BEER: Thank you. 10 Sir, I wonder if we could take the break there. 11 We're about to change topic. 12 SIR WYN WILLIAMS: Yes. 13 MR BEER: Until 1.55, please. Thank you very much. 14 (1.07 pm) 15 (The Short Adjournment) 16 (1.55 pm) 17 MR BEER: Thank you, sir. 18 Mr Jenkins, good afternoon. 19 A. Good afternoon. 20 Q. Can we turn to the issue of hardware failures at the 21 West Byfleet branch. Can we start by looking at 22 FUJ00083737. This is your witness statement of 23 8 October 2010. Can we turn to page 4, please. At the 24 top of the page you say: 25 "I understand there is a suggestion that the 117 1 equipment in the branch might be faulty. I am not aware 2 of any fundamental issues though this being covered by 3 Mr Dunks. Specifically, in his witness statement he 4 states that 'All the calls are of a routine nature and 5 do not fall outside the normal working parameters of the 6 system or would affect the working order of the 7 counters'." 8 By saying that, are you saying that you agreed with 9 him that all of the calls were of a routine nature, that 10 they didn't fall outside the normal working parameters 11 of the system or they wouldn't affect the working order 12 of the counters? 13 A. No, just quoting what he had said. 14 Q. What was the purpose of quoting him? 15 A. I was really referring that to his level of expertise 16 rather than mine. 17 Q. What expertise did he have in the malfunctioning of 18 equipment within branches? 19 A. Well, he had been examining the Helpdesk calls. 20 Q. What -- 21 A. And that's really what I -- that's really what -- he -- 22 I believed that he had an understanding of what were 23 routine calls and, therefore, I was believing -- if he 24 said that they were of a routine nature, I was believing 25 what he said. 118 1 Q. What was Mr Dunks' job? 2 A. He was part of the Security Team. 3 Q. Had he got any expertise in the functioning of equipment 4 within branches, to your knowledge? 5 A. I thought that he was experienced in analysing Helpdesk 6 calls. 7 Q. Experienced in -- what, good at reading? 8 A. Yes. 9 Q. Can we turn to some of the Helpdesk calls then, please. 10 POL00061793. Do you recognise these as being the logs 11 from the Helpdesk for the branch? 12 A. I recognise them as being that now, yes. 13 Q. Did you examine them at the time? 14 A. No, I did not. 15 Q. You've said in your witness statement, "I am not aware 16 of any fundamental issues", ie with equipment in the 17 branch. How were you able to say that you were not 18 aware of any fundamental issues? 19 A. Because I'd read Mr Dunks' witness statement. 20 Q. Did anyone suggest that that was an appropriate thing to 21 do, in answer to a question or a suggestion that there 22 was faulty equipment in the branch, you would read 23 somebody else's witness statement who said they'd read 24 the logs, and therefore you form a qualitative 25 conclusion that there weren't fundamental issues? 119 1 A. No one has suggested to me that that was inappropriate. 2 Q. You just did that yourself? 3 A. Yes. 4 Q. Did it occur to you that, in order to give evidence to 5 a court as to whether or not there were fundamental 6 issues with equipment in the branch, you should actually 7 look to see whether there was evidence of fundamental 8 issues with equipment in the branch? 9 A. That was outside my area of expertise. 10 Q. Why didn't you answer the question, "I understand 11 there's an issue with or a suggestion that the equipment 12 in the branch is faulty, or might be faulty. That's 13 outside my area of expertise", full stop? 14 A. I was just referring to Mr Dunks, who I thought had done 15 that sort of analysis. 16 Q. Yes, but you allied with it, didn't you, by mentioning 17 it, by quoting from it and saying, "I'm not aware of any 18 fundamental issues"? 19 A. I'd not seen it that way at the time. 20 Q. Do you see it that way now? 21 A. Potentially. 22 Q. If we look at some of these Helpdesk records, then 23 please. Can we look at page 37, please, at the bottom. 24 If we just go to the bottom of the page, thank you, 25 I think this is a record of a report on 2 May. Do you 120 1 see "Opened and closed", 2 May on the middle of the 2 page, that's it. Thank you. 3 A. Yes. 4 Q. The "Problem" text is: 5 "Critical Alert Received ... at [then there's a date 6 and a time]. Event description: A fatal error has 7 occurred. A corrupt storage unit was detected on volume 8 %1 with LPN %2. Unit type %3 ... The message store will 9 be shut down abnormally." 10 Can you see that? 11 A. I can. 12 Q. Can you help us as to what that is? 13 A. That is sounding like a hardware problem on the disk at 14 that time but that was out of hours but I don't think 15 that would have caused any problem to the branch 16 accounts because that would mean that Riposte just 17 wouldn't be able to function. So the counter would 18 effectively be dead, until either the message store was 19 deleted and recreated or the base unit swapped. 20 Q. So a corrupt storage unit at the branch being detected 21 and the message store was going to be shut down 22 abnormally? 23 A. Yes. 24 Q. If we go to the top of page 40, please, do you see the 25 entry at the very top: 121 1 "BASE UNIT SWAPS 'Warning!! Do not change the 2 mirror disk and PC together -- if you are in doubt 3 contact SMC or HIT'. 'Please can engineer print balance 4 snapshot and contact SMC to synchronise message store 5 before swapping out the base unit'." 6 Can you help us, what does that mean? 7 A. The simple answer is I don't know. That looks like some 8 sort of instruction to the engineers as to what they 9 need to do. 10 Q. To print a balance snapshot and then synchronise the 11 message store; why would that be necessary? 12 A. I think it is to do with the case of where there's 13 a single-counter branch, where there is only one counter 14 with what's called a mirror disk, removable hardware 15 drive, and what's supposed to happen in those cases is 16 that the removable drive is supposed to put into the new 17 counter -- 18 Q. This was a multiple counter branch though? 19 A. So, therefore, I don't think that's relevant to a branch 20 like West Byfleet, which had more than one counter. 21 Q. Do you know why it would be recorded on a Helpdesk call 22 log? 23 A. No idea. 24 Q. Is what we've read so far evidence that nothing happened 25 that fell outside the normal working parameters of the 122 1 system? 2 A. I believe so. 3 Q. Can we look at page 40 -- 4 A. Again, I'm not aware of what was normal. I was relying 5 on Mr Dunks knowing what was normal. 6 Q. Did you know how he knew or not knew what was normal? 7 A. No, but the fact that he was saying it was normal, 8 I took his word for it. 9 Q. Thank you. If we scroll down the page, please, and look 10 for the entry at, thank you. It looks, at 13.03, as if 11 an engineer arrived, and then the third entry of 13.03, 12 the one above that, please, "Engineer swapped ADSL 13 cable"; can you see that? 14 A. Yes. 15 Q. Then if we look, please, at page 41, the same day, can 16 we see a further record of a call: 17 "Critical NT event [this is under 'Problem Text'] 18 Riposte Error ... The ... server is unavailable." 19 Can you see that? 20 A. Yes, I can see that. 21 Q. Then, under the call activity log, "New call taken by 22 David Sweetlove", and then the substance of the message 23 again. So it looks like, continuing on the same day, 24 there is still a problem, doesn't there? 25 A. Yes, whether it's the same problem or a different one. 123 1 The other one was to do with a comms cable. This is to 2 do with the a problem with the background process. 3 Q. Are they unrelated? 4 A. I don't think so but I don't know. 5 Q. Again, would you be able to say whether this is evidence 6 of something falling within or outside the normal 7 working parameters of the system? 8 A. From my experience, I wouldn't be able to tell one way 9 or the other. 10 Q. It seems, over the next few pages -- I'm not going to 11 read them all in the interests of time -- but the base 12 unit was replaced because Mrs Misra's screen had frozen 13 on an error message which said, "Please wait whilst the 14 desktop connects to Riposte" and successive reboots had 15 failed to rectify the problem. Would that be the system 16 working within or outside its normal parameters? 17 A. Certainly, base units did fail occasionally and 18 therefore I would see that as being a normal sort of 19 occurrence. 20 Q. Would you want to check data that happened at about that 21 time to see whether it was affect by the failure of the 22 base unit? 23 A. I didn't have data for that period. I was focusing my 24 attention on the time that I'd been asked to look at, 25 which was December 2006 to December 2007, and this was 124 1 before that time. 2 Q. Do you know why you didn't have data for the period of 3 the indictment, which was about two and a half years, 4 for the theft count? 5 A. The Post Office had -- I know now that Post Office had 6 decided that on the grounds of costs but, at the time 7 then, I just knew that I'd been asked to look at data 8 for that period of 13 months. 9 Q. One thing you'd want to do if you were examining these 10 records is correlate the problems that are shown in the 11 Helpdesk to the transaction data to see whether there 12 was any connection between the two? 13 A. I realise now that that is something that I should have 14 done. At the time, I was not aware that I needed to 15 look at the Helpdesk calls and correlate them against 16 the message store. 17 Q. In your witness statement, you had said, we saw, 18 "I understand there is a suggestion that equipment in 19 the branch might be faulty, I'm not aware of any 20 fundamental issues, it's being covered by Mr Dunks". 21 Did I not occur to you, "I need, on the one hand, the 22 Helpdesk data or the logs and, on the other, the 23 underlying data to correlate them and I've got neither"? 24 A. I realise that now but I didn't at the time. 25 Q. Are you aware of any discussion within Fujitsu as to 125 1 whether that would have been the appropriate response to 2 the suggestion of faulty hardware on a continuing basis 3 within the West Byfleet branch: you need to do both of 4 those two things? 5 A. I'm not aware of any such discussion. 6 Q. Can we turn to page 45 a few weeks later. This 17 June 7 2006, page 45. If we scroll down. 17 June, new call 8 taken, this is under the "Call Activity Log". Thank 9 you: 10 "... the online services are down, PM has ..." 11 Do you know what that abbreviation means? 12 A. No, is the simple answer. 13 Q. It looks like she was being told to reboot; is that sort 14 of turning on and off again? 15 A. Yes. 16 Q. If this was the same error messages as Mrs Misra had 17 previously received, would that be evidence of the 18 system working within or outside its normal parameters? 19 A. I think it was expected to be normal that you would get 20 occasional hardware failures in branches. I have had no 21 experience of the -- of how often they occurred, and 22 I was happy to rely on the fact that Andy Dunks had said 23 that the hardware failures in this branch were normal. 24 Q. If I carried on pointing out reports of hardware 25 failures in the branch going right up to December 2006, 126 1 including where a base unit replacement was necessary 2 because the counter kept turning itself off in the 3 course of transactions, your answer would be the same: 4 you were relying on Mr Dunks; is that right? 5 A. I think so, yes. 6 Q. Did you consider the possibility that, on any one of the 7 occasions where the engineer was replacing hardware, 8 they may have failed to take necessary steps to 9 synchronise the message store? 10 A. I didn't consider that. 11 Q. That would be problematic, if they failed to do that, 12 wouldn't it? 13 A. But it would also probably have been very obvious at the 14 time. 15 Q. On the data that you didn't have and therefore couldn't 16 analyse? 17 A. Yes. 18 Q. So, in answer to the question, what did you do to 19 respond to the suggestion that hardware faults at branch 20 were persistent and affected balancing, is that you 21 relied on reading Mr Dunks' statement? 22 A. Yes. 23 Q. Thank you. Can I turn the investigations into 24 shortfalls at West Byfleet by looking at the transcript 25 POL00029406. Again, this is the transcript we looked at 127 1 earlier. Can we look at page 58, please, at D, thank 2 you. This is Mr Tatford asking you questions. He asks 3 you: 4 "... have you -- in your analysis of these just 5 under half a million transactions, have you seen any 6 sign, even the slightest symptom of any computer fault? 7 "Answer: No, but then I've been doing very sort of 8 high level rough analysis on the stuff. But to find -- 9 to do any detailed investigation you need to have some 10 sort of idea about a fault happened at that particular 11 time." 12 That was a reference to the ARQ data, the half 13 a million transactions, wasn't it? 14 A. Yes. 15 Q. I think you knew and you realised that the Post Office 16 had only authorised the extraction of ARQ data for the 17 period of the 1 December 2006 until 31 December 2007, so 18 13 months? 19 A. Yes. 20 Q. The indictment period, did you understand, was the 21 29 June 2005 until 14 January 2008, in relation to the 22 theft count, so a period of about two and half a years? 23 A. I think I may have seen that sometime but, as far as 24 I was concerned, the only data I had to analyse was for 25 the limited period. 128 1 Q. Did you clock, did you realise, I've got data for only 2 about half the relevant period for the theft? 3 A. That was the data I was given and that was the data 4 I was asked to look at. 5 Q. But did you at the time realise -- 6 A. And I think I made that clear that that was the period 7 I had -- I thought it was well known that that was the 8 data that I had to look at and, therefore, the only data 9 I could comment on. 10 Q. Did it strike you as incomplete, in the sense that it 11 didn't match the period of the theft charge? 12 A. That was what I'd been asked to look at. 13 Q. You say here: 14 "... I've been doing very sort of high level rough 15 analysis on the stuff ... to do any detailed 16 investigation, you need to have some sort of idea about 17 a fault happened at that particular time." 18 So do I understand this to be what you were saying: 19 although you say that there would be evidence of faults 20 in the data, you did not conduct any analysis to try and 21 locate such faults in the data because you didn't have 22 a data range to narrow down the search. 23 A. The data that I had was -- it wasn't really possible to 24 do that because I was trying to correlate the cash 25 movements against the cash declarations, and they just 129 1 didn't correlate and, therefore, I wasn't able to track 2 down where losses had actually happened to investigate 3 in detail. 4 Q. So the answer -- 5 A. And that's really what I was trying to say there. 6 Q. Sorry? 7 A. And that's what I was trying to say with that answer 8 there. 9 Q. So the "No, but", the "but" is very important, because 10 you're saying, "I actually haven't been able to do any 11 real investigation because I need to know when the fault 12 is supposed to have occurred, so that I can use that 13 information to narrow of the search down and look in 14 this half a million transactions"? 15 A. Yes, yes. 16 Q. Had Fujitsu developed any sort of knowledge bank about 17 the kind of events or patterns that you could do 18 searches on in order to identify evidence of known 19 faults? 20 A. I don't know. 21 Q. None to your knowledge? 22 A. I can't think of any at the moment. 23 Q. The questioning continues: 24 "Have you been given any information at all from 25 Professor McLachlan as to any particular fault that 130 1 might have been happening at a particular time? 2 "Answer: Other than the Callendar Square problem, 3 no." 4 Would you agree that that approach really throws the 5 burden onto the defence for coming up with suggestions 6 as to what to look for, and then only then would you be 7 able to search through the 500,000 transactions. 8 A. Yes, and I think I'd indicated that in some of the 9 emails that I exchanged with Jarnail Singh earlier in 10 the year. 11 Q. "We need the defendant to identify where the particular 12 fault is before I can look for the particular fault"; 13 that's what you were saying? 14 A. I suppose so, yes. 15 Q. But Fujitsu did keep PinICLs, PEAKs and KELs, didn't it? 16 A. Yes. 17 Q. Could you not have used those in the indictment period 18 for two-and-a-half years to find out what other known 19 faults were in play during the relevant time? 20 A. I think I did look for whether there were PinICLs 21 associated with that branch. 22 Q. Yes, not PinICLs associated with that branch, PinICLs or 23 PEAKs or KELs that identified known faults that were in 24 play, which were in evidence, during the time that the 25 theft was alleged to have taken place? You could use 131 1 that as a narrowing tool, couldn't you? 2 A. I realise that now. I didn't think of it at the time. 3 Q. Isn't that a straightforward or an obvious approach, 4 rather than saying, "It's up to you, defendant, to tell 5 me where to look in my system for faults, I've got 6 a Knowledge Base which identifies faults, I'll use that 7 to look in the data"? 8 A. When you put it that way, I realise that now, but that 9 was not the way I was thinking at the time. 10 Q. That would have given you, even on the limited 13 months 11 that you were given, a clue or ideas as to what to look 12 for in the ARQ data, wouldn't it? 13 A. I understand that now. 14 Q. But what you were saying is that there was too much data 15 to do any detailed investigation, the only thing that's 16 been suggested is Callendar Square, which they knew 17 about because of Lee Castleton's case, "That's the only 18 thing that I've got to go on"? 19 A. That's what I am saying, yes. 20 Q. Those things, the PinICLs, PEAKs and KELs, they would 21 have given you an idea what to look for in the ARQ data 22 but the defence couldn't do that, could they, because 23 they didn't have the PinICLs, PEAKs or KELs, did they? 24 A. That's true. 25 Q. Do you know that Professor McLachlan repeatedly asked 132 1 for disclosure of what he called Fujitsu's records of 2 known faults? 3 A. I'm aware that he asked for that, yes. 4 Q. That had been repeatedly denied him, that request, 5 hadn't it? 6 A. Post Office had decided not to meet that -- respond to 7 that request. 8 Q. So the very tool, if the burden was going to be thrust 9 on to the defence, that might have enabled the defence 10 expert to carry out the task that didn't occur to you, 11 that was denied them, as well, wasn't it, by the Post 12 Office? 13 A. Yes. 14 Q. Was there any prohibition or nervousness or ban on 15 revealing the existence of the Known Error Log in legal 16 proceedings, that you're aware of? 17 A. Not that I'm aware of. I certainly discussed the 18 existence of it with Professor McLachlan. 19 Q. Is that why he was asking for it, do you think? 20 A. It could well be. After we had a -- I think it was 21 shortly after a meeting I had with him that he actually 22 requested that sort of information from Post Office. 23 Q. Did you form a view about the denial of that material to 24 him? 25 A. I didn't think that it would actually help him a lot but 133 1 I was prepared to sort of -- if Post Office had come 2 back and said to give that material for him, then it 3 wasn't something that I would have been concerned about. 4 Q. Can we go forward in the transcript to page 59, please. 5 Mr Tatford asked you, at A: 6 "... if a computer problem led to a problem with the 7 actual figures on the accounts, would you -- would that 8 problem manifest itself to the staff at the post 9 office?" 10 "Answer: Clearly, if there's a problem in the 11 accounts then -- and there were losses and things like 12 that showing, I would expect the staff to be complaining 13 to the Helpdesk to investigate what's gone on and that 14 could -- that might trigger an investigation by 15 ourselves." 16 A. That was my understanding of how things were supposed to 17 work. I now realise that the Post Office Helpdesk 18 wasn't actually very good at passing things on when they 19 should have been. 20 Q. The trouble with the answer, therefore, is that it 21 assumed that the Helpdesk handled calls as they should, 22 rather than batting them off? 23 A. Yes, that was the assumption I had. 24 Q. Can we look at some call records, please. At 25 POL00061793. Page 22, please, and if we scroll 134 1 forwards, please, and look at the bottom of the page, 2 a call on 20 February 2006 at 10.42. It's opened at 3 10.42 and closed two minutes later. If you look at the 4 text, the caller, on the left, is Mrs Misra -- 5 A. Yes. 6 Q. It says "Mrs Seema" but it's Mrs Misra. 7 A. Yes. 8 Q. The problem is recorded as: 9 "[Postmistress] states that showing £6,000 down from 10 balance ..." 11 Yes? 12 A. I see that, yes. 13 Q. Then text underneath it, "Call closed". So two minutes 14 after the call, after Mrs Misra says that the system is 15 showing £6,000 down from balance, she's told it's 16 an NBSC issue. 17 A. I see that, yes. 18 Q. Then if we scroll down, please. 10.45, can you see 19 that, the next call record? 20 A. Yes. 21 Q. Mrs Misra again, this one is closed down in three 22 minutes: 23 "[Postmistress] states that showing £6,000 down from 24 balance. Advised [it's an] NBSC issue. [Postmistress] 25 stated she was talking to the NBSC and got cut off ..." 135 1 Then the fuller text in the next box down: 2 "call close ... I advised [postmistress] I would put 3 her through. [Postmistress] was happy with this." 4 Then scroll down, please -- and keep going. 3.40, 5 same day, call opened and closed within two minutes. 6 Seema Misra, the call. 7 "[Postmistress] states that her system is showing 8 different values for certain products." 9 Then if we scroll down, just stop there: 10 "Call close by Matthew Fry: [postmistress] states 11 that her system is showing incorrect values. 12 [Postmistress] transferred." 13 Next day, the 21: 14 "... last couple of weeks they have had problems 15 with Horizon kit and it is always showing that they are 16 down in money." 17 Carrying on: 18 "Call closed ... she has been advised by the NBSC, 19 advised [postmistress] to follow this ..." 20 Just stopping there, I'm not going to go through all 21 of these, there's a lot of them. Did you access these 22 call records as part of your work preparing to give 23 evidence in witness statements and then orally in Seema 24 Misra's case? 25 A. No, I did not. 136 1 Q. Did you therefore not correlate any of this information 2 with the data in the limited period of the 13 months 3 that you had? 4 A. No, I did not. 5 Q. I think from your earlier answers, I took that you 6 didn't know that the Helpdesk was perhaps not as 7 effective as it should have been, that people were, for 8 example, continually rerouted to the NBSC? 9 A. Yes, well, it was the NBSC, I think, that was 10 responsible to handle that sort of call but I now 11 understand that they didn't do it very well. 12 Q. Does any of this help or hinder us with assessing 13 Mr Dunks' statement that all calls are of a routine 14 nature and don't fall outside the normal working 15 parameters of the system? 16 A. I didn't do that sort of analysis of the calls at the 17 time. I just took his word for what he had actually said 18 as the summary. 19 Q. Do you think you probably should have looked at it 20 before allying yourself with what Mr Dunks was saying? 21 A. With hindsight, I can see that, yes. 22 Q. I think once the trial started and Mr Dunks was 23 testifying about the calls, you started make some 24 enquiries then, didn't you? 25 A. I did indeed. He referred to one call that had been 137 1 referred through to SSC and Anne Chambers had actually 2 examined the data for the branch at a particular time. 3 I think that was sometime in 2006 and I did contact her 4 to ask if she had any more useful information, and I did 5 look at the PinICL that was associated with that. 6 Q. Let's just look at that briefly, FUJ00153395. So this 7 is in the course of the trial. 8 A. Yes. 9 Q. You're emailing Anne Chambers on 13 October 2010 at 10 3.15: 11 "Anne, 12 "In court today Andy Dunks was asked about all sorts 13 of Helpdesk calls. 14 "One of them looked interesting ..." 15 Then I think that should read "so": 16 "... [so] I searched PEAK for more info." 17 A. Yes. 18 Q. I think that's the PEAK you attached. There's a pdf at 19 the top -- 20 A. I believe so. 21 Q. -- ending in 673: 22 "Do you remember anything about it?" 23 "NB I've not been asked to get this info officially 24 so you're in the clear still! 25 "I'm at ..." 138 1 A. "Home", I think that's supposed to say. 2 Q. "... if you want to ring me." 3 Why did the PEAK look interesting? 4 A. Because it was one that had been raised to do with 5 a discrepancy in the branch that had been investigated 6 by her. So I can't remember now exactly what the PEAK 7 said. I think she said she did a full investigation and 8 couldn't find any problem but I was just wondering if 9 she had anything to add to that. 10 Q. The "you're still in the clear", presumably, do you mean 11 by that that she was keen not to give evidence again? 12 A. I think that's what that's referring to, yes. 13 Q. Therefore, you hadn't been asked to get this officially 14 so there wasn't a prospect of her being dragged into 15 these proceedings? 16 A. Correct, yeah. 17 Q. The defence didn't have access to the PEAKs, did they? 18 A. No, they didn't. 19 Q. So you had a privileged position in relation to 20 Fujitsu's records? 21 A. Yes. 22 Q. Did you think at the time that that may give the Post 23 Office an unfair advantage over the defence? 24 A. I don't think I thought it through at the time but 25 I understand now why you're asking the question and, 139 1 yes, I accept that that wasn't right. 2 Q. That aside, do you think that whatever had piqued your 3 interest in the PEAK, you should have discussed it with 4 Professor McLachlan or at least shown him the document, 5 together with any information that Anne Chambers gave 6 you in response to your request to ring her if she 7 wants. 8 A. Again, I realise now that that's what I should have done 9 but I wasn't aware then that that was something that 10 I needed to do. I'd not been asked to look into this at 11 all. I just happened to have picked up listening to 12 Andy Dunks' evidence that there was this software call 13 that had come through and, therefore, I thought that 14 I would do a little bit of digging myself into it, and 15 no one had asked me to do that, so that's what was 16 behind this, to see if it would actually shed any light 17 of what had been going on. 18 Q. I think when you came to give evidence, you were 19 actually asked about Mr Dunks' oral evidence when you 20 were giving evidence? 21 A. I can't remember that at the moment, sorry. 22 Q. Let's have look. POL00029426, at page 7. If we scroll 23 down, please, at H: 24 "... you are aware, are you, of the witness 25 statements of Mr Dunks -- 140 1 "Answer: Yes, I am." 2 Then this is Mr Tatford: 3 "-- who provided a rather exhaustive list of all the 4 calls to the helpline. Do any of those calls to the 5 helpline -- I do not know if you heard his evidence 6 yesterday? 7 "Answer: Yes, I did indeed. 8 "Question: Did any of what you have read or of what 9 you heard yesterday from Mr Dunks, did that cause you 10 any concern as to your view that there is no evidence of 11 any computer fault? 12 "Answer: No, I -- I've not -- I haven't got 13 Mr Dunks' experience in examining call logs and things 14 like that but I was quite happy with his comment that 15 the level of calls from the branch were typical for 16 other branches." 17 Do you think, on reflection, that reply obscured 18 more than it revealed? 19 A. I didn't at the time. I understand now that maybe that 20 would have been an opportunity to talk about the 21 conversation -- well, I'm not sure whether I spoke to 22 Anne or not but, certainly when I looked at the PEAK, 23 there was nothing there to show that there was actually 24 a problem at the time and I believe the PEAK showed that 25 she'd done some examination and not found any faults. 141 1 So, therefore, I didn't see the need to mention it in 2 that sort of context. 3 Q. But the level or the number of calls, whether or not 4 that was typical of other branches, was beside the 5 point, wasn't it? 6 A. I thought that what I was saying there was -- I'm not 7 quite sure what you would have expecting me to say there 8 because, as I say, I was doing as I'd done in my witness 9 statement, saying that I thought that Mr Dunks knew 10 better how many calls would be typical in a branch such 11 as this. 12 Q. Well, you say, you don't know what I would have expected 13 you to say -- and I'm not going to put words into your 14 mouth -- but something like "I know that hardware faults 15 can lead to accounting irregularities, one needs to go 16 through the Helpdesk call records very carefully to look 17 for those hardware irregularities and then cross-compare 18 them to the ARQ data, using that as a key to look for 19 errors. I haven't done that and, in any event, I've 20 only got 13 months' of data which is about half the 21 relevant period". 22 A. I'd not thought it through in that way but, with 23 hindsight, maybe I should have said something like that. 24 Q. That would be the complete answer, wouldn't it? 25 A. Possibly. 142 1 Q. Can we look, please, at FUJ00153390, 11 October, which 2 I think is the second day of trial? 3 A. The first day, I think. 4 Q. The first day? 5 A. The first day, yeah. 6 Q. First day of trial at 7.00 at night. 7 A. Yes. 8 Q. You're sending, essentially, is this right, the NT event 9 logs? 10 A. Yes. 11 Q. That's because of his requests or focus on Callendar 12 Square; is that right? 13 A. Yes. I think during the trial we'd actually gone 14 through the event logs on my laptop and he'd asked me 15 could I send him a copy, so I did so. 16 Q. You answered some questions about these logs -- if we 17 look at those in the trial -- as you just said, 18 POL00029406, at page 67 at C. This is you being asked 19 questions by Mr Hadrill. 20 Just scroll up a little bit, so we can see the 21 question before. Thank you. So at C: 22 "Questions are asked about Callendar Square and, lo 23 and behold, you summon up then a third log and it is 24 called an NT log. Yes? 25 "Answer: Yes. 143 1 "Question: Which you had not bothered looking at 2 before? 3 "Answer: I had looked at them before when I was -- 4 produced my statement on what happened in Callendar 5 Square back in March. 6 "Question: You had never given Professor McLachlan 7 the benefit of looking at them? 8 "Answer: That's true. 9 "Question: You are compiling a -- 10 "Answer: He hadn't asked me for them. 11 "Question: Sorry? 12 "Answer: He hadn't asked me for them. I had passed 13 on the information from Callendar Square", et cetera. 14 A. I think that's just an example of me not understanding 15 that I had a responsibility for doing any disclosure. 16 Q. Aside from that, is it right to say that neither you nor 17 Professor McLachlan had time in the course of the trial 18 to analyse the data for anything other than the signs of 19 the Callendar Square bug. 20 A. I can't remember exactly what we had. I did give him 21 the full NT event logs for the full period that had been 22 provided to me back in February. Back in February, Anne 23 had done an analysis of the NT events and -- to see 24 whether there were any problems there, and we hadn't 25 seen anything that was startling about -- from the NT 144 1 events and now I was sharing these logs with Professor 2 McLachlan. 3 Q. You said that the pair of you sat in front of your 4 laptop in the course of the trial? 5 A. I believe so, yes. 6 Q. Was that searching for, or analysing the data for, signs 7 of the Callendar Square bug? 8 A. That was the main reason that we were looking at things 9 then, yes. 10 Q. You and he didn't have time to analyse all of the data 11 for any signs of bugs? 12 A. No, but he was -- as I say, having given him the things, 13 then he was free to look for anything he wanted and come 14 back to me with specific questions. I don't believe 15 that he did so but believe that Mr Hadrill did, as part 16 of the dialogue here. 17 Q. Can we go back to page 58, please, just go down to D. 18 Would you agree this was an important part of your 19 evidence, between D and G? It's essentially the core of 20 what you were saying. 21 A. Yes, I think so. 22 Q. The question: 23 "[Has] your analysis of ... just under half 24 a million transactions ... seen any sign, even the 25 slightest symptom of any computer fault? 145 1 "Answer: No, but then I've been doing very sort of 2 high level rough analysis ... to do any detailed 3 analysis you'd have to have some sort of idea about 4 a fault happened at that particular time." 5 "Question: Have you been given any information ... 6 from Professor McLachlan [about a particular fault]? 7 "Answer: Other than [Callendar Square], no." 8 Do you agree that a complete answer would have 9 included the following: however, Professor McLachlan 10 hasn't been provided with all of the records that are 11 available to me"? 12 A. I wasn't aware that I needed to say that and, by then, 13 he had had the information. But, as I say, I didn't 14 know that he needed -- I didn't know that he hadn't got 15 it until the Monday. 16 Q. A complete answer would be, "He has asked for, but has 17 not been given Fujitsu's records of known errors"? 18 A. I didn't know that I needed to say that. I thought he 19 could have said that himself, if he thought it was 20 important. 21 Q. "And that it's possible that if he or I or of both of us 22 had examined the records for all known faults operative 23 in the two and a half years of the alleged theft, we 24 would have had a better idea what to look for in the 25 500,000 transactions"? 146 1 A. I understand that now. I hadn't at the time. 2 Q. The testimony continues by Mr Tatford asking you: 3 "No. And, in relation to computer problems, you 4 have told us at the beginning of your evidence that 5 a computer problem displays symptoms?" 6 "Answer: In my experience, yes. 7 "Question: ... would a computer problem on the 8 Horizon system display symptoms for the people working 9 in the post office? 10 "Answer: Not necessarily for the people working in 11 the post office but if there were investigations carried 12 on, there would be symptoms in the logs. 13 "Question: And if a computer problem led with the 14 actual figures on the accounts ... would that problem 15 manifest itself to the staff at the post office? 16 "Answer: Clearly, if there's a problem in the 17 accounts then -- and there were losses and things like 18 that showing, I would expect the staff to be complaining 19 to the Helpdesk to investigate what's gone on and that 20 could -- that might trigger an investigation by 21 ourselves." 22 Do you accept that a complete answer would have 23 included, "Mrs Misra, I understand, did call the 24 Helpdesk about shortfalls in her accounts"? 25 A. I don't think, at that stage, I was -- yes, I probably 147 1 would have heard Mr Dunks saying that but I didn't think 2 that I would have needed to have said that. 3 Q. "I've investigated an investigation by a member of the 4 SSC staff and she, Anne Chambers, decided there was no 5 sign of a system problem and so the call wasn't taken 6 any further at that stage"; ought you to have disclosed 7 that? 8 A. Looking at that now, maybe. But, at the time, it didn't 9 occur to me that that was relevant. 10 Q. A complete answer would be "Professor McLachlan and 11 I haven't been able to check whether that investigation 12 by Anne Chambers was adequate because the Horizon data 13 that we've got begins in December 2006 and, therefore, 14 it doesn't cover the period of those calls in February 15 2006"? 16 A. Again, I understand that now but I didn't at the time. 17 Q. A complete answer would be, "I haven't examined the call 18 logs themselves but, even on what I know from Andrew 19 Dunks' evidence, there were numerous calls regarding 20 hardware failures, which I know from my experience can 21 lead to accounting irregularities"? 22 A. Again, I understand that now but I didn't think of it at 23 the time. 24 Q. And a complete answer would be, "I've given, a couple of 25 days ago, Professor McLachlan the NT event log which 148 1 might help about this possibility but he's only had it 2 since the first day of the trial, I've not myself looked 3 at it for any other signs other than the Callendar 4 Square bug"? 5 A. I think I had looked at the events for anything else 6 that was unusual back in February. 7 Q. Okay. What do you mean, "unusual"? 8 A. There are some events that you get sort of on a fairly 9 regular basis, so I was looking for ones that I could 10 see that would have caused a problem. 11 Q. Overall, were you untroubled by, and happy with, the 12 evidence you'd given at trial? 13 A. At the time I was. I clearly appreciate now that it 14 wasn't as good as it should have been but, at the time, 15 I felt happy with it. 16 Q. Can we look at FUJ00225196. Look at the second email on 17 the page from you on 22 October, subject "Court told 18 postmistress had 'her fingers in the till'": 19 "Penny, I did a Google search last night and found 20 this from last week." 21 It's an article from Surrey Online, essentially 22 getsurrey.co.uk. 23 A. The reason I was interested in that was because it 24 referred to me as "Professor Jenkins", rather than 25 Mr Jenkins, and that was the reason I was expressing 149 1 an interest in there. It was nothing to do with the 2 actual trial, as such, it was that I had magically 3 acquired this title of Professor in the article, as 4 obviously a typo. 5 Q. You say: 6 "They don't seem to have updated it yet to add in 7 the guilty verdict. 8 "Looks like I've acquired a new title!" 9 A. Which is what that last comment was about. 10 Q. Then if we scroll up: 11 "Nice one Gareth. Looks like you now have 12 a sideline of resident expert witness in future Post 13 Office fraud cases." 14 Which was indeed the case, wasn't it? You did have 15 a sideline as resident expert witness? 16 A. I did in a few cases. 17 Q. Did you subsequently pass this on to anyone else, that 18 you had -- 19 A. I probably sent it home, so that my wife could see the 20 article and, again, the reason for that was to do with 21 the fact that I'd been referred to as a professor. 22 Q. Thank you. That can come down. 23 Following your evidence in the Seema Misra case and 24 the guilty verdict, you were asked to provide a series 25 of generic witness statements, weren't you? 150 1 A. This was about a couple of years later, yes. 2 Q. Yes, in October 2012? 3 A. Yes. 4 Q. Can we look, please, at POL00096978. This is from 5 Mr Singh to you, copied to others. 6 "Dear Mr Jenkins 7 "Welcome from your annual leave ..." 8 I'm going to read this, as I usually do, exactly how 9 it's written: 10 "... and your assistance advice in the past 11 prosecution cases and I understand you are assisting my 12 colleagues at present. I need your urgent assist judge 13 has this morning ordered the prosecution to have the 14 following report ready to be served within Seven days. 15 On advise Post Office Limited have appointed one of 16 their investigators, Helen Rose as Disclosure Officer 17 dealing with Horizon challenges. She has prepared 18 a document/spreadsheet detailing all such cases, past 19 and present, approximately 20 in total, although none 20 thus far successfully argued in court. Post Office 21 Limited have been advised to obtain, an experts report 22 from Fujitsu UK, the Horizon system developers, 23 confirming the system is robust. Post Office Limited 24 maintain the system is robust, but in light of adverse 25 publicity, from legal viewpoint is that defence should 151 1 be given opportunity to test the system, should they 2 still wish to do so, on consideration of our report. 3 "You will need to consider Disclosure Officers 4 document/spreadsheet (see attachments) and need to 5 address in your report the followed issues: 6 "1) A description of the Horizon system (in laymen's 7 terms so that a jury can understand what it is and what 8 it does). 9 "2) A declaration that it is yet to be attacked 10 successfully. 11 "3) A summary of the basic attacks made on the 12 system concentrating on any expert reports served in 13 past cases. If there are none then state that no expert 14 has yet been found by any defence team civil or criminal 15 to attack the system (at the moment there seems to be 16 little more than [gripping] by defendants that the 17 system must be at fault without saying how). 18 "4) Plainly, like all accounting systems, there is 19 room for human error (keying in wrong amounts etc) are 20 you able to state that innocent human error is unlikely 21 to produce the types of discrepancies of many thousands 22 of pounds over many months. 23 "If you require any further information or wish to 24 discuss please do not hesitate to contact me." 25 As far as you're concerned, is this essentially the 152 1 origin of the request for the generic witness statement? 2 A. Yes, it is. 3 Q. You'll see from the top of the page that this was sent 4 to you on 1 October 2012, and you ended up signing the 5 generic witness statement on 5 October 2012? 6 A. Yes. 7 Q. So, in a relatively short period of time, and the 8 message says that this was urgent? 9 A. Yes. 10 Q. The request was framed in a way that the generic witness 11 statement, as perhaps is obvious from its title, was not 12 going to address the specifics of any particular case? 13 A. Yes. 14 Q. Did you form a view at the time as to the adequacy of 15 the instructions that you were being given here? 16 A. Not really. It was -- I suppose, a better way to 17 describe it: it was a typical garbled email from Jarnail 18 Singh. 19 Q. Given that it was a "typical garbled email" from Jarnail 20 Singh, didn't that cause you concern? 21 A. I tried to make the best sense of it that I could. 22 Q. Can we look at your reply, please. POL00096983. 23 A little later that day -- can we just go back, please, 24 to POL0006978. Yes, that was at 11.04 in the morning. 25 Let's go back to POL00096983, your reply at 2.10, 153 1 you say: 2 "... here are the two existing reports I mentions 3 ..." 4 It looks like it may have been catching! 5 A. Yes, sorry. Typing is not one of my strong points. 6 Q. "... I [mentioned] regarding Horizon and Horizon Online 7 integrity. 8 "I'll try and produce a further short report 9 addressing your specific points below in the next few 10 days. 11 "Hopefully that will cover off your requirements." 12 So you were sending back the Horizon data integrity 13 reports about Legacy Horizon and Horizon Online that we 14 looked at yesterday? 15 A. Yes. 16 Q. Why were you sending those back as the answer to his 17 question, essentially, or part of the answer? 18 A. Because I thought they were part of the answer to his 19 question, and they would give him something to be going 20 on with whilst I tried to frame a report addressing the 21 specifics that he raised. 22 Q. Can we look at your third witness statement, please, at 23 paragraph 541, at page 186. You say: 24 "I provided my two data integrity reports to 25 Mr Singh ... because I thought they were a good starting 154 1 point for the purposes of grabbing Legacy Horizon and 2 Horizon Online and the mechanics of their respective 3 audit trails. As I have explained in my second 4 statement to this Inquiry, these reports were not 5 a survey of bugs or other issues which had affected 6 Horizon and had not been prepared for that purpose." 7 A. Yes. 8 Q. Mr Singh had asked you to prepare a witness statement, 9 essentially, that said, amongst other things, that the 10 Horizon system was robust, hadn't he? 11 A. I can't remember exactly what he was saying but what 12 I thought -- I thought this was a good start in terms of 13 trying to address the robustness of Horizon, yes. 14 Q. If we just go back, it might not have been a witness 15 statement. I just want to check that. 16 A. He was asking for a report, I think. 17 Q. A report? 18 A. There was a report I produced, which later got turned 19 into a witness statement. 20 Q. Okay, so he was asking for a report that addressed, 21 amongst other things, to confirm that the system was 22 robust. 23 A. So what I -- so the next day I think I did produce 24 a Word document in the form of a report which covered 25 what I thought was the scope of what he was asking. 155 1 Q. Confused as Mr Singh's request might have been, why 2 didn't you write back and say, "No, I can't just provide 3 a generic witness statement that's to be used in support 4 of a prosecution. I can only look at branch-specific 5 data to examine whether there's evidence of a specific 6 problem at a specific branch"? 7 A. I think I had sought guidance from my colleagues within 8 Fujitsu as to how I should address Mr Singh's request. 9 Q. Your approach in the past was, "It's all about the 10 branch, it's not about the system in general"? 11 A. Yes. 12 Q. Why wasn't that your reply to him immediately, "I can't 13 give you a genericised statement. You need to be 14 looking at individual data which is branch-specific"? 15 A. I'd not thought of it that way. I thought I was just 16 providing a high level overview of how Horizon was 17 working in general. 18 MR BEER: Thank you. If we can stop there and take the 19 afternoon break until 3.15, please. Thank you very 20 much. 21 (2.58 pm) 22 (A short break) 23 (3.15 pm) 24 SIR WYN WILLIAMS: Have you had a sufficient break 25 Mr Jenkins? 156 1 A. I think so. I'd like to just get the day over with, 2 please. 3 SIR WYN WILLIAMS: Fine. 4 MR BEER: I'm aiming to finish by 4.00. 5 Can we look at your third witness statement, please, 6 at page 184. 534, you're referring to Mr Singh's email, 7 the one that we looked at before the break. 8 A. Yes. 9 Q. You say: 10 "[It] didn't mention any specific prosecution, name 11 any [subpostmaster] defendant or provide any information 12 about any prosecution that my statement was required 13 for. I understood that what [Post Office] wanted was 14 a generic statement that did not go into the specifics 15 of any case and that responded to the four issues 16 Mr Singh had set out in his email." 17 A. Yes. 18 Q. Then you say at 535: 19 "Mr Singh's email ... described and attached two 20 documents ..." 21 Then you give the references, the second which was 22 an Excel spreadsheet list interesting number of previous 23 prosecutions, that's the one ending in 468 -- 24 A. Yes. 25 Q. -- sorry, ending in 089. 157 1 A. Yes. 2 Q. "[648] was a Word document describing five of those 3 prosecutions. These five prosecutions [including] two 4 which I knew about (Rinkfield and West Byfleet). 5 I didn't know about [any of] the others at the time." 6 I'm not going to take you to the underlying 7 documents, to the Excel, but the spreadsheet had 25 or 8 so branches listed in it where there had been 9 a prosecution and Horizon issues had been raised. On 10 that list, Hughie Thomas' branch was not listed. 11 A. I can't remember. I'm sorry. 12 Q. Would you have gone through that list -- 13 A. I would have skimmed through looking for ones that 14 I recognised and I think I recognised about three or 15 four of them, something like that. I think I listed 16 them in the first draft of my report, the ones that 17 I did recognise. 18 Q. So you didn't see that the list was incomplete because 19 it didn't list Hughie Thomas' branch in Wales? 20 A. I took it at face value. 21 Q. I see. Therefore, there was nothing triggered in you 22 that this list that the Post Office are providing me 23 with doesn't provide a complete picture of all 24 prosecutions in which Horizon integrity had been raised? 25 A. No, I -- it was presented to me as a list of -- like 158 1 that, so I took it on face value and just didn't 2 consider it in that much detail. 3 Q. Okay, and if we go down the page to 537, you say: 4 "Whilst I understood that Mr Singh wanted a generic 5 report that did not address the specifics of any given 6 case and that responded to the four issues he had 7 identified, I was concerned that I did not know exactly 8 I how to approach this task." 9 I think that's a reference to -- that you made 10 earlier in your evidence before the break -- 11 A. Yes. 12 Q. -- when you said, "Yes, and I went off to the Fujitsu 13 lawyers"? 14 A. Yes, actually looking at this, it wasn't the lawyers at 15 this point but it was certainly -- 16 Q. Managers? 17 A. -- managers and things, yes. 18 Q. You say: 19 "That is why at 11.52 I emailed Peter Thompson (Head 20 of Customer Services), Howard Pritchard (Head of 21 Security) and Ian Turner (Development Manager and my 22 boss) copying Pete Newsome and James Davidson, saying 23 'help please' and seeking 'urgent guidance'." 24 Can we look at that please, FUJ00156645. Thank you. 25 This is the email you're referring to. 159 1 A. Yes. 2 Q. So Jarnail Singh sent the email with the four requests 3 in it, and you say: 4 "Help please. 5 "I've looked through the attachments on Jarnail's 6 email and it isn't at all clear exactly what he wants. 7 I suspect he wants a further report explaining why 8 Horizon integrity is okay. I had hoped that somebody 9 would have sent him the report I produced ... for Dave 10 Smith 3 years ago while I was on leave. 11 "However before I contact Jarnail I need some 12 guidance as to how this fits into my priorities. Should 13 I drop everything else and concentrate on [keeping] 14 Jarnail happy (one extreme) or should I wait for ..." 15 Is that a change proposal? 16 A. Yes, because I could see that this could involve quite a 17 bit of my work and, therefore, it needed to be paid for 18 by Post Office. 19 Q. A change proposal, ie a formal mechanism raising this 20 work? 21 A. Yes. 22 Q. "... to be raised and processed before I start (the 23 other extreme) or more likely, some in-between course. 24 "As a matter of courtesy, I would like to get some 25 kind of response to Jarnail today at least indicating 160 1 what we (I) propose to do. 2 "I'd appreciate some urgent guidance. Until I get 3 some I shall do nothing further on this other than send 4 a holding email to Jarnail around 5.00 if I've heard 5 nothing by then." 6 So your concern appears to be to secure management 7 sign-off as to where this sits within your day-to-day 8 priority -- 9 A. Yes. 10 Q. -- rather than concern with complying with the substance 11 of the request that was made to you? 12 A. At that stage it was really a case of, yes, is there 13 something I should be doing urgently or not, or should 14 I be doing at all, even. 15 Q. You don't question in this, where you're seeking 16 guidance, the appropriateness of the requests that are 17 being made to you? 18 A. I'd not thought of that but, clearly, it was open for my 19 management to say that it was inappropriate. 20 Q. Does that reflect the fact that you didn't see the task 21 as fundamentally inappropriate, namely providing 22 a generic witness statement that wasn't branch-specific, 23 and might be deployed without any reference to 24 an analysis of underlying data? 25 A. I've now realised that was totally inappropriate but 161 1 I didn't at the time. 2 Q. That's why we don't see you raising that -- 3 A. Yeah. 4 Q. -- in this email; is that right? 5 A. Yes. I was assuming that he knew what he was doing and 6 was asking me to do something that was appropriate. 7 Q. We saw, I think, at 3.10 that day, you sent over the two 8 Horizon integrity reports? 9 A. Yes. 10 Q. Does that mean that that was essentially the holding 11 response that you are referring to in the last paragraph 12 here? 13 A. It could be. I can't remember the consequences of what 14 happened at that time but that was clearly a sort of 15 holding response. 16 Q. What assistance, if any, did your managers provide you? 17 A. I can't remember. I may have had a conversation with 18 one of them or something in the meantime to agree that 19 that was something that should be done. I just can't 20 remember the details. 21 Q. What about more generally, the plea for help that you 22 were making in this? I don't think we've got a -- 23 A. I don't think that I've seen an email responding to that 24 but I think, later on, some of the Fujitsu lawyers got 25 involved in the report that I was drafting over the next 162 1 couple of days but I can't remember the exact details of 2 that. 3 Q. Can we go forward in your witness statement, please, to 4 page 187. Third witness statement, page 187, and 5 paragraph 545. You say: 6 "I understand that the question is being asked in 7 this Inquiry of why I didn't, in my draft report, refer 8 to bugs, errors [and] defects which had affected 9 Horizon. Quite simply, I didn't think [Post Office] 10 wanted a report which did anything apart from addressing 11 the four issues Mr Singh had asked me to address. In 12 terms of Mr Singh's general reference to wanting 13 a report 'confirming the system is robust', my draft 14 report referred to the components of Horizon that 15 ensured that the writing and storage of data had 16 integrity. My belief was that if and when there were 17 problems in the system, that would be detectable and 18 leave traces and we would pick those up in any specific 19 case. My view (and I think this was the common 20 understanding in Fujitsu) was that there would be 21 problems (there always are) but they were spottable and 22 there were mechanisms to spot and fix them, and all of 23 these components made the overall system robust." 24 If that was your state of mind, if that was your 25 appreciation, did you take any steps to check that all 163 1 bugs were being picked up, handled and fixed? 2 A. Not at that particular time, no, but I thought that the 3 processes that we had in place were doing that sort of 4 thing and I still believe that they did. 5 Q. You don't reply to Mr Singh by saying, "Look, the system 6 plainly has bugs but we hope that the system we have in 7 place works. It doesn't work 100 per cent of the time, 8 or at least I can't say it does. I can't say anything 9 more than that, ie that's how the system is designed to 10 work. I need to check underlying raw data to see 11 whether or not there is any evidence of a bug, known or 12 unknown, impacting on the data"? 13 A. I didn't think I needed to make a response along those 14 lines, though I clearly understand with hindsight that 15 perhaps I should have done. 16 Q. If we go forward to page 548, please, which is over the 17 page, you say: 18 "It appears that there was legal oversight of the 19 provision of my report at Fujitsu. No one from [Post 20 Office] or Fujitsu suggested that my report needed to 21 detail any bugs, errors or defects which had arisen in 22 Legacy Horizon or Horizon Online. David Jones ..." 23 That's the lawyer? 24 A. Yes. 25 Q. "... had been involved in Mrs Misra's case, in which he 164 1 had passed on to [Post Office] my raising the Craigpark 2 issue (which lawyers in Fujitsu and [Post Office] knew 3 about anyway). The Callendar Square bug that featured 4 in Mr Castleton's case and Mrs Misra's case. Mr Singh 5 was well aware that my approach in Mrs Misra's case was 6 that the underlying data needed to be considered for any 7 evidence of a system issue at a particular branch. 8 I now know that [Post Office's] Criminal Law Team were 9 aware of the Receipts and Payments Mismatch bug. 10 Plainly, there was a far greater exchange of information 11 between Fujitsu and [Post Office] about the problems 12 that arose from time to time in Horizon than this. But 13 even in terms of just those individuals dealing with 14 this request for a report, there was clearly knowledge 15 within [Post Office] of a number of Horizon-related 16 problems. No one suggested that these should be 17 included in the report that [Post Office] had asked me 18 to produce." 19 You say, "No one suggested this should be included 20 in the report"? 21 A. Yes. 22 Q. Forgive me whilst I take a lozenge. 23 SIR WYN WILLIAMS: Yes. 24 MR BEER: This report was going to form the basis of 25 a witness statement that you were in due course to sign? 165 1 A. I don't think I knew that at this time. 2 Q. By the time that you signed the witness statement, you 3 knew that the report that you drafted formed the basis 4 of that witness statement? 5 A. Yes. 6 Q. That witness statement was going to be backed by 7 a statement of truth? 8 A. And I believed that the report, as it was written, was 9 true. 10 Q. Did you not know that the picture was more complex than 11 was presented in the final draft of your 5 October 12 generic statement? 13 A. I'm not sure what you mean by "more complex". 14 Q. Well, for example, you didn't include a caveat linking 15 your limited information on any individual case, or 16 saying that Horizon was not fault free? 17 A. I accept that I did not include those caveats. I didn't 18 think I needed to, at the time. 19 Q. This tends to suggest, this paragraph, that you thought 20 perhaps that it was the responsibility of others to tell 21 you what qualifications you needed to include in your 22 report and then witness statement. Is that what you 23 were intending to say here? 24 A. I think I would -- I was seeking guidance. I didn't 25 receive any guidance to do other than that which I did 166 1 and, therefore, I think there were a number of 2 opportunities where others who were more knowledgeable 3 of these things than me to actually suggest that I added 4 further things in to there, and they didn't suggest that 5 to me. 6 Q. Thank you. Now, I think it's right that you prepared 7 a draft report on 2 October 2012, and then a witness 8 statement on 5 October 2012, and between those two 9 events there had been some revisions. One of those 10 revisions was that somebody had inserted the words 11 "I understand that my role is to assist the court" in 12 the opening paragraph of the witness statement? 13 A. Correct. 14 Q. Those words weren't in the draft report that you had 15 submitted originally back on 2 October? 16 A. Correct. 17 Q. Do you know who inserted them? 18 A. I think I've recently understood that it was one of the 19 lawyers in Cartwright King but I'm not 100 per cent 20 certain of that. 21 Q. Did you ask a question about why somebody had written 22 something into your witness statement that said, 23 "I understand my role is to assist the court"? 24 A. No, it didn't seem to -- it didn't seem to contradict 25 anything that I'd said in the report, so I thought it 167 1 was just something that needed to be included in it as 2 a legal document. 3 Q. You, I think, had had some concerns about both Post 4 Office's and Fujitsu's management of the prosecution 5 cases and your requirement to give input into them? 6 A. I'm not quite sure -- 7 Q. Taking a step back, by October 2012, you said that their 8 management of the case was chaotic, the Post Office. 9 A. And certainly the subsequent events over the next few 10 months with this use of the generic statements was even 11 more chaotic, yes. 12 Q. By this time, did that concern about their management of 13 the cases trigger anything in you in thinking why is 14 somebody adding "I understand my role is to assist the 15 court"? 16 A. No, it didn't. I just assumed everyone -- that others 17 knew what they were doing. 18 Q. Did you ask yourself: are there any legal reasons for 19 including this? 20 A. I assumed that the reason it had been added in was it 21 needed to be done and I couldn't see any harm in it 22 being included in there. 23 Q. If we can go back to your witness statement, please, 24 page 194, paragraph 571. You're dealing with the last 25 paragraph of your signed witness statement of 5 October 168 1 2012, the generic witness statement, and you say: 2 "The concluding paragraph of my witness statement 3 addressed the fourth and final issue." 4 That's the fourth and final issue in Jarnail Singh's 5 email of 1 October? 6 A. Yes. 7 Q. "To recap, Mr Singh had asked me to confirm that 8 innocent human error was unlikely to produce large 9 discrepancies over a number of months. The wording 10 I used in response was ... 11 "'In summary I would conclude by saying that I fully 12 believe that Horizon will accurately record all data 13 that is submitted to it and correctly account for it. 14 However, it cannot compensate for any data that is 15 incorrectly input into it as a result of human error, 16 lack of training or fraud (and nor can any other 17 system)'." 18 By this, were you seeking to imply that Horizon was 19 both robust and infallible? 20 A. I certainly wasn't saying it was infallible. 21 Q. Were you seeking to imply that it was robust? 22 A. I was saying that it would accurately record what had 23 happened and I believed that it did. 24 Q. You knew, I think, that the Post Office wanted 25 statement, to use this statement, to rely on it to show 169 1 that Horizon was robust. That had been what was said in 2 Mr Singh's email to you? 3 A. I can't remember the exact wording but I'll take your 4 word for it. 5 Q. What did you think, then, the purpose of your generic 6 statement was? What was it going to be used for? 7 A. To show that you could rely on the data that was found 8 in the audit trail from Horizon. 9 Q. That limited purpose? 10 A. And that's what I was trying to say here, yes. And the 11 description in the statement is primarily about 12 describing how the audit trail worked and how it was 13 actually recording the data that had been input into the 14 system and how it was stored into the audit servers and 15 retrieved for ARQ data. 16 Q. What this doesn't do is look at the other side of the 17 coin and look at instances where Horizon doesn't 18 accurately record all data because of bugs, errors and 19 defects in the system, does it? 20 A. I am not aware of bugs, errors and defects that stop it 21 recording things, other than the hardware issues that we 22 spoke about this morning, and that was already being 23 referred to in the integrity paper that was associated 24 with this statement. 25 Q. Can I turn briefly to look at the use of the generic 170 1 statement, POL00097061. If we scroll down, please. 2 I should give you the context by looking still further 3 down. The case of Patel: 4 "Hi Gareth ..." 5 This is from Sharron Jennings in the Post Office: 6 "This is the one that you supplied the expert report 7 and witness statement for the week before last. 8 Apologies for not explaining that properly in the 9 previous email, it was a blanket email for all 10 witnesses! It is unclear at this stage who will be 11 required as witnesses and which evidence will be 12 accepted without the need for attendance. I just 13 thought if I let everyone know they can pencil it in and 14 then I can let you know nearer the time." 15 Then scroll up. 16 A. Well, in fact, if you go further down there, then what 17 had happened was I'd had this email out of the blue 18 asking me to reserve some diaries for a court date for 19 something I'd never heard of before. 20 Q. Yes. 21 A. And that was what I was flagging up here at this point. 22 Q. Let's look at that, then, if you want to look at it. 23 I think you accurately summarised it. Keep going. 24 There's an email from Ms Jennings, saying that Patel is 25 up for trial at Peterborough Crown Court, yes? 171 1 A. Yes. 2 Q. Then you reply further up, saying: 3 "[I don't know anything] I'm not aware of this case 4 ..." 5 Yes? 6 A. Yeah, yeah. 7 Q. Then she gives that explanation: 8 "This is the one you supplied the expert report 9 for", et cetera. 10 A. Yeah. 11 Q. Then further up, please, which is where I want to look 12 here, your reply: 13 "Thanks for the clarification. I had not understood 14 that that related to a specific case." 15 There, the "that that related to a specific case" -- 16 A. By that I meant the witness statements. 17 Q. -- ie, the generic witness statements? 18 A. The generic witness statements, yes. 19 Q. So she was saying you supplied a witness statement for 20 the case of Patel and you're saying, "Well, hold on, I'd 21 understood that my generic witness statement was not in 22 relation to a specific case". 23 A. Correct. That was my understanding. 24 Q. Continuing: 25 "... I thought that was a general statement. If 172 1 I am required to go to court for that, I think I need to 2 have some more background on the specific case and 3 exactly what is [required]. I appreciate that is not 4 covered by my statement, but if I need to be an expert 5 witness, I need to understand what is happening." 6 By this time, had you realised that you were 7 performing the function of an expert witness? 8 A. Yes, but without understanding what that really meant. 9 Q. What did you understand it to mean? 10 A. Someone who had good knowledge of how Horizon operated. 11 Q. This is a slightly different formulation from what we've 12 seen before, where you are referring to yourself as 13 an expert witness, isn't it? 14 A. Yes, because -- well, I knew people had been referring 15 to me as an expert witness in the past and I now realise 16 that I just didn't know what an expert witness actually 17 meant. But I assumed that what I'd -- the function I'd 18 performed in the past was what was required of me again. 19 Q. By this time, were you aware that an expert witness had 20 a special status in legal proceedings? 21 A. I was aware that there were some differences but 22 I didn't understand all the legal niceties involved in 23 such things. 24 Q. What was the extent of your understanding of the "some 25 differences"? 173 1 A. Well, for example, I think I was aware that I was 2 allowed to sit in court and listen to other witnesses, 3 which I understand a witness of fact was not allowed to 4 do. 5 Q. From whom did you gain that understanding? 6 A. Because I was asked to do so during Mrs Misra's trial. 7 Q. Who explained that the reason you were allowed to sit in 8 was -- 9 A. I can't remember -- 10 Q. -- because you were being treated as an expert witness? 11 A. Yes, but that was all I knew that made it any different 12 from anyone else, but I didn't really understand the 13 legal niceties of it, I'm afraid. 14 Q. Who explained that to you? 15 A. I can't remember. Possibly Mr Tatford, I'm not sure; or 16 it could be Mr Longman. I don't know. I don't think 17 Jarnail Singh was actually around very much during the 18 trial. 19 Q. By this time, 19 October, did you understand that 20 an expert witness had particular responsibilities? 21 A. No, I did not. 22 Q. Did you draw a link between describing yourself as 23 an expert witness and the line that had been inserted in 24 your generic witness statement that you had a duty to 25 the court? 174 1 A. No, I didn't realise there was any link between them. 2 I do now but I didn't then. 3 Q. Did you, when the line was inserted in your witness 4 statement, with you saying to the defence and to the 5 court that you understood that you owed a duty to the 6 court, ask anyone, "What duty do I owe to the court?" 7 A. I just assumed it meant to tell the truth, which is what 8 I'd do anyway. 9 Q. Did you question why somebody was inserting something 10 into your witness statement that was very obvious to you 11 already? 12 A. No, I didn't. 13 Q. Did you ever ask, "What are these duties that I owe"? 14 A. No, I did not. 15 Q. If you were completely oblivious to the status and 16 responsibilities of an expert witness, why were you 17 describing yourself as one? 18 A. Because that's what other people were describing me as. 19 SIR WYN WILLIAMS: In Mrs Misra's trial, were you present 20 throughout the whole time? 21 A. No, and, in particular, I was only present during the 22 first week. I never heard the evidence that she gave, 23 and the first time I saw that was about two or three 24 years ago, when preparing for -- 25 SIR WYN WILLIAMS: Yes. So it follows from that that you 175 1 weren't present when the judge summed up to the jury? 2 A. No, because that was the second week. I'd gone home 3 after -- 4 SIR WYN WILLIAMS: That's all right. 5 MR BEER: You were expressing some concern, in this chain, 6 that your generic report was being used, was being 7 deployed, in a specific case? 8 A. Yes. 9 Q. Can we turn forward, please, to POL00097137 -- can we 10 scroll down, please, and again, please -- an email from 11 Rachael Panter to you, under the heading "Horizon expert 12 report": 13 "As you may already be aware, your expert report 14 detailing the reliability of the Horizon system ..." 15 First of all, did you understand that you had 16 produced an expert report? 17 A. I had produced a report which then got converted into 18 a generic witness statement. 19 Q. Did you understand that it was the statement of somebody 20 who was being treated as an expert witness, or did you 21 still regard yourself as a witness of fact, or didn't 22 you make the distinction at all? 23 A. I didn't make the distinction. As I say, the first time 24 I became aware of the duties of an expert witness was in 25 about 2020/2021, something like that. 176 1 Q. "... your expert report detailing the reliability of the 2 Horizon system ..." 3 Did you understand that the document you produced 4 was a document which detailed the reliability of the 5 Horizon system? 6 A. I took that as a description of the document I'd 7 produced, which I thought was more of a description of 8 the reliability of the audit data that came out of the 9 Horizon system, but I didn't quibble about the 10 difference between those two. 11 Q. She continues: 12 "... has been served as evidence in a number of Post 13 Office cases that are at various stages of the court 14 process, most of which are listed for trial in the early 15 part of next year. 16 "It should be noted that most, if not all cases 17 raising the Horizon system as an issue, have been 18 unable/not willing to particularise what specific issues 19 they may have with the system, and how that shapes the 20 nature of their defence. 21 "As we already have your detailed report, I would 22 like to serve it in each case below ..." 23 Then six cases are set out: 24 "Your expert report has already been served in the 25 Nemesh Patel case. I would like to serve your report in 177 1 the remaining cases, and have attached a case summary of 2 each case listed above so that you may familiarise 3 yourself with the facts of each case. 4 "... Please could you read the case summaries 5 attached and send 5 original signed and dated copies of 6 your report to me as soon as possible." 7 Did you understand what you were being asked to do 8 here? 9 A. Not really. I think I was just being asked to actually 10 send her five copies -- signed copies of the statement 11 and -- as I responded, saying, "Well, why can't you use 12 the one you've already got?" 13 Q. This was a month or so after your pushback on the use of 14 the generic statement in a single case -- 15 A. Yes. 16 Q. -- saying that you hadn't seen the underlying data. 17 A. No. 18 Q. You're now being told that it's being used in six cases. 19 In fact, it had already been served in one. 20 A. I think that one was the one that I was already aware 21 of -- 22 Q. The same? 23 A. -- yes. 24 Q. What was your reaction? Did you not think, "Well, hold 25 on, I'm similarly unhappy that my generic statement is 178 1 being deployed in individual cases when I've seen none 2 of the underlying data in any of those cases"? 3 A. I think it was -- I sort of took that as read, having 4 already made that point, admittedly to someone else, and 5 the pushback was really, "Well, why don't you just get 6 on with it? You've already got the signed statement, 7 why do I need to be involved?", because I'd been told 8 before that I didn't need to be involved. And I also 9 sought some guidance from management as to the fact that 10 this had been sprung on me out of the blue. 11 Q. Can we scroll up, please. We see your reply: 12 "Can't you use the report I have already sent to 13 you? There is no mention of the case on the report." 14 By that, are you meaning that you could just serve 15 it times six -- 16 A. Yes. 17 Q. -- because it's not a case-specific report; there's no 18 defendant name in it? 19 A. Correct. 20 Q. Then you say: 21 "You need to be addressing such requests through 22 Post Office rather than to myself." 23 Then: 24 "... there is no commercial cover in place", ie that 25 there isn't any allocated spend to this? 179 1 A. Yes. 2 Q. Is that right? 3 A. Yes, that's right. And I then copied the email to the 4 people who should have been involved in sorting out 5 those commercial niceties, and so on. 6 Q. Penny Thomas and James Davidson? 7 A. Yes, and also Jane Owen, who I knew was Penny's 8 counterpart in Post Office. 9 Q. So at this point, what was your state of mind, "I've 10 produced this expert report, as it's described, which is 11 generic; I've been told that it's been deployed in the 12 Patel case; I've pushed back against that, saying I've 13 seen none of the underlying data; and then I'm told it's 14 being deployed in five other cases, can I provide some 15 more copies of it"? 16 A. I was even more confused because it implied that it had 17 already been employed in an unspecified number of cases, 18 so -- in which case, why was I being asked to sign it 19 five times? So I was just generally confused. 20 Q. What about the more fundamental issue that you'd raised 21 beforehand about the appropriateness of using a generic 22 statement without looking at underlying data? 23 A. I thought I'd already made that point and didn't feel 24 the need to repeat it at this point. 25 Q. Thank you. That is all I want to ask you about the use 180 1 of the generic statement. 2 Just one last topic before we conclude my questions, 3 and it concerns the decision not to continue to use you 4 as a witness in support of Post Office prosecutions. 5 A. Okay. 6 Q. Can you recall what you were told about that? 7 A. I can't remember the exact detail. I've covered it in 8 my witness statement but, basically, it was along the 9 lines that I was been told that I couldn't be used any 10 more for some sort of legal reasons, and it was never 11 explained to me what those reasons were. I now 12 understand, I think, a bit more of the background but 13 I didn't find that out until 2020/2021. 14 Q. I'm going to come to the email in a moment. 15 A. Yeah. 16 Q. Were there any conversations with you explaining why the 17 Post Office didn't want to use you as an expert witness 18 in legal proceedings against subpostmasters in the 19 future? 20 A. Not that I can remember, other than the email which 21 I suspect is the one you're about to show me. 22 Q. Yes. Can we look, then, at the email, FUJ00156923, and 23 look at the email at the bottom of the page. We're 24 a year and a bit on now, December 2013. 25 A. I think there had been some emails between myself and 181 1 Rachael Panter over the intervening months saying, 2 "Where are we in terms of are there any more cases 3 I need to worry about, and diary management?" And I'd 4 effectively been fobbed off on those. 5 Q. Exactly so -- 6 A. But there was nothing to say, "We can't use you 7 because"; it was just a case of "nothing for you to do". 8 Q. I haven't gone to those because they don't say anything 9 of substance and whether or not you're to be used and, 10 if not, why not. 11 A. Yeah, yeah. 12 Q. 3 December 2013, James Davidson to a wide group of 13 people, including you -- 14 A. Yes. 15 Q. -- with the subject "Horizon Integrity Challenges": 16 "Folks, 17 "Mike, Pete and myself ..." 18 Just decoding who that is. 19 A. I assume that is Michael Harvey. 20 Q. And Pete Newsome? 21 A. Pete Newsome, I would think, yes, as opposed to Peter 22 Thompson. 23 Q. Remind us, Michael Harvey? 24 A. He was a lawyer. Well, I think he was a commercial 25 manager but I believe he was also a qualified lawyer. 182 1 But I think his role at this point was as a commercial 2 manager. 3 Q. And Pete Newsome? 4 A. He was a liaison between Fujitsu and Post Office, and 5 particularly involved with the prosecution side of 6 things. 7 Q. "... met with Post Office lawyers last Friday to discuss 8 the proposal Post Office have come forward with for 9 an Independent External Expert review of the integrity 10 of the Horizon system. In essence, the problem that 11 Post Office have is that they are being given legal 12 advice that the 'rules of evidence' mean that 13 submissions which in the past have sufficed to support 14 prosecutions (namely the input from Gareth Jenkins), 15 cannot be used in future prosecutions. They are 16 therefore seeking the view of an external expert who can 17 warrant that the data they are basing prosecutions upon 18 has integrity. 19 "In the meeting, we were successful in moving the 20 discussion away from a full 'system' review (which would 21 not achieve their aims and is unworkable in practice) to 22 focus it on what we are calling the 'Core Audit 23 Process'. This scope is aligned to the audit we were 24 intending to do in 2012 and for which we had already 25 agreed a terms of reference with KPMG." 183 1 That was the paper for which you had written some 2 terms of reference? 3 A. Indeed. 4 Q. "The attached slides describe the Core Audit Process and 5 sets out a terms of reference for an audit that is 6 achievable and that we could support. We have made it 7 clear to Post Office that whilst we will cooperate and 8 provide access to our experts (in a controlled way) we 9 reserve the right to challenge any findings of 10 an external expert. The support will also be chargeable 11 and a CT ..." 12 What's that? 13 A. Commercial Terms. So it's basically a contract between 14 Fujitsu and the Post Office. 15 Q. "... is being drafted by myself in parallel. All 16 content has been reviewed by Mike, Pete, Gareth." 17 Scroll down: 18 "Can you review the terms of reference?" 19 I'm not going to read the rest of it. 20 A. Yes. 21 Q. The first paragraph of that email explains that the 22 problem that the Post Office has is that they are being 23 given legal advice that the rules of evidence mean that 24 the input from you cannot be used in future 25 prosecutions? 184 1 A. And I took it at face value as that, and I thought that 2 meant okay, I was off the hook and didn't need to get 3 involved in any more legal proceedings. 4 Q. Were you content with that? 5 A. Oh, yes. 6 Q. You were happy not to be involved? 7 A. Yes, because I preferred to be working on my day job, 8 which was designing changes to the system, rather than 9 being involved in prosecutions. 10 Q. Was that the extent of what you were told as to why you 11 were no longer the Post Office's expert witness of 12 choice in their prosecutions? 13 A. As far as I can remember, yes. 14 Q. Did they ever tell you about a legal advice that they 15 had received which concluded, in broad terms, that you 16 had broken your duties to the court and had given 17 evidence which was unreliable and, therefore, you could 18 not be relied on for those reasons? 19 A. No, I didn't understand any of that until 2020. 20 MR BEER: Thank you. 21 Sir, they are the only questions that I ask. 22 SIR WYN WILLIAMS: Thank you. 23 Just on this, the only persons who communicated with 24 you in this email, they are your fellow employees of 25 Fujitsu? 185 1 A. Well, they're more senior than me -- 2 SIR WYN WILLIAMS: More senior, but Fujitsu people? 3 A. Yes. 4 SIR WYN WILLIAMS: So the explanation to you came from your 5 own employers, in effect? 6 A. Yes, I'm just checking that. As far as -- yes, all 7 those people there, I think, were Fujitsu people. 8 SIR WYN WILLIAMS: Right. Did you ever have any direct 9 reason from Post Office employees as to why you were 10 going to not be a prosecution witness any more? 11 A. Not that I can recall. 12 SIR WYN WILLIAMS: Right. Fine. Thank you. 13 MR BEER: Thank you very much indeed, sir. They are the 14 questions that I ask. I think we're going to start with 15 Core Participants tomorrow at 9.45. 16 SIR WYN WILLIAMS: All right. That's it for the day. 17 (3.57 pm) 18 (The hearing adjourned until 9.45 am the following day) 19 20 21 22 23 24 25 186 I N D E X GARETH IDRIS JENKINS (continued) ..............1 Questioned by MR BEER (continued) .............1 187