1 Wednesday, 26 June 2024 2 (10.03 am) 3 GARETH IDRIS JENKINS (continued) 4 Questioned by MR BEER (continued) 5 SIR WYN WILLIAMS: Yes, Mr Beer. 6 MR BEER: Thank you. Good morning, Mr Jenkins. 7 A. Good morning, sir. 8 Q. Before we carry on looking at bugs errors and defects, 9 which is where we left off yesterday, I'd like, if 10 I may, to go back over an issue that we addressed 11 yesterday which is whether you saw the Bond Pearce 12 letter of 18 November 2005. 13 A. Okay. 14 Q. You remember that letter from -- 15 A. I remember discussing it yesterday, yes. 16 Q. Yes, from Bond Pearce to Fujitsu. It was the one that 17 asked Fujitsu to prepare an expert report, which asked 18 Fujitsu to address in the report the issues raised by 19 Lee Castleton in his defence and which explained 20 an expert's duties and the necessary inclusions in 21 an expert report; do you remember? 22 A. I remember discussing it yesterday, yes. 23 Q. I showed you a copy of that letter yesterday. Can we 24 have that on screen, please, FUJ00152573. If we scroll 25 through, please, you remember the way this pack of 1 1 documents was constructed and, if we carry on 2 scrolling -- and keep scrolling -- to the letter of 3 18 November 2005 from Bond Pearce to Fujitsu. I showed 4 you a copy of this letter yesterday; do you remember? 5 A. I remember that, yes. 6 Q. Thank you. When I asked you about the letter and if it 7 was sent to you, you said no, you had got no 8 recollection of that. You remember that? 9 A. I certainly -- that is certainly what I said yesterday. 10 Q. Yes. You said yesterday that, if the letter had been 11 sent to you, then you would have remembered it because 12 it clearly sets out what duties there are, and you 13 weren't aware of them until the end of 2020? 14 A. I believe I must have been mistaken in what I said 15 yesterday, in that respect. 16 Q. Before we come to whether you were mistaken or not, I'm 17 just going to recap what you said. 18 A. I accept that that is what I said yesterday. 19 Q. So you said that you would have clearly remembered it 20 because it sets out the duties of an expert and you 21 weren't aware of them until the end of 2020, yes? 22 A. Correct. 23 Q. You said yesterday that you did not see the letter 24 because, if you had seen it, then you would have learned 25 about the existence of an expert witness's duties? 2 1 A. That is what I said yesterday, yes. 2 Q. You agreed yesterday that Bond Pearce had set out in 3 this letter very clearly, and in an easy-to-understand 4 way, the duties of an expert witness? 5 A. I accept that's what I said, yes. 6 Q. Lastly, you said that you were sure that the letter 7 never made its way to you because you would, and I quote 8 here: 9 "... have done things differently, not necessarily 10 in this case [the Lee Castleton case] but certainly in 11 later cases, if I'd been aware of those 12 responsibilities." 13 You remember saying that? 14 A. Yes. 15 Q. I also asked you yesterday about your attendance 16 a conference with lawyers at Fujitsu's headquarters on 17 6 June 2006; do you remember? 18 A. I do. 19 Q. You said, in summary, that there was no discussion about 20 giving expert evidence or the duties of an expert at 21 that conference? 22 A. And I've no recollection of that, no. 23 Q. Indeed, that's reflected in the attendance note of that 24 conference: there appears to have been no such 25 discussion? 3 1 A. I agree. 2 Q. Can we turn to your witness statement, please. Your 3 third witness statement, at paragraphs 251 and 252, 4 which is page 86. I'm going to read those in full: 5 "251. The Inquiry's Rule 9(2) Request has asked me 6 how I first came to be involved in Mr Castleton's case. 7 It has also asked what I understood Stephen Dilley to be 8 asking ... to provide by way of his letter dated 9 18 November 2005." 10 Then you give that URN that we've just looked at, 11 okay? 12 A. Yes. 13 Q. "I don't believe that I had ever seen this letter until 14 it was shown to me for the purposes of preparing this 15 witness statement. It appears that at some point Brian 16 Pinder emailed me and asked me to address a single 17 question about human error (which I can now see came 18 from the list of questions asked in the letter) and that 19 my response to this single question seems to have been 20 my first involvement in Mr Castleton's case." 21 We looked over all of that yesterday. 22 A. Yes. 23 Q. "252. I don't think that I had any further involvement 24 until, over six months later, I was invited to attend 25 a meeting with [the Post Office's] solicitors on 6 June 4 1 2006. My lawyers have shown me an email from Brian 2 Pinder to me dated 5 June 2006", and you give the 3 reference number ending in 601. 4 A. Yes. 5 Q. "This email attached an agenda for the next day's 6 meeting and a scanned extract from a document", and you 7 give a reference number 602. 8 A. Yes. 9 Q. "... which recorded issues Mr Castleton had experienced 10 with Horizon. I probably read both attachments in 11 advance of the next day's meeting although I don't 12 recall doing so. Brian Pinder's email stated that he 13 was 'not sure how this affects ourselves but reading the 14 scanned document attached (his queries re Horizon) we 15 may well require [your] expertise in this'. Looking 16 back at this email, I cannot recall what my 17 understanding of the purpose of the next day's meeting 18 was but I would probably have assumed that I would be 19 asked to respond to the Horizon issues identified by 20 Mr Castleton. However, I don't think I knew at this 21 stage that [the Post Office] intended me to become 22 a witness in any legal proceedings." 23 Now, you say, if we go back to 252, at the foot of 24 the page, "My lawyers have shown me an email", can you 25 see that in the second line into the third line? 5 1 A. Yes. 2 Q. "Dated 5 June", and you give that the number ending 601. 3 So that's the day before the conference. 4 A. Yes. 5 Q. If we look at that, please, FUJ00152601. 6 Look at the bottom part of the email first. 3.49 7 from Mr Dilley to Brian Pinder and Graham Ward and 8 others, not you at this stage: 9 "Dear Brian and Graham, 10 "I attach a suggested agenda for tomorrow's 11.30 am 11 meeting, just for discussion purposes. I also attach 12 an extract from Mr Castleton's Part 18 Replies setting 13 out his allegations about Horizon. I think Mared may 14 also add in a few issues." 15 So that suggests that the things attached were an 16 agenda and an extract from some part 18 replies. Okay. 17 A. And that was the assumption I made when making my 18 witness statement. 19 Q. So ie you read what was in this email and effectively 20 translated it into your witness statement? 21 A. Correct. 22 Q. Thank you. Then if we go up, please, that's forwarded 23 to you. There must be a time difference because the 24 time given is 40 minutes before the earlier email is 25 sent. 6 1 A. It says UTC time, so that's June, summertime. 2 Q. So that would have been 4.19? 3 A. Yeah. 4 Q. "Please be advised in the email below [this is directly 5 to you] and the attachments." 6 Just stopping there, we can see the attachments, one 7 is called "letter_doc" and the second attachment is 8 an eCopy of a scanned copy document.pdf; can you see 9 that? 10 A. Yes, I see that. 11 Q. "Not sure how this affects ourselves", and then there's 12 the sentence that you set out in your witness statement; 13 can you see that? 14 So, again, essentially two documents, it is 15 suggested are attached, an agenda and a scanned copy of 16 Mr Castleton's Part 18 replies. 17 Now, in you witness statement you said that the 18 email attached an agenda and a scanned extract from 19 a document, and you gave that the reference number 602. 20 Can we look at that, please. FUJ00152602. You'll see 21 these are statements of case, so legal documents which 22 used to be called pleadings, in which the parties 23 exchanged information about their case. 24 A. I didn't understand that at the time but I understood 25 that it was asking technical questions that I thought 7 1 I would need to address. 2 Q. Okay, so that's not relevant to us at the moment but in 3 your witness statement that's the only document you 4 refer to as being attached -- 5 A. I agree. 6 Q. -- and that's not relevant. Now, it's true to say that 7 there was an agenda for that meeting, for 6 June 2006, 8 but it wasn't attached to this email. The reference 9 number, we needn't turn it up, for the agenda is 10 POL00071418. 11 Now, can we please switch to the native version of 12 the email, FUJ00152601, and it will be necessary to drop 13 the YouTube feed for this. You understand, this is the 14 original version, essentially, of the email that 15 I showed you yesterday in an imaged form. 16 A. Yes, I understand that. 17 Q. So you'll see it's exactly the same email, the text is 18 exactly the same in it, but the difference is we can 19 see, actually, what the two attachments were? 20 A. Yes, I understand that. 21 Q. If we look at the second attachment first, if we can 22 just open that on the screen. We'll see that's the 23 Part 18 replies. 24 A. I understand, yes. 25 Q. So that can be closed, thank you. Then if we open the 8 1 first attachment, we can see that, in fact, the first 2 attachment of the email to you was the letter of 3 18 November 2005; can you see that? 4 A. I can see that and I now understand that to be the case. 5 Q. Now, when you made your witness statement, you would not 6 have been able to refer to this document, the 7 18 November 2005, as being an attachment to the email 8 because it hadn't been disclosed by the Inquiry on its 9 eDisclosure platform in a way that associated it with 10 the covering email; do you understand that? 11 A. I accept that, yes. 12 Q. Thank you. The association between the email and this 13 attachment, the letter, was only realised yesterday. So 14 I'm obviously not going to ask you why you said what you 15 did in your witness statement about the other attachment 16 being an agenda, when it in fact wasn't. 17 A. Yes. 18 Q. Okay? Nor am I going to ask you why you didn't refer to 19 this letter of 18 November 2005 in your witness 20 statement as being provided as an attachment to the 21 email because you didn't have the evidence then before 22 you to make that connection? 23 A. Correct. 24 Q. Understood? 25 A. Yeah. 9 1 Q. Can we go back, please, to the imaged version of this 2 letter, FUJ00152573. I think it was page 8 or 9 -- 3 maybe 10. Just scroll on, please. Thank you. I'm just 4 going to use this imaged version rather than that native 5 version. The text is exactly the same. So I'm not 6 going to ask you about why your witness statement was 7 written in the way that it was. The plain facts are, 8 I think you'll agree, however, that we can now see that 9 the letter of 18 November 2005 was sent to you? 10 A. I understand that and I accept that, and I apologise for 11 any mistake I made yesterday. 12 Q. It was sent directly to you and, therefore, I've got ask 13 to ask you some questions about that, okay? 14 A. I understand that, yes. 15 Q. So you accept, firstly, the simple fact that, in the 16 light of the evidence as it now stands, this Bond Pearce 17 letter was sent directly to you? 18 A. Yes, clearly. 19 Q. Do you remain of the view that you expressed yesterday 20 that, if the letter had been sent to you, you would have 21 read all of it and not just half of it? 22 A. I would probably have skimmed through it but, clearly, 23 it hadn't stuck in my mind beyond the time that I read 24 it because, at the time, I was concentrating of the fact 25 I was being invited to a meeting the following day and 10 1 the substance of what I needed to look at for the 2 following day was the other attachment, not this one. 3 So, yes, I would probably have opened it up and read it 4 but I have no recollection of doing that. 5 Q. Why do you think your focus of attention would have been 6 the Part 18 replies and not this letter? 7 A. Because what I was being asked to do was answer 8 questions. If we go back to the covering email from -- 9 Q. FUJ00152601. Is it the second paragraph there you're 10 referring to? 11 A. "Not sure how this affects ourselves but reading the 12 scanned document attached (his queries re Horizon) we 13 may well require you're [sic] expertise on this." 14 Are you saying that, because of that sentence, you 15 would have focused only on the scanned document and not 16 on the letter? 17 A. Yes. 18 Q. Really? 19 A. Yes. That was the one that actually had technical 20 information that I felt I could contribute towards. 21 Q. Why did you tell me yesterday that, if you had been sent 22 this letter, you would have read all of it and not half 23 of it? 24 A. I probably would have read all of it but I probably 25 didn't read it in detail and having -- and, obviously, 11 1 it didn't stick in my mind, and sorry if I misled you 2 yesterday to suggest that it would have stuck. I think 3 it's a problem of understanding the difference between 4 it clearly strikes me, reading it now, as being 5 important, whilst at the time, it probably didn't strike 6 me as being important. 7 What I was being asked to do was to go to a meeting 8 to talk about technical issues with Horizon and so 9 I would have concentrated on the technical scanned 10 document, not the letter talking about putting together 11 the report and, at that stage, I didn't think I was 12 being asked to put together a report and, in fact, 13 I don't believe I was ever asked to produce a report in 14 that sort of form, for this case. 15 Q. Or is it that you said yesterday that, if you'd received 16 the letter, you would have read all of it and not just 17 half of it, because it was acceptable to say that when 18 you didn't think you'd been sent the letter? 19 A. I don't know. As I say, I would probably have read the 20 whole thing. I'm not suggesting I didn't read the whole 21 thing but, clearly, it didn't stick in my mind. 22 Q. Okay, thank you. Can we go back to the letter, please. 23 FUJ00152573, and I think page 11 -- 12 -- 13, thank you. 24 Do you remain of the view that you expressed 25 yesterday that the letter sets out, in a detailed but 12 1 easy-to-understand way, the duties that would need to be 2 complied with in the case of a Fujitsu employee 3 providing evidence in at least the Lee Castleton case 4 about the matters raised in the letter? 5 A. Yes, I have to accept that. 6 Q. Do you accept now, in the light of the evidence, that 7 you were sent a copy of this letter, that you were 8 therefore informed, in a detailed but easy-to-understand 9 way, the duties that would need to be complied with in 10 the event that you provided evidence in the Lee 11 Castleton case in answer to this letter? 12 A. Yes, but I don't believe I was actually asked to provide 13 that sort of report in the Lee Castleton case. 14 Q. Well, I'll come to that a little later. 15 A. But yes. 16 Q. If we scan through the letter just slowly -- I think you 17 will have read it carefully overnight -- the letter 18 doesn't refer, would you accept, to the provider of 19 a report from Fujitsu as being an expert witness? 20 A. I'm not sure that I'd have spotted that in terms of -- 21 I was concentrating more at looking at what else 22 happened in the case, rather than the contents of this 23 letter, last night. 24 Q. Look at it carefully now. If we just go back to the 25 top, thank you. If you just read it to yourself slowly, 13 1 and just indicate when you've read that which is on the 2 screen. 3 A. Yes, I've read what's on the screen. 4 Q. Thank you. If we go down, please. 5 A. Yes, I've read that page. 6 Q. Thank you. Over the page, thank you. 7 A. Yes, I accept that it doesn't say that the expert's 8 report is a formal report. 9 Q. Thank you. Maybe you can just read the first two 10 paragraphs under (4), and then go over the page, and 11 then read after 14 the next three paragraphs. 12 A. Yes, I've read those three paragraphs. 13 Q. Then just read under (5). 14 A. Yes, I've read that. 15 Q. So the letter does not refer to the provider of the 16 report that was being sought by the letter by Fujitsu, 17 at once, as being an expert witness, does it? 18 A. No, it doesn't. 19 Q. Instead, the letter, would you agree, works on the basis 20 that any person providing evidence from Fujitsu about 21 the matters mentioned in the letter needs to comply with 22 the duties set out in the letter? 23 A. I understand that now but I almost certainly would not 24 have realised that at the time. 25 Q. You said yesterday that you were sure that the letter 14 1 never made its way to you, because: 2 "[You] would have done things differently, not 3 necessarily in this case, [the Lee Castleton case] but 4 certainly in later cases." 5 A. If I'd understood what that was all about. Now, 6 I clearly had no recollection that I'd received it at 7 this point. You've now shown that I did receive it at 8 this point but I still have no recollection of 9 remembering anything about that at any later stage. 10 Q. But you said yesterday that, had you received the 11 letter, you would have done things differently, 12 certainly in later cases, if you'd been aware of the 13 responsibilities set out in this letter. 14 A. Maybe I need to qualify that to say, if I'd received and 15 understood the detail of the letter. 16 Q. Well, that's not what you said. 17 A. I accept that and I was mistaken. 18 Q. Why were you mistaken that you think that yesterday you 19 thought, had you received the letter, you would have 20 certainly done things differently in later cases; today 21 you don't think, now you know you received the letter, 22 that you would have done things differently in later 23 cases? 24 A. I think the difference is in terms of understanding what 25 it was all about. What I was thinking of yesterday was, 15 1 with my knowledge now, of what -- of the implications of 2 what it actually says in that letter, then I would have 3 done things differently. If I'd just skimmed through it 4 and not taken too much -- made too much attention of 5 what it actually said and it had then gone from my mind 6 when I was busy concentrating on the other attachment, 7 then clearly I had not remembered anything about it much 8 beyond those couple of days. 9 Q. What were the later cases that you had in mind yesterday 10 when you told us that, if you'd seen this letter, if 11 you'd received this letter, you would have done things 12 differently? 13 A. If I'd understood my duties as an expert, and I accept 14 they're laid out here, and understood at the time, then 15 I would clearly have approached the reports that 16 I produced for Seema Misra's case in a different way. 17 And, as I've said before, I had no recollection of 18 having been briefed in this way and, after all, that was 19 about three or four years later and I'd clearly 20 forgotten all about having seen this letter briefly at 21 that time. 22 Q. So when you said yesterday, had you seen this letter, 23 you would have done things differently in later cases, 24 you had in mind Seema Misra? 25 A. Yes, I think so that was -- and, clearly, there was 16 1 a later case as well but that was the next one that came 2 along. 3 Q. What about the Hughie Thomas case? 4 A. I think that had happened letter. 5 Q. You'd made a witness statement earlier in the year in 6 March or April? 7 A. But I didn't think that that was an expert report at 8 all. That was just looking at one specific question of 9 fact, which was whether -- what was the reason for zero 10 value transactions. I didn't -- I had no concept of the 11 fact that that was an expert report in any sort of way. 12 Q. I should just say, whilst it's on the screen there, just 13 look under (5), can you see that? 14 A. Yes. 15 Q. The words "expert witnesses" is included; can you see 16 that? 17 A. Yes, yes. 18 Q. "In performing all your duties ..." 19 A. Yes. 20 Q. So it's not quite right what I suggested to you earlier 21 that "expert witness" or "expert witnesses" are not 22 mentioned in the letter. 23 A. No. 24 Q. They are in that paragraph there. But do you still 25 agree that, reading the letter now, the requirements 17 1 imposed by or suggested by Bond Pearce don't depend upon 2 the person giving the report or making the report being 3 classified as an expert witness? 4 A. I'd not really taken that into account. So I think it's 5 very difficult to talk about hypotheticals in that sort 6 of way. 7 Q. So you were told about these responsibilities because 8 you'd been provided with this letter. What's the 9 explanation now, then, for not doing things differently 10 in later cases? 11 A. Because I'd clearly not remembered about the fact that 12 I'd seen this letter in 2006. 13 Q. Your evidence in your witness statement and in your oral 14 evidence yesterday was founded on the suggestion that 15 you didn't need to comply with all of the duties set out 16 at length by, for example, Professor McLachlan in his 17 report, nor Mark Turner in his report because they were 18 independent experts and you were from Fujitsu. 19 A. Yes, and I stand by that. 20 Q. Doesn't you receiving this letter rather undermine that 21 point? 22 A. It could do if I'd remembered that I'd received this 23 letter. But I -- I-- where I'm coming from is that 24 I know that I acted honestly at all times and, 25 therefore, I'm -- the only way that I can explain that 18 1 is by the fact that I just had no recollection of this 2 letter. 3 Q. If we go to the top of the previous page, please. 4 There. Thank you. The six topics set out there are 5 just asking questions about Horizon, aren't they? 6 A. Yes. 7 Q. The advice from Bond Pearce was that the person that 8 answers them in a report needs to comply with everything 9 that follows in the letter? 10 A. Yes, but I don't think I was being asked to produce 11 a report at that time. 12 Q. Can we look at then what happened subsequently, in 13 fairness to you, at FUJ00122283. It might be that's 14 an errant reference -- ah no, perfect. 15 Can we just look at the email at the foot of the 16 page, please, and then go up, please. You say in this 17 email, and we're now in the beginning part of August 18 2006, so a couple of months after the conference at 19 Fujitsu? 20 A. Yes. 21 Q. "I've had a go through this. It isn't a complete 22 statement and I won't have time to sort out the 23 outstanding bits for a while ... 24 "I'm happy for you to send it as it is to Stephen 25 ..." 19 1 Who was the Stephen you had in mind? 2 A. Stephen Dilley. 3 Q. "... you may decide it isn't complete enough for that 4 ... I may be able to research some of the areas (such as 5 Auditing) that I'm claiming ignorance of if that is 6 required." 7 You will see there is an attachment, which is 8 a statement? 9 A. Yes. 10 Q. Can we look at that statement, same reference, but 11 ending 84 [FUJ00122284]. This was the attachment to 12 your email and was a draft witness statement, is this 13 right, that you compiled? 14 A. The draft witness statement was compiled by Stephen 15 Dilley and he'd left various questions in it and, as 16 I explain on that first page, I made various annotations 17 and we're back to the problem about being able to tell 18 the difference between colours in a black and white 19 document. 20 Q. I see. So the format of the witness statement was 21 compiled by Mr Dilley and some of the formal parts of it 22 were compiled by Mr Dilley. He included some text in it 23 and he asked you some questions that you were to 24 responding to and fill in? 25 A. Correct. 20 1 Q. And you did that in this version of it? 2 A. Yes. 3 Q. If we just scroll down, please, and look at paragraphs 1 4 and 2, and then scroll on, you address the Horizon 5 system which you'll remember is question 1 in the letter 6 of 18 November 2005: 7 "What's the Horizon system? We need to explain to 8 a judge how it works." 9 A. I don't think I would have been thinking of it in the 10 context of that letter, at the time I saw this, because 11 this was clearly a couple of months later. 12 Q. Right, okay. Then, if we just scroll on, we can scan 13 it. The detail is not important for the moment. If we 14 scroll on, and keep scrolling, you can see some examples 15 there of him asking questions and you replying? 16 A. Yes. 17 Q. Carry on. 18 A. And there's also examples of him asking questions and me 19 saying I don't think I'm able to reply. 20 Q. Yes. If we scroll on, and then stop there. You'll see 21 that it was proposed that you signed off with a standard 22 declaration: 23 "I believe that the facts stated in this witness 24 statement are true." 25 Can you see that? 21 1 A. Yes. 2 Q. So this draft provided by him and further expanded by 3 you didn't contain any of the declarations or the 4 matters mentioned by Mr Dilley in his letter of 5 18 November 2005 as an expert witness? 6 A. Correct. 7 Q. Did you give any thought to that at the time? 8 A. No, is the simple answer. I was given this document, 9 which was supposedly a draft of something that I should 10 be saying, and what I did was I commented on that as to 11 whether I could truthfully say what was in that draft or 12 not and corrected any factual errors that I saw within 13 there. I gave no thought to what format it was in. 14 Q. So, by that time, had the letter of 18 November been 15 lost by you in your memory? 16 A. I believe so, yes. 17 Q. Was there any suggestion by Mr Dilley that you needed to 18 comply with the duties set out in his letter of 19 18 November, or say that you had complied with the 20 duties set out in his letter of 18 November, for the 21 purposes of giving this evidence? 22 A. Not that I'm aware of and I don't believe there is 23 anything in the covering email that -- I'm not sure if 24 we went to there now. And I'm also aware of some emails 25 which I've seen as part of the Inquiry where he was 22 1 saying that I was to be treated as a witness of fact and 2 not an expert evidence but, obviously, I didn't see 3 those at the time. 4 Q. Thank you very much. 5 SIR WYN WILLIAMS: Mr Beer, can I take it, because I've only 6 seen this as we're going along, that, in fact, it 7 contains a mixture of factual content and opinion? 8 MR BEER: Very much. 9 SIR WYN WILLIAMS: So there would have been a need for all 10 the necessary declarations? 11 MR BEER: Correct. 12 SIR WYN WILLIAMS: Fine. 13 MR BEER: It's not a witness statement of fact alone. You 14 can see that, for example -- 15 SIR WYN WILLIAMS: That's no reflection on you, Mr Jenkins. 16 A. I -- 17 SIR WYN WILLIAMS: I'm just trying to ascertain the true 18 legal status -- 19 A. I understand. 20 SIR WYN WILLIAMS: -- of what this would be. That's all. 21 A. Yes. 22 MR BEER: To round this topic off, Mr Jenkins, you didn't, 23 in fact, sign this witness statement and you weren't 24 required to give evidence in court? 25 A. Correct. And, in fact, I don't believe that this 23 1 version of it ever went back to Mr Dilley, and he 2 decided he didn't want a statement from me at all, 3 but -- and I think we worked through yesterday the 4 communication I had from Brian Pinder as to the fact 5 that it was no longer required. 6 Q. Thank you, that can go back down. 7 Right, can we go back to where we were? 8 SIR WYN WILLIAMS: Before we do, Mr Beer, I want to say this 9 openly, since obviously it is unfortunate that this 10 series of documents or these documents you've shown 11 Mr Jenkins this morning had not been disclosed earlier. 12 I want the Inquiry to make a further investigation, 13 just so I can be satisfied about what exactly occurred. 14 The letter of 18 November, as originally sent, as 15 I understand it, attached experts' reports obtained by 16 Mr Castleton. 17 MR BEER: Yes. 18 SIR WYN WILLIAMS: So I would like us to bottom out whether 19 those reports made their way, in some way or another, to 20 Mr Jenkins and/or were discussed at the conference which 21 he attended. 22 MR BEER: Thank you, sir. We can pursue that. 23 SIR WYN WILLIAMS: Yes. 24 MR BEER: Can we go back to bugs, errors and defects, 25 please, and turn to the Callendar Square bug, please. 24 1 A. Yes. 2 Q. You addressed this at significant length in your second 3 witness statement, it's paragraphs 53 to 82, where you 4 recount your knowledge of the Callendar Square bug. Can 5 we just look at a couple of parts of that, please. 6 Second witness statement at page 16, please, and 7 paragraph 54. You say: 8 "In the Technical Appendix to the Horizon Issues 9 judgment, Mr Justice Fraser grouped together number of 10 problems under a single heading called 'Callendar 11 Square/Falkirk'. I understand why the court and now the 12 Inquiry has taken this approach but it risks creating 13 the erroneous impression that the bug was homogeneous in 14 the sense that it was caused by the same factors, and 15 manifested itself in the same way, in every branch it 16 affected. This was not the case. As I explain below, 17 what happened at [the] Callendar Square branch in 18 Falkirk in 2005 was caused by an underlying bug in the 19 Riposte software that manifested itself when 20 a combination of unpredictable factors, including (in 21 that particular case) steps taken by the [subpostmaster] 22 to rectify the issue, came together. In other words, 23 there was an underlying software bug, but how it 24 manifested itself and whether it caused harm was 25 contingent on the presence of other circumstances. 25 1 These circumstances were not fixed. That is why the 2 problem was hard to reproduce." 3 Then page 28, please, and paragraph 80b, which is 4 beyond the there. You say: 5 "While many incidences of the locking errors were 6 not associated with any discrepancies in [branches] 7 (eg they resulted in operational problems such as screen 8 freeze), I became aware that it was not just the 9 Callendar Square branch that had suffered from 10 discrepancies but other branches where there were 11 counter occurrences of the timeout waiting for lock 12 problem." 13 Can we start with -- 14 A. Can we qualify that last thing because at the top of the 15 page it says that this comment arose from the work I did 16 in 2018/2019, and not at the time. 17 Q. Yes, sorry, this is later-discovered information. 18 A. Yes, yes. 19 Q. In what respect was the Callendar Square bug not 20 homogeneous? 21 A. I think it comes down to understanding exactly what 22 people mean by the Callendar Square bug. If we're 23 talking about the circumstances that actually occurred 24 in Callendar Square, then that was a storm of events 25 which came regularly at every ten seconds for a period 26 1 of time until the particular PC or server was rebooted. 2 But there were other cases where there was a single 3 event saying time-out waiting for lock, and what I was 4 getting at, in the first paragraph that you read to me, 5 is that, as far as the civil case, they were all 6 considered as being basically the same problem, and 7 I didn't feel that that was a correct representation of 8 how it should be taken. 9 Q. What was the nature of the other circumstances that were 10 required when interacting with the underlying Riposte 11 bug that produced lock errors? 12 A. The simple answer is I don't know. 13 Q. The reason why you don't know is? 14 A. It was in the Escher software and it was their propriety 15 software, they as I understand it, fixed the problem in 16 2006 and, certainly, the storm of events, which was what 17 was seen at Callendar Square, did not reoccur, as far as 18 I'm aware. 19 Q. But you didn't know then, and you don't know now -- and 20 even in 2018/19 didn't know -- what circumstances were 21 required in order, essentially, to trigger the existence 22 of the bug? 23 A. Correct. 24 Q. Can we look, please, at POL00081928 and can we go down 25 the page, please, to Ms Chambers' email. Scroll down, 27 1 please. Hmm. Just scroll back up, please -- and 2 again -- and again. That's obviously an erroneous 3 reference. I'll come back to that maybe later this 4 morning. 5 Can we turn instead to FUJ00083712. Thank you. 6 This is an email thread, if we scroll down, please, and 7 if we keep going, we can see an email from you there. 8 If we go to the end, thank you, and then scroll up from 9 the bottom. We can see the start of an email chain. If 10 we scroll up and keep going, you're not included at the 11 moment. You come in here, and just in what context 12 would you be sent an email like this? 13 A. I don't recognise this specific email but I recognise 14 the background to it as being part of the investigation 15 we were doing for an issue that I believe occurred in 16 Craigpark and one other branch, which I think is 17 Tonbridge. So, as a result of that, we then did 18 a retrospective search of NT events associated that were 19 related to ARQ data that had previously been supplied to 20 Post Office. 21 Q. Why were you doing that? 22 A. To make sure that there weren't any problems that we'd 23 not spotted that could cause any -- there weren't 24 problems that were occurring at the branch at the time 25 that the ARQ data was being extracted for. 28 1 Q. Was this a hangover, essentially, from the supposed 2 resolution of the Callendar Square bug in 2006? 3 A. No, this was a separate issue, that was -- as I say, 4 it's -- I've referred to it in my witness statement as 5 the Craigpark issue. In fact, having re-read them over 6 the last few weeks, I'm now aware that, actually, the 7 first occurrence of it was at a different branch called 8 Tonbridge but it was Craigpark that was what caused Post 9 Office to investigate things further, when I actually 10 got involved in this, in I think it was August 2008. 11 Q. I see. So if we carry on scrolling up, please -- and 12 keep scrolling -- and again -- and stop there. Are you 13 essentially saying that this chain is about another kind 14 of Riposte lock issue that's separate from the Callendar 15 Square issue in 2006? 16 A. I'm not 100 per cent sure about that as to whether this 17 was a storm of events or just individual events. If 18 it's just individual events then it was a separate issue 19 and it was important, as was done here, to check to see 20 if anyone was actually logged on to the system at the 21 time. And that was the sort of check that was being 22 done when those events were being detected in our 23 checking of NT events associated with ARQ data. 24 Q. But I shouldn't see this, I shouldn't read this, as 25 essentially a continuation of the Callendar Square issue 29 1 discovered in 2006? 2 A. Not if it's an isolated, a Riposte lock issue, no. 3 Q. How would you know whether it's an isolated Riposte lock 4 issue or a continuing manifestation of the Callendar 5 Square bug? 6 A. If it was a Callendar Square issue there would be 7 hundreds of thousands of events occurring. 8 Q. Why? 9 A. Because the problem was continuing. So a lock was 10 being -- there was an attempt get a lock, and then 11 Riposte was continually to do the same thing, process 12 every ten seconds, and it was failing every ten seconds, 13 and will carry on doing that until the Riposte service 14 on that box was restarted. 15 Q. So you said earlier that you didn't know the full 16 circumstances that would cause -- 17 A. What I don't understand is why the lock was being held 18 all that time. The symptoms of it I do understand, in 19 that it would be showing all these events, but why the 20 thing was being locked and not able to process things is 21 what I don't understand. 22 Q. Was it possible that the Riposte lock problem had 23 continued and was impacting on a branch in a way that 24 was previously unknown to Fujitsu? 25 A. I think what we were -- what we'd decided was the safest 30 1 thing to do was to check, if we saw the Riposte lock 2 problem, to see if anyone was actually using the 3 terminal at the time and, if no one was using the 4 terminal at the time, then it was deemed to be safe and 5 that -- nothing to worry about. 6 Q. We know you were asked some questions about Callendar 7 Square in 2010 for the purposes of the Seema Misra case, 8 yes? 9 A. Correct. 10 Q. And that you in turn asked Anne Chambers for information 11 about the problem? 12 A. Correct. 13 Q. Did you make any connection with your previous 14 involvement in Riposte lock issues such as this? 15 A. No, I think is the simple answer to that. 16 Q. Why would that be? 17 A. Well, at that stage, I didn't know, actually, what the 18 root cause in Callendar Square. I'd just been told 19 "Please can you look into what the Callendar Square 20 issue was". When I asked Anne for the information about 21 it, she passed me references to a PEAK and I think 22 she -- some old emails from the time that it had 23 occurred, and said that, if I remember rightly, that the 24 main cause of the issue was these what she called "event 25 storms", which was repeated events, which is separate 31 1 from these isolated events that we're looking at here. 2 Q. But would, in 2010, all of the earlier Riposte lock 3 matters -- I'm going to call them -- have been available 4 to you on PinICLs, PEAKs and KELs, if you had searched 5 for relevant key terms? 6 A. Yes, I guess they were but, again, in what we were 7 looking at, in terms of what had happened in West 8 Byfleet, we did actually extract all the NT events from 9 2005 to 2008, and we checked all the events, including 10 lock issues, and I believe there were some found. 11 I believe there is an email from Anne where she actually 12 summarises what she had actually found there, and there 13 were some further checks that Penny Thomas made to see 14 whether anyone was logged on at the times that these 15 events had occurred. 16 Q. I'm just thinking more broadly at the moment, without 17 getting into the detail of the individual cases, what 18 you would do if you were preparing a witness statement 19 in a case, or preparing to give evidence in a case: 20 would you go back and look at PinICLs, PEAKs and KELs in 21 order to conduct research about what was known about an 22 issue, or would you, as I think you just described, just 23 email or talked to Anne Chambers? 24 A. I would probably have asked the SSC because I wasn't 25 particularly adapt at being able to do searches on 32 1 PinICLs and KELs. I'd probably rely more on PinICLs 2 than KELs because I found those were useful in terms of 3 giving me the information I wanted. 4 Q. But if you were to give evidence about a bug, your 5 primary source of reference would have been to speak to 6 people in SSC; is that right? 7 A. Correct. 8 Q. Can we go forwards, please, to 2015 to FUJ00083775 and 9 look at page 2, please. In 2015, Anne Chambers was 10 asked to provide a description of the Callendar Square 11 problem, for the purposes, I think, of some litigation. 12 We can see here she says to Pete Newsome -- do you 13 remember him? 14 A. Yes. 15 Q. What function did he perform? 16 A. He was a -- I'm trying to think exactly what his job 17 title was. He was a liaison between Fujitsu and Post 18 Office, at a fairly senior level. 19 Q. "I have found a summary of the problem that Gareth 20 produced when preparing evidence for a court case in 21 2010." 22 That's a reference to the Seema Misra case, I think. 23 A. I assume so. 24 Q. Yes: 25 "I have made a few amendments to make it more 33 1 generally applicable. 2 "1. The problem occurred when transferring cash or 3 stock between stock units. 4 "2. It manifested itself by the receiving stock 5 unit not being able to 'see' the transfer made by the 6 'sending' stock unit and could be compounded by 7 attempting to make a further transfer. Note that such 8 transactions usually reappeared the next day. 9 "3. It was clearly visible to the user as 10 a 'Receipts and Payments mismatch' at the time that one 11 of the stock units was balanced. 12 "4. The problem was also visible when looking at 13 the system events associated with the branch. 14 "5. The problem was fixed in S90 release which went 15 live in March 2006. 16 "Receipts and payments mismatches normally showed up 17 on one of the reconciliation reports and would have been 18 investigated and reported to Post Office at the time 19 (though of course we don't know what action they took)." 20 Does that summary by Anne Chambers match your 21 understanding? 22 A. Yes, and I've certainly deferred to her to have 23 a clearer understanding of the detail than I did. 24 Q. You'll see in the last paragraph, after the numbered 25 paragraph, she says that receipts and payments 34 1 mismatches normally showed up. Is it the case that 2 that's essentially a reflection: that neither you nor 3 Ms Chambers would be able to categorically to state that 4 every incident or error would be caught by the 5 reconciliation process and directed to the Post Office 6 to take action? 7 A. I can't remember now exactly what the reconciliation 8 report said so I'm afraid I'm unable to help you. 9 Q. Could you tell, if a problem which led to a discrepancy 10 was referred to the Post Office, whether it was, in 11 fact, actioned through the issue of a correction, 12 a transaction correction? 13 A. That was up to the Post Office. There was 14 an organisation called Management Support Unit, I think 15 it was, MSU, and they raised things called BIMS reports, 16 and I can't remember what "BIMS" stands for. But if 17 a reconciliation error or something unusual was found 18 they would produce reports for Post Office describing 19 what had gone on, and what Post Office did with those 20 reports, in terms of how they affected the branches, 21 would be something for Post Office to answer. 22 Q. If we scroll up, please. A further email from 23 Ms Chambers, you're not copied in. 24 A. I'd retired by then. 25 Q. Yes. In the third paragraph, you'll see that she says: 35 1 "I've found a long email chain ... which I've 2 attached ..." 3 She says: 4 "Having looked at these, I slightly disagree with 5 Gareth's description of the problem below. My wording 6 would be: 7 "2. It manifested itself by the receiving stock 8 unit not being able to 'see' that the transfer made by 9 the 'sending' stock unit had already been accepted in, 10 and compounded by accepting it again. Other 11 transactions might have been omitted from the balance 12 but would have been included in the following week's 13 balance instead. 14 "But this is all a very long time ago and it is hard 15 to be entirely accurate." 16 What do you understand her to be saying is the 17 difference between you and her? 18 A. I'm not quite sure, actually. It would be useful to be 19 able to see the two statements side by side, I don't 20 know if that's possible -- 21 Q. I don't think -- 22 A. Or if we find out what the URN is, maybe I can have 23 a look at it during a break. 24 Q. Well, let's have look at the lower email. Keep going, 25 and you'll see how 2 is written there and, if we go up, 36 1 how she rephrases it. 2 A. I think it's the extra "has already been accepted in and 3 compounded by accepting it again", is from -- what 4 I think is the change, but I can't be 100 per cent sure 5 of that. 6 Q. What's the relevance of that? 7 A. That's why there's the receipts and payments mismatch, 8 that you have to actually repeat the transfer for the 9 problem to occur and, clearly, I'd missed that out of 10 the description that she was quoting below, but I accept 11 that that is a -- an important difference to the issue, 12 and -- in terms of describing what actually happened in 13 the Callendar Square/Falkirk branch at the time. 14 Q. Okay. If we scroll up, please. Thank you. I think 15 that's the end of the chain. Would you agree that this 16 was a problem, looking at it from the subpostmaster's 17 perspective, that they would be faced with different 18 figures on different counters as a result of the bug? 19 A. Yes, I accept that would be a way it would show up, yes. 20 Q. And that would be a situation which would rightly cause 21 them considerable difficulties and confusion? 22 A. I accept that, yes. 23 Q. This wasn't a problem that could be explained away by 24 user error? 25 A. Correct. 37 1 Q. Were you ever party to the explanations given by the 2 Service Support Centre or the lines of support below 3 them? 4 A. I'm not quite sure what you mean by that. 5 Q. Yes. In your work, did you see the kinds of things that 6 the Helpdesks or the SSC on occasions said directly to 7 subpostmasters? 8 A. Occasionally but, in general, no because what -- when 9 I saw PEAKs and PinICLs, they were being passed on to 10 me, or the teams that I was involved in, by the SSC 11 looking for an explanation, and what was being responded 12 back by the SSC to the Helpdesk, and also to the 13 branches, would be based on the information that had 14 come from the developers or the designers as to what 15 they had seen and gone on. So we wouldn't necessarily 16 see what -- the comments they made. 17 Q. Was there a formalised process for providing the Post 18 Office with information about bugs, errors and defects 19 and their effects on their live estate? 20 A. I don't know. There were a number of meetings that were 21 held between the Customer Services team and their 22 equivalents in Post Office but the scope of what those 23 discussed, and so on, I'm afraid I don't know. I wasn't 24 involved in them, usually. 25 Q. So you, on each bug, error or defect, weren't in 38 1 a position to know precisely what information was 2 provided to the Post Office -- 3 A. Correct. 4 Q. -- as to the existence of the bug, the extent of the 5 problem or the fixes proposed? 6 A. In general. There were a few cases where I was involved 7 in the detail because I was called in to actually help 8 explain the problem to Post Office but, in general, no. 9 Q. Where did responsibility for that lay within Fujitsu? 10 A. There was a team called Problem Management, which was 11 part of the Customer Services Team, and I think those 12 have been referred to in some of the -- well, in my 13 witness statements. 14 Q. Were you responsible for liaison with the Problem 15 Management Teams? 16 A. No. I was occasionally contacted by the Problem 17 Management Team or by the SSC to get involved in 18 discussions with things but, on a regular basis, no. 19 Q. Thank you. Thank you. Can we turn, please, to the 20 topic of remote access. 21 A. Okay. 22 Q. Can we look at your third witness statement, please, at 23 paragraph 473, which page is 164. You are here dealing 24 with your understanding of remote access at the time of 25 Seema Misra's case, ie in about, I think, the autumn of 39 1 October 2010. If we read it, having set out an extract 2 from a document you had created, you say: 3 "In this paragraph, I addressed Professor 4 McLachlan's hypothesis about transaction corrections 5 (which were the most common means by which [Post Office] 6 could, externally, impact the accounts of a branch). 7 I noted this could only happen with the consent of the 8 [subpostmaster] or an authorised supervisor or manager. 9 I was not addressing the issue of remote access as it 10 has been considered in the Group Litigation and [the] 11 Inquiry. I understand that remote access by Fujitsu is 12 now regard as an issue of some significance. But at 13 that time, remote access by Fujitsu didn't occur to me 14 in the context of responding to Professor McLachlan's 15 hypothesis, which related to [the Post Office's] ability 16 to impact branch accounts through transaction 17 corrections. Aside from the fact that I was responding 18 to his hypothesis, I understood then that remote access 19 by Fujitsu was an exceptional and regulated course. 20 There was nothing in Mrs Misra's case to [suggest] that 21 a need for remote access that ever arisen or that it 22 could in any way explain ongoing losses in the region of 23 £70,000. It was just not something that I had 24 considered." 25 So at that time, in the autumn of 2010, was it your 40 1 understanding that remote access was, firstly, 2 controlled? 3 A. It depends exactly what you mean. I believed that there 4 were processes that the SSC had by which they recorded 5 when they did carry out remote access to insert messages 6 into the message store. 7 Q. That it was recorded and auditable? 8 A. I understood that they had a process whereby it was 9 recorded. I didn't know the details of what that 10 process was. I believe it involved things called OCPs 11 or OCRs, and I don't know that I ever knew the 12 difference between them, but I have seen both phrases 13 used and I believe they were different mechanisms. 14 Q. Was it your understanding that remote access was visible 15 and identifiable by a subpostmaster who could, for 16 example, tell that there had been remotely inserted 17 balancing transaction? 18 A. That was my view at the time, though I've since come to 19 realise, as part of the Group Litigation, that it wasn't 20 always done in quite the way that I thought it was done. 21 Q. That was what I wanted to ask you about. It was your 22 understanding at the time -- 23 A. Yes. 24 Q. -- in autumn 2010 that the remote access you're speaking 25 about here was controlled, recorded, auditable and 41 1 visible by the subpostmaster? 2 A. Correct. That was my view then. 3 Q. What was your understanding based on? 4 A. Informal conversations with SSC staff. 5 Q. Anyone in particular? 6 A. People like John Simpkins and Anne Chambers. And 7 I believe they have both given evidence to the Inquiry 8 about their understanding of how those processes worked 9 and how frequently they occurred. 10 Q. Was that understanding, that any incidences of remote 11 access would be visible to a subpostmaster, a reason why 12 you did not the address the issue of remote access in 13 any of your reports or witness statements as a possible 14 cause of shortfalls? 15 A. It didn't occur to me that that was relevant. With 16 hindsight, I realise that perhaps it was but it was 17 something that I didn't think of as being relevant at 18 the time. 19 Q. Did you give thought to it and dismiss it or did you not 20 give thought to it? 21 A. I just didn't give thought to it, I'm afraid. 22 Q. Thank you. Can we turn, please, to INQ00001079. This 23 is a transcript of Anne Chambers' evidence. There's 24 a disclosure issue at the start of the transcript. If 25 we go to page 17, please, and internal page 65. 42 1 Perfect. 2 Then if we just scroll down to page 67. So if we 3 just go back up to 65. I should read this, just to put 4 it in context. Line 5, page 68, I said: 5 "Back when you gave evidence on 3 May, you described 6 an occasion when the Post Office had wanted to insert 7 transactions without the branch being aware. I'll just 8 read back the [Q&A]. The question was: 9 "'Was there any method to alert others that 10 corrective action had been taken to insert data or extra 11 messages into a branch's account?' 12 "You said: 13 "'The ARQ data would contain both the original 14 transaction and the corrective transaction, at the point 15 at which they were done. If the full, unfiltered data 16 was retrieved and inspected, then that would show the 17 comment, for example. Certainly, sometimes, for counter 18 corrections -- and it wasn't done consistently -- but we 19 often might use a counter number that didn't exist to 20 make it clear that it was something out of the ordinary 21 or a username, including SSC, again to show that it was 22 something out of the ordinary. 23 "'That wasn't done on this specific one and I cannot 24 remember whether that was because I was specifically 25 asked not to or I was just producing a transaction that 43 1 was absolutely a mirror of the one that shouldn't have 2 been there in the first place and all I did was change 3 the signs of the values, effectively, and ... I left all 4 the other data as it was.' 5 "So remembering back to what you were saying and 6 judges to synthesise it, you were saying, sometimes 7 a fictitious counter number was used to mark out the 8 transaction correction? 9 "Answer: As I recall, yes. 10 "Question: Secondly, you were saying [that] that 11 wasn't done consistently, ie the use of a fictitious 12 counter number to mark out the fact that the SSC had 13 made a correction? 14 "Answer: Yes. 15 "Question: Thirdly, you were doing that, or the SSC 16 was doing that, deliberately, ie using the fictitious 17 counter number, because you would want to show that an 18 SSC member had been making the correction? 19 "Answer: Yes." 20 Then page 67: 21 "Question: But, fourthly, there might be occasions 22 when you were specifically asked not to use the 23 fictitious one? 24 "Answer: I don't think we were ever asked not to 25 use the fictitious one. 44 1 "Question: It was just passage of your evidence 2 where you said 'That wasn't done on this occasion and 3 I can't remember whether that was because I was 4 specifically asked not to', which tended to indicate 5 that you may have been asked not to use the fictitious 6 number? 7 "Answer: I don't recall ever being asked not to do 8 that and I can't remember which specific instance we're 9 talking about here. Sorry. 10 "Question: Overall, does it mean that it was 11 possible for members of the SSC to insert transactions 12 using the branch user ID? 13 "Answer: Right, you're talking about user ID here 14 now, rather than counter number, but, yes, it was -- 15 I mean, the messages that we inserted could have 16 contained the branch user ID. 17 "Question: Would it follow that standard filtered 18 ARQ data would not distinguish those insertions from 19 those that were, in fact, carried out in the branch? 20 "Answer: The standard ARQ data, yes. You might not 21 be able to see the difference. 22 "Question: So transactions which appeared in the 23 standard filtered ARQ data, for example, in 24 Mr Castleton's case, with his ID user number next to 25 them, would not necessarily mean that they were carried 45 1 out by him? 2 "Answer: It would have been possible, yes, for SSC 3 to put transactions in that -- 4 "Question: Using his ID? 5 "Answer: Using his ID. 6 "Question: Without leaving a fingerprint on the 7 standard, filtered ARQ data that that had been done? 8 "Answer: Yes, I think that would have been 9 possible. 10 "Question: It shouldn't have been done but it could 11 have been done? 12 "Answer: Yes, I don't think there's anything that 13 would have prevented that. I don't believe that was 14 done but I can't say it's an impossibility." 15 Then we moved on. 16 When did you become aware that the SSC were able to 17 insert transactions into branch accounts in the way 18 Ms Chambers describes here? 19 A. I was aware that technically they could but my 20 understanding was that they always injected the 21 transactions at the correspondence server, rather than 22 the counter. In 2018, as part of the Group Litigation 23 Order, it was explained to me by John Simpkins, I think, 24 that, in some circumstances, they actually did it 25 directly at the counter, rather than at the 46 1 correspondence server and that was different from the 2 understanding I had as to how they operated up until 3 that time. 4 Q. Do you know why you hadn't been told that previously? 5 A. No, is the simple answer. 6 Q. You said that you'd had conversations with people in the 7 SSC but they hadn't revealed that to you in previous 8 conversations? 9 A. As I understand it, and from what John Simpkins 10 explained to me in 2018, normally, they would insert 11 transactions at the correspondence server in the way 12 that I was expecting, but there were a few specific 13 cases where they were doing it at the counter, instead, 14 in the way that Anne is describing in this exchange. 15 Q. Did you learn about this practice of generally using 16 an SSC ID in order to mark them out but that not being 17 a consistent practice? 18 A. When things were inserted in the correspondence server, 19 I would have expected the details of the SSC user and 20 why they were doing it to be inserted as part of that, 21 even though it was also very obvious that it was being 22 done at a strange counter. 23 Q. But what about what Ms Chambers has described here? 24 A. I -- it was only when I realised that they were actually 25 really doing things at the counter in 2018 that I was 47 1 aware that that was a possibility -- 2 Q. In -- 3 A. -- or something that was actually done in practice, as 4 opposed to in theory. 5 Q. In 2018, did you also learn not only that they were 6 doing it at the counter, which is what is described 7 here, but they were doing it in a way that might 8 attribute a change to the subpostmaster, rather than 9 a change being made by the SSC? 10 A. If it was done at a counter and a postmaster was logged 11 on at the time, then the way Riposte worked, it would 12 always be attributed to whoever was logged on at the 13 time. So it all comes down to whether they -- when they 14 were inserting the transaction, whether they were doing 15 it when someone was logged on or not. If there was no 16 one logged on at the counter where they inserted it, or 17 they were using a fictitious counter, then there would 18 be no use for ID associated with it. 19 Q. But if they were doing it in the hours of business when 20 the subpostmaster was logged on, their work might be 21 attributed to him or her? 22 A. I accept that, yes. 23 Q. When did you come to learn of that? 24 A. As I say, I always knew it was theoretically possible 25 but it wasn't until 2018 that I realised that SSC on 48 1 some occasions actually did that. 2 Q. Were actually doing it? 3 A. Yeah. 4 Q. Thank you. 5 A. I think I cover that more in my fourth witness 6 statement. 7 MR BEER: You do. 8 Sir, that might be an appropriate moment for the 9 morning break. 10 SIR WYN WILLIAMS: By all means. What time shall we resume? 11 MR BEER: If we can have a longer break and do one break, if 12 that's possible. So maybe 20 minutes until 11.40. 13 SIR WYN WILLIAMS: Fine. 14 MR BEER: Thank you, sir. 15 (11.20 am) 16 (A short break) 17 (11.40 am) 18 MR BEER: Thank you. 19 Mr Jenkins, can we look, please, at INQ00001058. 20 This is a transcript of the evidence of Richard 21 Roll. Can we look at page 21, please, and then internal 22 pagination page 81, which is the top left, from line 11 23 onwards, which I think is a question by me. I say: 24 "... can we go back to you witness statement to 25 page 6, and look at (c) ... 49 1 "'We could go directly through the communication 2 servers to the [Post Office] gateway and then the 3 counter -- if the [postmaster] wasn't logged in then 4 there would be no ID attached to the database entries, 5 which sometimes caused the batch processing to fail 6 overnight; if the [postmaster] was logged on then any 7 changes we made would have their ID attached -- so as 8 far as the system (and any auditing) was concerned the 9 [subpostmaster] would have been responsible for the 10 transactions.' 11 "Answer: That's what I was trying to say. I think 12 that's what I was trying to say in the Post Office 13 transcript we just looked at. 14 "Question: Thank you. Was this a method that you 15 used frequently ... 16 "Answer: We were all pretty adept at it, yeah. 17 "Question: Whether you were adept at it -- 18 "Answer: Fairly frequently, yes. 19 "Question: Fairly frequently? 20 "Answer: Yes. 21 "Question: ... Why did you use that method? 22 "Answer: It was the only way to rebuild the 23 counters to get the data off the counters. 24 "Question: The footprint that was left would have 25 been the subpostmaster's footprint and not yours? 50 1 "Answer: Yes. 2 "Question: Was there any visibility that you or 3 somebody else in the SSC had done this as opposed to the 4 subpostmaster themselves having done it? 5 "Answer: Sometimes yes, sometimes no. 6 "Question: What would distinguish? 7 "Answer: We would -- sometimes it would be 8 recorded. I'm a bit rusty on this now, I'm afraid, but 9 sometimes we told the postmaster we were going to do it. 10 While we were doing this, the postmaster couldn't use 11 the counter. It was very important that nobody used it. 12 At other times, especially if maybe the postmaster -- 13 I'm just thinking. I'm just trying to remember 14 something else ... if the postmaster had gone to lunch, 15 [for example], we could have gone in and done things 16 without him knowing." 17 Just stopping there. I think you accept now that 18 the SSC staff could not only effectively hijack branch 19 user IDs in this way to insert transactions into branch 20 accounts but did so? 21 A. Yes, I accept that now. In fact, it was as a result of 22 seeing Mr Roll's witness statement for the Group 23 Litigation that then I enquired further of the SSC to 24 understand more as to exactly how they'd done the 25 injections. 51 1 Q. Did they tell you for how long they had been doing it? 2 A. It was something that was always possible within Legacy 3 Horizon. What I was told was that it wasn't done very 4 frequently but I don't think that figure was quantified. 5 I believe John Simpkins may actually have tried to 6 quantify it at some stage but I can't remember the 7 details. 8 Q. What about the duration of it, rather than the 9 frequency? Was it throughout the use of Legacy Horizon, 10 ie from about 2000 to 2010? 11 A. It was certainly possible throughout that period, yes. 12 Q. Did they tell you anything as to whether it had been 13 done throughout that period? 14 A. I can't remember whether I asked that question or not. 15 Q. That, as I've described it -- hijacking a branch user ID 16 to insert a transaction into a branch account -- would 17 that make the entries in the standard, filtered ARQ data 18 impossible to distinguish from the transactions that 19 were, in fact, entered by the subpostmaster in the 20 branch? 21 A. In the standard ARQ data, yes, if it was done on a real 22 counter. Though, again, as Anne Chambers was saying, 23 they often did it using false counters -- 24 Q. Yes, and -- 25 A. -- and then it would be obvious that it wasn't done on 52 1 a real counter. 2 Q. When did you discover that this could be done, was being 3 done and, therefore, when it was done, the entries in 4 the standard ARQ data were impossible to distinguish 5 from transactions in fact carried out by the 6 subpostmaster in branch? 7 A. As I say, I always knew it could be done but I didn't 8 realise it was being done until 2018. 9 Q. Okay, so you always realised that it could be done -- 10 A. Yes. 11 Q. -- right from the year 2000? 12 A. Yes. 13 Q. How did you know that it could be done? 14 A. Because I knew what functionality the Riposte provided 15 and the sort of interfaces it provided to allow it to be 16 done. 17 Q. But nobody had told you, from that time onwards, that 18 they were in fact doing it? 19 A. Correct. 20 Q. You discovered that in 2018/19? 21 A. That is my belief, yes. 22 Q. So you knew that the SSC had the facility to tamper with 23 branch accounts? 24 A. I don't think I'd call it tampering but, yes, 25 I understand what you're saying. 53 1 Q. Well, it would be tampering, wouldn't it? I know that 2 carries pejorative overtones -- 3 A. That's why I didn't like the term. I believe that the 4 SSC would have been using that facility responsibly and 5 doing it when they needed to, to correct errors that had 6 occurred in the branches. 7 Q. The standard filtered ARQ data used in court proceedings 8 wouldn't be able to distinguish between those occasions 9 when a transaction had been made by the subpostmaster 10 and when it had been made by the SSC? 11 A. Not if it had been done at the counter. 12 Q. Therefore, the standard ARQ data used in cost 13 proceedings couldn't, of itself, prove that the branch 14 accounts had not been, in my words, tampered with? 15 A. I accept that now but I hadn't thought of it that way at 16 the time. 17 Q. When you prepared your witness statements, both the 18 case-specific witness statements and the generic witness 19 statements, you didn't mention the facility that the SSC 20 had to undertake this function, did you? 21 A. No, because I didn't think it was relevant. Again, with 22 hindsight, I realise that perhaps I should have done 23 but, at the time, it was -- my understanding was, and 24 still is, that it was a very rare thing for the SSC to 25 actually do. 54 1 Q. Did you not realise that, whether the SSC told you that 2 it was being done rarely, wasn't the real issue; it was 3 the extent to which there was a tamper-proof evidential 4 chain being presented to a court? 5 A. I'd not thought it through in that way at the time, I'm 6 afraid. 7 Q. Can I turn to your own knowledge of remote access by 8 turning up WITN00420100. This is Ian Henderson's 9 witness statement. Do you remember him from Second 10 Sight? 11 A. Yes. 12 Q. Can we turn to page 13, please, and look at paragraph 43 13 at the bottom. He says: 14 "In September 2012, I met with Gareth Jenkins, the 15 lead engineer for Post Office Horizon, at the Head 16 Office of Fujitsu in Bracknell. He told me that 17 approximately 10 members of staff from Post Office were 18 permanently based in Bracknell, dealing with various 19 issues including bugs, errors and defects. 20 "44. Gareth Jenkins told me that Fujitsu routinely 21 used remote access to branch terminals for various 22 purposes. This was often without the knowledge or 23 specific consent of individual subpostmasters. He also 24 told me that members of his team could connect remotely 25 to branch terminals and generate keystrokes that were 55 1 indistinguishable from a subpostmaster accessing the 2 terminal directly. They did this for various purposes, 3 including collecting log files directly from branch 4 terminals." 5 Is what Mr Henderson says there accurate? 6 A. I don't think so. I don't think that is how I would 7 have explained things. I mean, particularly the last 8 bit about collecting log files directly. That's nothing 9 to do with adding -- changing transactions into the 10 accounts. 11 Q. Do you agree that in September 2012 you met Ian 12 Henderson? 13 A. Yes, I agree. 14 Q. Do you agree that was at the Head Office in Bracknell? 15 A. Yes. 16 Q. Did you tell him that members of staff from Post Office 17 were based in Bracknell permanently, dealing with issues 18 including bugs, errors and defects. 19 A. I was referring to some of the testers, not people -- 20 those Post Office staff members did not have access to 21 the live system. 22 Q. You weren't, in your conversation, referring to the SSC 23 then? 24 A. No. Not in terms of paragraph 43. 25 Q. Did you tell him that Fujitsu routinely used branch 56 1 access to branch terminals for various purposes? 2 A. I probably said that SSC had access to the branches for 3 pulling back things like diagnostic logs, and things 4 like that, and I may well have suggested that it was 5 possible for SSC members to inject correcting 6 transactions. But I would have seen that as being done 7 through the correspondence server at the time -- 8 Q. Did you -- 9 A. -- but whether I spelt that out in detail at the time, 10 I just can't recollect the details of the conversations. 11 Q. Did you tell him that such remote access was often 12 without the knowledge or specific consent of individual 13 subpostmasters? 14 A. Well, if it's referring to actually pulling back log 15 files, there would be no need for the knowledge or 16 consent of the subpostmasters. 17 Q. Did you tell him that members of your team could connect 18 remotely to branch terminals and generate keystrokes 19 that were indistinguishable from a subpostmaster 20 accessing the terminal directly? 21 A. I don't think I would have said those because the only 22 people who had that sort of access were the SSC, and 23 they weren't part of my team. 24 Q. Did you tell him that anyone in Fujitsu could connect 25 remotely to branch terminals and generate keystrokes 57 1 that were indistinguishable from a subpostmaster 2 accessing the terminal directly? 3 A. I wouldn't be generating keystrokes, so any injection of 4 data wasn't done by generating keystrokes. 5 Q. Did you tell him that Fujitsu staff could connect 6 remotely to branch terminals and alter branch accounts 7 in a way that made it indistinguishable from the 8 subpostmaster doing so? 9 A. I don't think I would have said that because my 10 understanding was it was actually done at the 11 correspondence server at the time. 12 Q. In general terms, you say you wouldn't have said this 13 because you didn't know it at the time: you only 14 discovered this in 2018/2019? 15 A. That's my understanding, yes. 16 Q. What is he's attributing to you, in September 2012, 17 never mind the precise details, but generally speaking, 18 is accurate, isn't it? 19 A. That there was remote access, yes, but not in terms of 20 the gory details of it. 21 Q. Well, not only that there was remote access, there was 22 remote access of a kind which made it indistinguishable 23 from the subpostmasters' activities? 24 A. Well, I'm not sure that would be the case because what 25 I was thinking of was remote access to the 58 1 correspondence server, which would be distinguishable. 2 Q. He was, it seems, told, that there was a form of remote 3 access that made what had happened indistinguishable 4 from the subpostmaster carrying out the relevant 5 functions. That, you say, you didn't know at the time 6 was happening but you subsequently found to be true in 7 2018/2019? 8 A. Again, I knew it was possible but I didn't know that it 9 was being regularly done. 10 Q. What was the context of your conversation with him in 11 September 2012; why were you talking about remote access 12 at all? 13 A. I can't remember. My recollection was that we had 14 a meeting when I was talking to him about how he could 15 actually interpret and analyse the logs. 16 Q. Do you think you would have spoken about remote access 17 in that connection, ie what the logs did or didn't show? 18 A. I can't remember, I'm afraid. 19 Q. Can we go back in time a little to 2002 by looking at 20 FUJ00088036. Can you see that this is a Fujitsu 21 document, a design document, version 1 of which is dated 22 2 August 2002? 23 A. I can see that, yes. 24 Q. It's described as 2Secure Support System Outline 25 Design"? 59 1 A. Yes. 2 Q. You're neither an originator nor a contributor? 3 A. Correct. 4 Q. You'll see the abstract: 5 "[It] describes the requirements and outline design 6 for secure operational support access to counters and 7 servers." 8 Then if we scroll down, please, look at the approval 9 authorities on that page, and then over the page -- and 10 scroll down -- you're not included in any of those? 11 A. Correct. 12 Q. Is this a document you would have seen at the time? 13 A. Probably not. 14 Q. If you can tell us in general terms why not? 15 A. It wasn't an area that I was particularly involved in. 16 Q. Because it's about the functions of the SSC; is that 17 right? 18 A. It was to do with some tooling that was being provided 19 for the SSC is what I understand from it from reading it 20 now. 21 Q. Can we look, please, at page 15, and look at 4.3.2. 22 This is describing the position in 2002: 23 "... support access to the Horizon system is from 24 physically secure areas." 25 You would know that to be the case: the SSC was, as 60 1 you've told us -- 2 A. Yes. 3 Q. -- was in a physically secure area? 4 "Individuals involved in support processes undergo 5 more frequent security vetting checks." 6 Were you were of that? 7 A. I wasn't aware of that but I'm not surprised that they 8 were. 9 Q. "Other than the above[,] controls are vested in manual 10 procedures, requiring managerial sign-off controlling 11 access to Post Office Counters where update of data is 12 required. Otherwise third line support has: 13 "Unrestricted and unaudited privileged access 14 (system admin) to all systems including post office 15 counter PCs ..." 16 Did you know that was the position in 2002? 17 A. I can't remember. I wasn't involved in the detail. 18 Q. Did you ever come to learn that they, in the SSC, had 19 unrestricted and unaudited privileged access to all 20 systems including counters? 21 A. I know that later on it was audited. I didn't realise 22 that this was something that had not been introduced at 23 day 1. 24 Q. When did you find out it was controlled in some way or 25 audited in some way? 61 1 A. Well, what this outline design is to do with the 2 mechanisms that were introduced to give that control. 3 What I didn't realise was that that hadn't been there 4 from the very beginning, until I saw this document as 5 part of my preparation for the Inquiry. 6 Q. Okay, so it wasn't until 2021 -- 7 A. Maybe later, yeah. 8 Q. -- or even later -- 9 A. Yes. 10 Q. -- that you realised that the SSC had unrestricted and 11 unaudited privileged access to counters? 12 A. As far as I can remember, yes. 13 Q. It says, secondly, third line support has: 14 "The ability to distribute diagnostic information 15 outside of the secure environment; this information can 16 include personal data", et cetera. 17 The document continues: 18 "The current support practices were developed on 19 a needs must basis; third line support diagnosticians 20 had no alternative other than to adopt the approach 21 taken given the need to support the deployed Horizon ... 22 "There are ... no automatic controls in place to 23 audit and restrict user access. This exposes Fujitsu 24 ... to the following potential risks: 25 "Opportunity for financial fraud; 62 1 "Operational risk -- errors as a result of manual 2 actions causing loss of service to outlets; 3 "Infringements of the Data Protection Act." 4 When you were giving evidence about the operation of 5 Horizon in your witness statements, is this the kind of 6 information that you ought to have known about? 7 A. With hindsight, probably, yes. 8 Q. When you wrote your witness statements and when you gave 9 evidence in the Seema Misra case, did you make any 10 efforts to find out about issues such as this? 11 A. My understanding was that, at that time, things were all 12 being audited and there were processes in place but 13 I didn't understand the details of the processes and 14 what auditing was taking place. 15 Q. When did you gain an understanding that everything was 16 being audited and processes were taking place? 17 A. I don't -- I can't remember, is the simple answer. 18 Q. Can we look, please, at FUJ00089535. If we just look at 19 the date of this by scrolling down -- over the page. 20 There, the foot of the page -- sorry, just at the top. 21 24 August 2016; do you see that? 22 A. Yes. 23 Q. If we go back up to the first page, "Remote Support 24 Secure Access Server High Level Design". Is this 25 speaking about a similar thing than as in the previous 63 1 document we looked at? 2 A. I don't know. I don't recognise that. I see that what 3 looks like my name is there but that is the other Gareth 4 Jenkins working on Post Office because I had retired at 5 that point. 6 Q. I was about to ask you, if we scroll down, please. The 7 "HDCR Solution Owner", that's not you? 8 A. No, I had retired by that time and I was aware that 9 there was another Gareth Jenkins working on the account, 10 which did cause some confusion. 11 Q. So you didn't see this document at the time? 12 A. No. 13 Q. So the references that I would find in it backed the 14 2002 document that we looked at, it's a reference 15 document and there are some extracts from it cited in 16 the body, you wouldn't have seen at the time in 2016? 17 A. No. 18 Q. Thank you. Can we turn, please, to POL00097217. If we 19 scroll down, please, if we go to the foot of the 20 document -- and scroll up, keep going so we get that 21 first email. Thank you. 22 An email from Rachael Panter to you of 26 November 23 2012, copied to others, asking for an expert report. 24 This is in the case of Kim Wylie, yes? 25 A. Yes. 64 1 Q. She says: 2 "Dear Gareth 3 "Request for an urgent expert report 4 "Further to my previous email, I am writing as 5 a matter of urgency to request a signed and dated copy 6 of your report in the case of Kim Wylie ... 7 "I have attached a copy of [defence solicitors'] 8 request for further disclosure ... 9 "I have also attached some documents ... 10 "All of which set out the background to this case. 11 Please can you consider [this] and include this 12 paragraph at the beginning of your report: 13 "'I have been asked to prove statement in the case 14 of Kim Wylie. I understand that the integrity of the 15 system has been questioned and I can say the following.' 16 "As there have been no specific criticisms of 17 [Horizon] raised by the defence, your previous report 18 can be used with the addition of the opening line 19 referred to above. Please see the Defence Case 20 Statement which contains the nature of the defendant's 21 case." 22 If we scroll up, please, to see your reply: 23 "I have been advised that I can try to help you [you 24 say], if it isn't too late. 25 "I've had a read through the info you sent me 65 1 yesterday and I have also taken my previous witness 2 statement which I produced in early October and amended 3 it ..." 4 You're there, I think, referring to your generic 5 witness statement of 5 October 2012. 6 A. Correct. 7 Q. We're going to come back to that generic witness 8 statement later: 9 "Reading through the Defence Statement I see it does 10 make some specific points which my statement doesn't 11 currently address. Specifically the challenges 12 regarding Robustness and Remote Access to the system. 13 Do you want me to try and address these specifics?" 14 Then scroll down. Then some admin. Then if we go 15 up, she replies: 16 "[Thank you]. If you feel you are able to deal with 17 the issues of robustness and the Remote Access system 18 fairly swiftly then I would like you to address [these] 19 points that have been raised, so that we can deal with 20 every area that they have criticised." 21 Then scroll up, you say: 22 "What I propose is [adding] the following: 23 "I have been asked to provide a statement in the 24 case of Kim Wylie." 25 Just stopping there, this was going to be a bolt-on 66 1 to your generic statement of 5 October? 2 A. Correct. 3 Q. "I understand the integrity of the Horizon system has 4 been questioned and this report provides some general 5 information regarding the integrity of Horizon." 6 Then if we skip a paragraph and look at: 7 "I also note a comment has been made about it being 8 possible to remotely access the system. It is true that 9 such access is possible; however in an analysis of data 10 audited by the system, it is possible to identify any 11 data that has not been input direct by staff in the 12 branch. Any such change to data is very rare and would 13 be authorised by [the Post Office]. As I have not had 14 an opportunity to examine data related to this Branch, 15 I cannot categorically say this has not happened in this 16 case, but would suggest this is highly unlikely." 17 Dealing with that proposed insertion into the Kim 18 Wiley statement, which I'm not going to deal with it 19 now, you in fact entered into the Kim Wiley statement? 20 A. Yes. 21 Q. You say: 22 "It is true that [remote access to the system] is 23 possible ..." 24 A. Yes. 25 Q. You say that: 67 1 "... in an analysis of data audited by the system, 2 it is possible to identify any data that has not been 3 input directly by staff in the branch." 4 That's not accurate, is it? 5 A. It was as I understood things at the time but it isn't 6 in my understanding of things now. Well, sorry, when 7 I say that, if the SSC had been following their 8 processes, then it would be -- even if they'd been doing 9 it at the branch, it should have been visible in the raw 10 data but not in the ARQ extracts. 11 Q. What data were you referring to there? You say: 12 "... however in an analysis of data audited by the 13 system ..." 14 A. Whatever data was there. So when I was looking at data 15 I tended to have the raw data as well as the ARQ data 16 and I appreciate, and I say here, that I'd not had the 17 opportunity to examine any data from the branch. 18 Q. On what basis did you understand, here in 2012, that any 19 change to data was very rare? 20 A. From informal conversations that I'd had with the SSC as 21 I mentioned earlier. 22 Q. So the people -- 23 A. Not saying I'd had a conversation as part of this but 24 just from what I'd picked up over the years. 25 Q. So informal chats? 68 1 A. Yes. 2 Q. Where did you get the understanding that any such change 3 would be authorised by the Post Office? 4 A. I thought that was part of the process. As I understood 5 it, then any changes -- when they inserted data, then 6 there would be a mechanism whereby Post Office would 7 sign it off and I believe that both Anne Chambers and 8 John Simpkins indicated that as part of their testimony. 9 Q. You knew, when you were making this addition to the 10 statement, though, that there was a form of remote 11 access that could occur that would not be shown in the 12 data? 13 A. I knew that it was theoretically possible. I didn't 14 think it was something that was routinely done. 15 Q. Why didn't you say so? 16 A. Because I didn't think of it. I was assuming that 17 things were done in the normal way at that time. 18 Q. Do you think that that was the right basis to approach 19 statements made to criminal courts: on the basis of 20 assumptions, of propriety? 21 A. I realise now that that is the wrong basis for doing it 22 but that is what I felt at the time. 23 Q. You do say you cannot categorically say this hasn't 24 happened in the case but suggest it was highly unlikely. 25 On what basis were you saying it was highly unlikely? 69 1 A. Because I knew that -- well, I understood that remote -- 2 that inserting transactions by the SSC was a very rare 3 thing and I believe that has been established, as well. 4 Q. Can we turn on to POL00098724. This is an email to you 5 from Simon Baker of June 2013, so the following year. 6 If we just scroll down, he cites from an email that you 7 had sent, where he says: 8 "... the key message in your response is as follows: 9 "'The [Post Office] Bracknell Team had no access to 10 the live system therefore could not conduct any of these 11 transactions'." 12 Is that something that you told him? 13 A. Yes, I think so. The Post Office did not have access to 14 the live system. Yes. 15 Q. Then -- well, hold on. The Post Office Bracknell team? 16 A. Did not have access to the live system. They only had 17 access to test systems. 18 Q. If you said that, would you be referring to Post Office 19 personnel -- 20 A. Yes. 21 Q. -- stationed -- 22 A. That's what I meant by "POL". 23 Q. -- at Bracknell -- 24 A. I would have said "POA", if I was referring to the 25 Fujitsu team. 70 1 Q. Then he continues: 2 "... the previously supplied email suggests that the 3 Bracknell Team could access, possibly on an overnight 4 basis, transactions on the live system. 5 "Please see the following extracts ... 6 "'Although it is rarely done it is possible to 7 journal from branch accounts. There are possible P&BA 8 concerns about how this would be perceived and how 9 disputes would be resolved'." 10 Then scroll down, please, and over the page. 11 "At the moment your response is inconsistent with 12 what is being said in emails sent between Bracknell Team 13 members. 14 "Can you look into this urgently and address this 15 apparent inconsistency in an updated response?" 16 So you're being asked, essentially, to comment on 17 questions from Second Sight about staff at Bracknell 18 having access to the live system? 19 A. Yes. 20 Q. Did you reply to this? 21 A. I can't remember. I may well have done, yes. 22 Q. If we scroll to the top, I'm not sure that I found 23 a reply. 24 Did you have a conversation with Mr Baker about it? 25 A. I can't remember. This is part of what I think was 71 1 SR05, one of the spot reviews, and there were emails 2 about things, but this particular bit of it was talking 3 about the Post Office team and it was very clear to me, 4 from the context, that it was referring to the Post 5 Office team in the basement. I notice there was 6 a reference to journalling and I think that was 7 referring to what Post Office could do in their back-end 8 systems in Chesterfield, which would have no impact on 9 the actual branch accounts. 10 Q. I see. So the context of this exchange or these 11 requests is Post Office staff at Bracknell, rather than 12 Fujitsu staff? 13 A. That is how I understand it from looking through this 14 now. I can't remember the details of this specific 15 email off the top of my head but it is one that 16 I recognise as having looked at as part of my 17 preparation for the Inquiry. 18 Q. Thank you. Moving on, can we turn to FUJ00087100. If 19 we scroll down, please, a little further -- thank you. 20 Just scroll up, thank you, we can see an email from 21 Michael Harvey to you and James Davidson, and he says: 22 "With respect to inserted data -- can you provide 23 a little more detail as to the audit trail that addition 24 of data leaves. So notwithstanding that we haven't used 25 it, if we were to use, would the subpostmaster or Post 72 1 Office be able to see it in the audit record. I assume 2 [that] this is the case and so can we make it explicit. 3 We need to be cognisant of the point that these two 4 individuals are attempting to make -- they are trying to 5 say that [Fujitsu] have changed data and so in answers 6 Post Office's explicit questions the underlying answer 7 must be: 8 "1. When we have and why (ie the 2010 incident); 9 and 10 "2. If we had, the following audit trail would be 11 left demonstrating we had done it -- no audit trail 12 means we could not have changed the data." 13 Scroll up, please. Your reply: 14 "The answer is slightly different depending on 15 whether we are talking about the old Horizon (2010 or 16 earlier) or new Horizon Online (2010 onwards). 17 "However in both cases the injected transactions 18 would be visible in any local reports the postmaster may 19 view and also in the audit extract." 20 I think you now realise that to be wrong, don't you? 21 A. Yes, I do. 22 Q. On what basis were you saying it? 23 A. Because my understanding was that, whenever the SSC were 24 injecting transactions, they did so at the 25 correspondence server and not at the counter. 73 1 Q. Was that based on informal conversations as well? 2 A. I can't remember on what basis it was. It was probably 3 informal conversations and also the fact that it's 4 actually a lot easier to inject transactions at the 5 correspondence server than the counter in terms of their 6 processes. 7 Q. The context in which you were being asked this, I think 8 James Davidson was the lawyer, wasn't he? 9 A. No, James Davidson was -- at that stage, I think he 10 was -- he had been employed by Fujitsu as head of 11 Customer Services. 12 Q. Okay. 13 A. He'd then left and gone freelance and was primarily 14 involved in liaison along -- for Second Sight between 15 Fujitsu and Post Office. Michael Harvey was a lawyer, 16 I believe. 17 Q. Thank you. You were being asked by a lawyer for quite 18 precise information, if we go back to his email. 19 A. Yes, and that was my belief at the time. I accept now 20 that I had misunderstood things. 21 Q. Well, it wasn't misunderstood. What enquiries had you 22 made in order to provide the answer that you did? 23 A. At the time, none. I was relying on what I thought 24 I knew. 25 Q. Is that reflective of a general approach when involved 74 1 in threatened or possible court proceedings, or court 2 proceedings themselves: you work on the basis of what 3 you had understood from conversations and what you 4 thought you knew, rather than properly researching 5 an issue? 6 A. Yes, I guess that has to be the case. 7 Q. If we scroll back up to your reply. Yes, sorry, we're 8 there. You say: 9 "In both cases the audit extract would show (perhaps 10 not obviously and in different ways depending on the 11 version of Horizon) that these transactions had been 12 inserted in the data centre and not taken place on 13 a normal counter. 14 That's not correct in relation to standard ARQ data, 15 is it? 16 A. It would be in the ARQ data because one of the bits of 17 information in the standard ARQ data is the node ID or 18 the counter position and anything injected at the data 19 centre would have a node ID that reflected that it was 20 being done at the data centre. 21 Q. You say: 22 "Our processes should ensure that [Post Office] have 23 signed off on any occasion when such data was inserted." 24 What enquiries did you make as to the nature of the 25 processes? 75 1 A. At the time, none but my understanding had been that 2 there were standard processes, which I didn't know the 3 details of, that the SSC followed, which involved 4 getting things signed off by Post Office. 5 Q. Again, you say that was based on your understanding. 6 Was that just through general chat? 7 A. Probably, yes. 8 Q. Can we move forward, then, to the Horizon Issues 9 judgment, please. RLIT0000005. Page 87, please, and 10 pick up at paragraph 316: 11 "He [that's Mr Godeseth] had given information in 12 his first witness statement, namely that in Legacy 13 Horizon, any transactions injected by [the] SSC would 14 have used the computer server address as the counter 15 position which would be a number greater than 32, so it 16 would be clear that a transaction had been injected in 17 this way by someone other than the [subpostmaster]. 18 This is important because it would be consistent with 19 the case originally advanced by the Post Office that any 20 such injections would be entirely visible as having been 21 done externally (ie, not within the branch) due to the 22 counter number used. 23 "317. However, this important incorrect, and was 24 corrected by both Mr Godeseth and Mr Parker in 25 subsequent statements before they were called and as 76 1 a direct result of Mr Roll's evidence. The information 2 that was incorrect, and therefore had to be corrected, 3 had come directly from Mr Jenkins. This shows that 4 Mr Jenkins did, in at least one very important respect, 5 give Mr Godeseth directly incorrect information about 6 the visibility of injected transactions, which not only 7 could have an effect on branch accounts, and whether 8 this would show (or rather, not show) that the impact on 9 those accounts had come from the injections made outside 10 the branch. Mr Godeseth only found out the true 11 position when Mr Parker was preparing his subsequent 12 witness statement in the weeks prior to the commencement 13 of the Horizon Issues trial, in other words in 2019. He 14 had not known that before. His explanation" -- 15 A. Sorry. I don't think that is quite an accurate 16 representation of what actually happened in the -- 17 Q. Just let me get to the end of the paragraph? 18 A. Sorry: 19 "He had known that before. His explanation was as 20 follows", and then he sets out an explanation. 21 Firstly, is it right that you told Mr Godeseth that 22 any transactions injected by the SSC would have used the 23 computer server address at the counter position using 24 a number greater than 32? 25 A. That was my understanding when I was having 77 1 a conversation with him -- I can't remember the exact 2 month now, I think it was probably August or September 3 in 2018, something like that -- when he was preparing 4 his first statement. 5 Q. Why were you having that conversation with him? 6 A. To explain my understanding of remote access at that 7 time. 8 Q. Why were you having to explain to him your understanding 9 of remote access? 10 A. I think I was commenting on a witness statement that he 11 was putting together for the GLO. 12 Q. So is this right: as a result of informal chat, you told 13 Mr Godeseth something that turned out not to be true, he 14 included it in a witness statement in the High Court? 15 A. I realised my mistake soon after he'd actually put his 16 witness statement in and I then attempted to correct my 17 mistake shortly after that. And I sent a -- I produced 18 a document about two or three weeks after his final 19 draft of his first witness statement, pointing out that 20 I'd realised that that was actually incorrect, and 21 I sent that to the various lawyers and to Torstein and 22 also, probably, to Steve Parker. 23 Q. Did your change in approach come about because of you 24 reading Richard Roll's witness statement? 25 A. Yes, and as a result of that, I had a conversation with 78 1 John Simpkins, which pointed out the error in my 2 previous understanding. 3 Q. We've seen Mr Henderson say that you said to him in 2012 4 that Fujitsu could remotely access branch terminals and 5 insert transactions which would be indistinguishable for 6 those in the branch? 7 A. And we've seen him say that, yes. 8 Q. You subsequently claimed otherwise in your emails that 9 we've looked at? 10 A. Yes. 11 Q. It appears when you gave your initial account to 12 Mr Godeseth, for the purposes of him writing a witness 13 statement, you maintained that position? 14 A. Yes, it was a genuine mistake. 15 Q. So is it the case that it turns out that what 16 Mr Henderson says you admitted to him in 2012 was, in 17 fact, in general terms true all along? 18 A. I don't believe that I would have said that at the time, 19 but I just can't remember the conversations that we had. 20 It was my genuine belief that injections of transactions 21 were done at the correspondence server, rather than at 22 the counter. 23 Q. Thank you. Can we turn to a new topic, please: the use 24 and reliability of ARQ data. You deal in various points 25 in your third witness statement with ARQ data. Can we 79 1 turn those up, please. Paragraph 127, which is on 2 page 41. 127, you say: 3 "The evidence before the Inquiry has shown that 4 issues sometimes arose in respect of the audit data." 5 You deal with the issues in greater detail below: 6 "... however, there is a distinction I think it is 7 important to be clear about at the outset. This is the 8 distinction that exists between (a) data which actually 9 gets committed to the audit server (which becomes the 10 audit data), and (b) data being affected by a problem 11 before it is committed to the audit server (which, at 12 that stage is not audit data)." 13 Then paragraph 128, you say: 14 "... overall the issues which have been considered 15 by the Inquiry do not change my mind that the audit data 16 (including the raw audit trail and the NT events logs) 17 was a highly reliable and effective means of diagnosing 18 system issues. I would still agree with Mr Justice 19 Fraser's judgment that audit data (once the true scope 20 of that term is understood) was a gold standard for this 21 purpose. The relevance of audit data not including 22 information generated by POLSAP or Credence ... is not 23 clear to me, since neither system was used by 24 [subpostmasters] in balancing their accounts." 25 Some general questions to start with then, please. 80 1 Did you, in general terms, consider that the ARQ data 2 that you were provided with, in each case that you were 3 asked to provide a witness statement, was sufficient to 4 enable you to reach conclusions about whether 5 a particular branch had or had not experienced issues 6 with Horizon? 7 A. It's not just the ARQ data; it's looking at the 8 associated NT event logs that come with it. 9 Q. Yes, so ARQ data plus NT event logs? 10 A. Yes. 11 Q. Did you consider that both species of data were 12 sufficient for you to reach reliable conclusions about 13 whether a particular branch had or had not experienced 14 issues with Horizon? 15 A. I was looking -- right. I think, with the use of the NT 16 events, then, yes, the ARQ data would be reflecting what 17 had actually been recorded at the branch and, if there 18 was a problem in recording it, then that would have been 19 visible in the NT events, so yes. 20 Q. So is the answer to my question "yes" then? 21 A. I think so yes. 22 Q. Has your view about adequacy of any such data in any way 23 changed, in the light of any of the evidence given to 24 the Inquiry by any of the Fujitsu witnesses? 25 A. I believe that it is also useful to be able to look at 81 1 the raw NT event logs as well, and I usually had access 2 to those raw event logs. 3 Q. Did you look at them? 4 A. Yes. At least when I was given evidence -- I mean, 5 there was clearly some cases in 2012 where it was 6 decided that there was no need to look at the ARQ data 7 at all. 8 Q. In some cases, in your witness statement, I think you 9 made clear that you hadn't seen the ARQ data and, 10 therefore, couldn't comment on the cause of the 11 shortfall? 12 A. Correct. 13 Q. In other cases in which you provided a witness 14 statement, you didn't explicitly point that out? 15 A. That is correct. I was expecting that, if there were 16 some specifics to be looked at, then I would be shown 17 the ARQ data. But, clearly, I -- that didn't happen 18 in -- 19 Q. In those other cases, in which you provided a witness 20 statement but didn't state that you were unable to 21 comment on the cause of the shortfalls without having 22 seen the ARQ data and NT event logs, should we take it, 23 nonetheless, that you felt able to offer definitive 24 views on the causes of shortfalls, without having seen 25 such data? 82 1 A. I think I was giving a general view of how I thought -- 2 how the system should be operating. I was expecting to 3 have some sort of comeback in terms of specifics, if 4 that was appropriate, and clearly in some of those cases 5 it didn't happen. 6 Q. Who were you expecting to give some sort of comeback? 7 A. The Post Office Investigators. 8 Q. Why did you make it clear, in some witness statements, 9 for example the Grant Allen one, that you were unable to 10 offer a view as to the cause of specific shortfalls in 11 that case, without having seen the ARQ data but not in 12 others? 13 A. Because I'd actually been given some specific 14 information there as to something that might be the 15 cause and I was agreeing that that could be a cause, and 16 that was something that I would have needed to 17 investigate. 18 Q. Did it ever occur to you to refuse to provide a view, 19 unless and until you were provided with all of the data 20 that you considered necessary in order to answer the 21 questions which you have been asked to address? 22 A. Well, taking that specific case of Grant Allen, I was 23 certainly concerned about that, and not only did I check 24 with the lawyers from Post Office that it was okay to 25 sign that statement, I also checked it with the lawyer 83 1 within Fujitsu that it was okay to sign that statement 2 before I did so. 3 Q. Did you assume that those at the Post Office responsible 4 for pursuing the criminal cases, ie investigating or 5 prosecuting them, who were asking you to provide reports 6 and witness statements, were aware of the fact that you 7 didn't automatically have access to ARQ data and NT 8 event logs in every case? 9 A. I had understood that it was the normal thing to 10 actually have ARQ data. It wasn't until later that 11 I realised that they didn't always provide it. 12 I thought that was the whole purpose of ARQ data, was to 13 have some sort of information to support any sort of 14 prosecution or investigation. 15 Q. Dealing with the distinction that you make in 16 paragraph 127 -- if we just scroll up -- between data 17 which is committed to the audit server and that which 18 isn't -- 19 A. Yes. 20 Q. -- if data is affected by a problem before it is 21 committed to the audit server, would that mean that 22 there is a risk that, once the data is committed to the 23 server, it's unreliable? 24 A. If taken into account with the NT events, then that 25 would make it clear that that has actually happened. 84 1 So, as long as you look at the NT events as well as the 2 ARQ data, then I felt it would be reliable. 3 Q. I don't think you addressed that point in any of your 4 reports or witness statements or any of your advices or 5 explanations to the Post Office? 6 A. I was expecting to be asked to look at the -- at the 7 audit data, which therefore would have been doing the 8 checks against the NT event data, but I accept that that 9 didn't happen. 10 Q. So again, the reason for not mentioning it is you were 11 expecting somebody to come back to you about it? 12 A. Correct. 13 Q. When you were looking at audit data which had been 14 committed to the server, how could you be confident that 15 that data hadn't been affected by a problem before it 16 got to the server? 17 A. Because I know -- I was aware that the process for 18 generating -- for generating the ARQ data would include 19 checking for unexpected NT events at the same time, 20 which should therefore have detected any such failure. 21 Q. Can we look, please, at POL00097480, and go to page 3, 22 please, and scroll down. This is quite a complicated 23 email chain between you and Helen Rose about the Lepton 24 report and the logs produced as a result. If you can 25 try and and help us distinguish or understand what's 85 1 being said to start with. You refer to some files and 2 you say: 3 "[They] are part of the standard ARQ returned." 4 A. Yes. 5 Q. Then in the third paragraph you say: 6 "If that [which is what you described in 7 paragraph 2] is sufficient for your purposes, then you 8 do have all you need in the standard ARQ. 9 "However what I was able to confirm from my look at 10 Live data a couple of weeks ago and is also held in the 11 underlying raw logs is confirmation that the reversal 12 was generated by the system (and not manually by the 13 user). What might also be available in underlying logs 14 is whether or not the system was rebooted -- I suspect 15 it was but have no evidence one way or the other ... 16 I can confirm that the [other] user did log on again ... 17 "Do you need anything further from me ..." 18 To summarise, you were explaining here, the 19 limitations of what was included in standard ARQ data; 20 is that right? 21 A. Yes. 22 Q. You were explaining some additional data that you were 23 able to find by looking at what you call Live data -- 24 A. Yes. 25 Q. -- which in fact revealed that the relevant transaction 86 1 was one not undertaken by the subpostmaster but was one 2 generated wrongly by Horizon itself? 3 A. No, it wasn't generated wrongly by Horizon itself; it 4 was one that was generated by Horizon itself as part of 5 the recovery process following a failure of the system, 6 and the whole purpose of that reversal being undertaken 7 by the system was to reflect what the postmaster would 8 have understood to have occurred at the time of the 9 failure. 10 Q. I see. So it was an automated process where the system 11 generated the reversal? 12 A. To reflect what should have been understood by the 13 subpostmaster in the branch at that particular time. 14 Q. Okay. Then if we scroll up, please, we see Helen Rose 15 replies: 16 "I can see where this transaction is and now 17 understand the reason behind it. My main concern is 18 that we use the basic ARQ logs for evidence in court and 19 if we don't know what extra reports to ask for then in 20 some circumstances we would not be giving a true 21 picture." 22 That's right, isn't it: that if they were using, in 23 court cases the basic ARQ logs, in some cases a true 24 picture would not be given to the court? 25 A. I would dispute that because, in fact, that reversal was 87 1 part of providing the true picture. The reason that 2 reversal had actually taken place was to reflect what 3 the subpostmaster would have seen in the branch at the 4 time. 5 Q. Well, hold on, if we go up to your -- well, hold on. We 6 should read the second paragraph first? 7 "I know you are aware of all the Horizon integrity 8 issues. I want to ensure that the ARQ logs are used and 9 understood fully by our operational staff that [we] have 10 to work with this data both interviews and in court." 11 What did you understand her to mean by saying to you 12 that you were aware of all of the Horizon integrity 13 issues? 14 A. She -- I'd been sent, a few months earlier, a report 15 that she'd produced of branches that were challenging 16 the integrity of Horizon. So I assumed that she was 17 referring to that report of hers, which -- which I'd 18 been sent a copy of. 19 Q. So not meaning you're aware of all of the integrity 20 issues with Horizon that in fact exist but, instead, 21 meaning all of the allegations made by subpostmasters 22 about Horizon integrity issues? 23 A. That is how I would have understood that statement to 24 have been made. 25 Q. She says: 88 1 "... I want to ensure that the ARQ logs are used and 2 understood by operational staff that have to [use them] 3 in interviews and in court." 4 If we go back up to your reply, please. You say 5 you: 6 "... understand your concerns. 7 "It would be relatively simple to add an extra 8 column into the existing ARQ report spreadsheet, that 9 would make it clear whether the Reversal Basket was 10 generated by Recovery or not. I think this would 11 address your concern." 12 So was it the case that the existing ARQ data being 13 relied on by investigators and prosecutors did not give 14 a full picture; you were less thing a solution to that? 15 A. It was giving a full picture of what had occurred. 16 There were some -- these system recoveries were purely 17 to actually put the branch accounts back into the state 18 that they should have been in following a failure. It's 19 a fairly complicated scenario to try to describe. 20 I don't know if you want me to get into the 21 technicalities, I suspect not, but it was all to do with 22 what could happen when a comms failure occurs at the 23 wrong time. 24 Q. You don't say in reply "Hold on, Helen [or Ms Rose], 25 you're wrong, the data you have does present a full 89 1 picture". You say there needs to be a change in order 2 for it to do so? 3 A. What I was saying was that, if she wanted to be aware of 4 those things specifically, then that could be added into 5 the ARQ -- into the standard ARQ extract. 6 Q. Was that done -- 7 A. I don't -- 8 Q. -- to your knowledge? 9 A. Not as far as I'm aware. 10 Q. Why wasn't it done, to your knowledge? 11 A. Because it required Post Office to request a change on 12 Fujitsu to do that and I don't believe they pursued it. 13 Q. So it was a Post Office decision? 14 A. That -- yes. 15 Q. Do you know why the Post Office did not pursue it? 16 A. No. 17 Q. Thank you. Can we move on, please, to your integrity 18 reports. 19 A. Yes. 20 Q. Thank you. That can come down. Just looking at it 21 generally, to start with, you authored two reports 22 addressing the integrity of data on Horizon? 23 A. I think they were narrower than that, in terms of what 24 their scope was. But I know the reports you're 25 referring to. 90 1 Q. The first -- excuse me. I've got a frog in my throat. 2 It's hayfever. 3 SIR WYN WILLIAMS: Take your time, Mr Beer. 4 MR BEER: You authored two reports, the first, 2 October 5 2009; do you remember? 6 A. I remember that report, yes. 7 Q. The title of that was "Horizon Data Integrity"? 8 A. Yes. 9 Q. That was about Legacy Horizon? 10 A. Correct. 11 SIR WYN WILLIAMS: Would you prefer an early lunch? It's 12 a 12.45? 13 MR BEER: That's most generous, sir. 14 SIR WYN WILLIAMS: Right. What time shall we resume? 15 MR BEER: Can we say 1.35, please. 16 SIR WYN WILLIAMS: Right. Fine. 17 (12.44 pm) 18 (The Short Adjournment) 19 (1.35 pm) 20 MR BEER: Thank you, sir. I think I'm back up to full 21 operating speed. 22 SIR WYN WILLIAMS: Good. 23 MR BEER: Mr Jenkins, we were going to look at the two 24 Horizon data integrity reports that you wrote and I was 25 summarising that you authored two reports, the first of 91 1 which was dated 2 October 2009, and the title of which 2 was "Horizon Data Integrity" -- 3 A. Correct. 4 Q. -- and that was about Legacy Horizon? 5 A. Correct. 6 Q. The second of which was dated 25 November 2011 and was 7 called "Horizon Online Data Integrity", and that was 8 obviously about Horizon Online? 9 A. There were various versions of it with different dates, 10 but, yes, that was the date of one of them. 11 Q. The last version, I think, that we've got is version 4 12 and is dated 25 November 2011? 13 A. There was a variant on that that I used later on in 2012 14 but, yes. 15 Q. Can I ask you some general questions about those 16 reports, then, before looking at the contents in detail. 17 Firstly, at the time that you prepared the first report, 18 the one about Legacy Horizon, called "Horizon Data 19 Integrity", what did you understand the purpose of the 20 report was? 21 A. I was asked to provide a report on some specific failure 22 scenarios that occurred on Legacy Horizon. I couldn't 23 remember initially what it was that had caused that but 24 when Dave Smith gave his evidence to the Inquiry, he 25 says that he phoned me up and instructed me to do that, 92 1 and I was happy to accept that. I've since seen some 2 emails saying that I was asked by people in Fujitsu on 3 his behalf to write that report but I don't think that 4 really contradicts what I said before. 5 Q. Did you therefore understand that you were being asked 6 to produce a report that was narrow in scope about very 7 specific issues? 8 A. Correct. 9 Q. In relation to the Horizon Online data integrity report, 10 at the time that you prepared that, what did you 11 understand its purpose was? 12 A. Initially, it was supposed to be a companion document to 13 that. It then got subsumed into documents to support 14 a proposed investigation that Fujitsu were having with 15 KPMG into the integrity of the audit trail. 16 Q. So originally it was a companion document to the -- 17 A. But with a similar sort of scope, in other words looking 18 at primarily at what -- the integrity of the audit 19 trail. 20 Q. Why did the reports not cover all bugs, errors and 21 defects known to Fujitsu at the time that they were 22 written? 23 A. Because that wasn't what I was asked to produce at the 24 time. 25 Q. Do you know why you weren't asked to produce a report 93 1 that was addressing all known bugs at the time? 2 A. No, I was asked to produce a report covering the scope 3 that the report actually covered. 4 Q. Were you aware, whilst employed by Fujitsu, that the 5 Legacy Horizon report was relied on by Mr Ismay of the 6 Post Office in preparing his report in August 2010 about 7 the general robustness of Horizon? 8 A. No, I was not. 9 Q. You didn't know it was going to be appropriated and used 10 for a different purpose than you had prepared it for? 11 A. No, I didn't. 12 Q. Have you seen now, in fact, how it was used by -- 13 A. Yes, I have seen Mr Ismay's report, as part of the 14 documentation I've been provided for the Inquiry. 15 Q. Do you agree that it has been used for a purpose 16 different from that which you intended? 17 A. Yes, it has been. 18 Q. Were you aware at the time that you wrote the Online 19 report that the Post Office may intend to use that 20 report as a basis for briefing or addressing concerns 21 raised by MPs? 22 A. When I wrote the Online report, I don't think it was 23 prepared for the purpose of the Post Office, that the 24 KPMG audit was not something that Post Office was 25 necessarily aware of at the time. When I produced the 94 1 cutdown version of it in 2012, I saw that purely as 2 being a description of the integrity of the audit trail, 3 and I wasn't -- I think there was an email exchange 4 wanting a report, effectively the equivalent of the 5 Legacy Horizon one, and I thought it was actually to do 6 with prosecutions, rather than anything to do with 7 briefing of MPs, or anything like that. 8 Q. Were you told at any time that the Post Office intended 9 to use your report as a basis to address concerns raised 10 by MPs? 11 A. I don't think I was even aware of that now! 12 Q. Can we look at the detail, then, of each of the two 13 reports. Can we start, please, with the Legacy Horizon 14 integrity report, data integrity report. FUJ00080526. 15 Thank you. We can see the document's title is "Horizon 16 Data Integrity", and we can see on this page here that 17 you are the author of it? 18 A. Correct. 19 Q. Correct. If we just go over to page 2, please, and 20 scroll down -- and over the page, and scroll down -- 21 after the index we can see the document history. Can 22 you see the version that we're working from here is 23 version 1.0, dated 2 October 2009? 24 A. Correct. 25 Q. If we just look at the foot of the page, we can see 95 1 that. That's on every page, bottom right; can you see 2 that? 3 A. Yes. 4 Q. Just go back up to the top of the page. Just out of 5 interest, version 0.1b, which appears before version 1 6 is described as a first informal draft that is dated 7 a year after the final version for release? 8 A. I think that must be a typo. 9 Q. Should that be -- 10 A. That should be 2009. I'd not noticed that until now. 11 Q. Okay. So does that reflect that you created the report 12 and finalised the report all on one day? 13 A. I think I -- I think the version 0.1a may have been 14 produced a day or two before that but, yes, the whole 15 report was produced in a fairly short timescale. 16 Q. We can see who is required to review it, "Mandatory 17 Review: Suzie Kirkham and Jeremy Worrell", both Fujitsu? 18 A. Yes. 19 Q. Then "Optional Review", are the first three of those 20 Fujitsu? 21 A. Yes. 22 Q. The fourth them is Mr Smith from the Post Office; is 23 that right? 24 A. Correct. 25 Q. Do you know whether he in fact reviewed it? 96 1 A. I can't remember now. 2 Q. It's not marked up with the asterisk, if you look at the 3 bottom of the page, which indicates that the reviewer 4 actually returned comments. 5 A. Yes. I'm not sure whether it had been sent to him but 6 the thing was -- I was asked to actually get it signed 7 off and reviewed quickly. So whether he'd actually seen 8 it or not, I'm not absolutely certain. 9 Q. If we go to the top of the page, and this appears on 10 every page, "Commercial in Confidence and Without 11 Prejudice". Why was it marked up "Without Prejudice"? 12 A. Someone asked me to put that there. I've no idea what 13 that actually means. I believe it's some sort of legal 14 thing but I don't understand what it means. 15 Q. Who asked you to put that there? 16 A. I can't remember now. 17 Q. Were you told that this was being prepared in the 18 contemplation of any litigation? 19 A. No, I don't believe so. 20 Q. Were you told that this was prepared for the purposes of 21 obtaining legal advice? 22 A. I don't believe so. 23 Q. Did anyone mention to you before, or as you were 24 preparing the document, that it was to be used or might 25 be used in litigation that was reasonably in 97 1 contemplation or for the purposes of obtaining legal 2 advice? 3 A. Not that I'm aware of. 4 Q. Can we go forward to page 5, please. We can see the 5 purpose of the document, it is: 6 "... submitted to the Post Office for information 7 purposes only and without prejudice." 8 You put that? 9 A. Again, I don't understand what "without prejudice" means 10 so someone must have said I needed to add that in but 11 I can't remember who. 12 Q. Would that be a lawyer? 13 A. I don't know. 14 Q. "In the event that the Post Office requires information 15 in support of a legal case, Fujitsu will issue a formal 16 statement." 17 Again, you wrote that? 18 A. And I think someone must have told me that I needed to 19 include that in the document. 20 Q. Do you know why: that this report couldn't be relied on 21 in support of a legal case; if there was a legal case, 22 then Fujitsu would issue a formal statement? 23 A. No, I'm not aware of why I was asked to include that in 24 the document. 25 Q. "This document is a technical description of measures 98 1 that are built into Horizon to ensure data integrity, 2 including a description of several failure scenarios, 3 and descriptions as to how those measures apply in each 4 case. 5 "[The] document only covers Horizon. It does not 6 cover [Horizon Online]." 7 Was that dictated to you or was that you writing? 8 A. That was me. 9 Q. So it describes measures that are built into Horizon 10 Online to ensure data integrity? 11 A. I think you mean Legacy Horizon. 12 Q. Sorry? 13 A. I said I think you mean Legacy Horizon. 14 Q. Did I say Horizon Online? 15 A. Yes. 16 Q. I'm so sorry, Legacy Horizon. 17 A. Yes. 18 Q. To ensure data integrity; is that right? 19 Then if we go over the page, please, there is, under 20 the heading, "Horizon Data Integrity" a description in 21 summary terms of data processing within Horizon; is that 22 fair? 23 A. Yes. 24 Q. Then, if we scroll down, please, after that summary, you 25 set out scenarios -- 99 1 A. Yes. 2 Q. -- and you say: 3 "It should be noted that these scenarios are all to 4 do with equipment failures and these will always be 5 visible to Fujitsu through event logs which are 6 retained." 7 A. Yes. 8 Q. The possible failure scenarios that you then discuss in 9 the report deal, essentially, would you agree, with 10 equipment failures -- 11 A. Yes. 12 Q. -- and not software failures? 13 A. Yes. 14 Q. So, for example, the Callendar Square bug is not 15 mentioned, is it? 16 A. No, it isn't. 17 Q. Nor the remming out bug? 18 A. No. 19 Q. Would you agree that, if this document was taken as 20 providing a picture of Horizon's good health, it only 21 gave a partial picture because it only dealt with 22 equipment failure? 23 A. That was only what I was actually asked to produce 24 a report on, so the scope was given to me and I believe 25 those scenarios were actually provided to me by Post 100 1 Office. 2 Q. Would you agree that a full and complete explanation of 3 Horizon data integrity would not only focus on equipment 4 failures? 5 A. I can see that now but, as I say, that wasn't the scope 6 that I was asked to produce this document for. 7 Q. Did you say, "Well, hold on, if I'm writing a report 8 about Horizon data integrity, it should be more than 9 about pieces of equipment that may fail"? 10 A. My -- I'd understood that that was what the report was 11 required to cover. 12 Q. Do you know why a narrow focus was selected? 13 A. No, I don't. 14 Q. Did you get any sense that a narrow focus was selected 15 to provide a commercially acceptable reality that 16 Fujitsu could present to the Post Office, which only 17 discussed equipment issues, without referring to bugs, 18 which, if mentioned officially in a report in a long 19 series, might lead to a reappraisal of the contract 20 between the parties? 21 A. That hadn't occurred to me at the time. As I say, I was 22 just asked to produce a document with this particular 23 focus, which is what I did. 24 Q. The information in the report, therefore, didn't reflect 25 a full picture of what you knew of concerns arising over 101 1 data integrity in Horizon. It's only about equipment 2 failures? 3 A. Yes. 4 Q. Can we go to POL00088935. Thank you. If we just look 5 at the top of the page, this is the Ismay Report of 6 2 August 2010. You're not on the distribution list -- 7 A. No. 8 Q. -- because it was a Post Office document? 9 A. Yeah. 10 Q. When did you first see it? 11 A. I think I first saw it as part of the preparation for 12 this Inquiry. 13 Q. Can we look, please, at page 7, under the heading 14 "Horizon": 15 "Horizon was developed as an Electronic Point of 16 Sale to replace archaic paper based accounting 17 processes. It utilised hardware in branches which had 18 strong security features and interfaces to a central 19 data centre with high security. Data was replicated 20 between units in the branch and once it reached the data 21 centres it was replicated between those. 22 "This gave three strengths which are important in 23 terms of allegations being made: 24 "1. Horizon infrastructure was robust from 25 a security and access perspective 102 1 "2. It was resilient in terms of being able to 2 continue customer service and hold data in a queue in 3 the event of incidents, and 4 "3. It had strong back up and integrity features 5 with data backed up in branch and centrally. 6 "This provided a strong audit trail. Further 7 narrative from Fujitsu is included at Appendix 3." 8 Then if we go to Appendix 3, which is page 17, can 9 we there see -- it is printed in a different way than 10 the document we just looked at -- but your Horizon data 11 integrity report for Legacy Horizon. 12 A. Yes. 13 Q. Is the entirety of the report set out there by Mr Ismay 14 as his Appendix 3? 15 A. I've not checked that it is entirety but I'm happy to 16 believe that it is. 17 Q. How did your report come to be included in the Ismay 18 Report? 19 A. I've no idea. 20 Q. Were you asked for your consent or permission for it to 21 be reused in this way? 22 A. No. 23 Q. To your knowledge, was anyone in Fujitsu asked if it 24 could be repurposed in this way? 25 A. I don't know. 103 1 Q. Were you consulted on any of the wider contents of the 2 Ismay Report? 3 A. No. 4 Q. To your knowledge, was anyone else in Fujitsu consulted 5 on the wider content of the Ismay Report to -- 6 A. Not to my knowledge, no. 7 Q. When he gave evidence to the Inquiry, Mr Ismay said that 8 he relied on your Legacy Horizon data integrity report 9 in drafting his own report and he accepted that, at the 10 time, he did not think that the Legacy Horizon data 11 integrity report had any limitations at all, and he 12 accepted that that was a failing. To your knowledge, 13 was anyone in Fujitsu, including yourself, involved in 14 any way in the use or the misuse of your report in the 15 Ismay Report? 16 A. Not to my knowledge, no. 17 Q. Thank you. Can we look, please, at FUJ00156064 and look 18 at page 3, please -- thank you, if we scroll up, 19 please -- from Suzie Kirkham; do you remember who she 20 was? 21 A. She was an account manager or something -- I think that 22 was the title that's given on the report. I have 23 a vague recollection of her but I didn't have very much 24 to do with her. I think my main interaction with her 25 was to do with this report. 104 1 Q. Who was Gavin Bounds? 2 A. He was also an account manager at some stage. How they 3 related to each other I just don't remember, I'm afraid, 4 but they were both sort of fairly high up compared with 5 me. 6 Q. She says: 7 "Gavin 8 "I have detailed below a brief summary of the 9 subpostmaster claim that Horizon fails to maintain the 10 integrity of branch accounts if the branch experiences 11 technical issues. 12 "Horizon has been running for around 10 years, 13 during which time Post Office has had to handle a number 14 of legal cases surrounding system integrity, mostly from 15 apparently disgruntled or fraudulent subpostmasters. 16 Fujitsu has twice appeared in court to support Post 17 Office. Only once has Post Office conceded its case, 18 and this was at a time when the audit trails were too 19 short to provide sufficient evidence to refute 20 a subpostmaster's claim. 21 "Recently a number of subpostmasters are again 22 querying the data integrity of Horizon and are claiming 23 that they are owed money by the Post Office due to 24 inaccurate branch accounts. These subpostmasters are 25 talking to the press and are trying to gain some 105 1 momentum to support their case -- it is even rumoured 2 that the BBC is considering a Watchdog programme on the 3 subject. Post Office is trying to head off this 4 possible escalation of interest by preparing a stock of 5 simple responses to possible questions from the media or 6 interest groups. I have summarised below the timeline 7 that details when Post Office asked Fujitsu for 8 assistance in compiling this information. 9 "The initial request from Post Office (Dave Smith 10 and David Gray -- POL) was made on 24 September asking 11 us to join a conference call ... hosted by Dave Smith 12 where Dave would outline his understanding of the 13 situation and outline his response. Fujitsu would only 14 comment if he made any factual errors on how Horizon 15 works. 16 "I [that's Suzie] spoke to Mike Wood, Dave Roberts 17 and our Press Office as soon as the request was made. 18 [Fujitsu] Press Office confirmed with [Post Office] 19 Press Office that the journalist would not be present at 20 this conference call and that it was only internal [Post 21 Office] attendees. They also confirmed that they did 22 not want [Fujitsu] responding to any questions from the 23 press or external organisations. This was a Post Office 24 issue as the solution has been formally handed over to 25 Post Office. They only need Fujitsu for 'expert' advice 106 1 if required." 2 Were you involved in any of those first two 3 paragraphs? 4 A. I don't believe so. 5 Q. Friday, the 25th, the conference call was completed; 6 Fujitsu's attendees, are set out, not including you? 7 A. Correct. 8 Q. Then over the page: 9 "During the conference call [the Post Office] 10 requested from Fujitsu a short paper to describe how 11 Horizon maintains the integrity of branch accounts when 12 certain issues affect the branch, eg blue screen, 13 hardware failure. Report requested to be given to Dave 14 Smith by 2 October." 15 That's, in fact, the date of your report? 16 A. Correct. 17 Q. "Gareth Jenkins ... spent the best part of one day 18 completing the report." 19 Is that right: you spent a day writing it? 20 A. It was something like that because there was fairly 21 short timescales for actually putting that together, 22 yes. 23 Q. "The finished report was passed to Amanda Craib 24 (Commercial -- at David Roberts' request) before being 25 sent to Dave Smith. Report was sent to Dave Smith on 107 1 Friday, 2 October late morning just prior to a second 2 conference call. 3 "Friday, 2 October -- second conference call -- Dave 4 thanked Fujitsu for the document -- no further actions 5 ... 6 "[She] spoke to Andy McLean on 5 October -- he 7 didn't expect any further activity for at least a couple 8 of weeks. 9 "[She] spoke with Dave Smith on 3 December and he 10 mentioned that the data integrity issue had reappeared. 11 It has been taken up by a Welsh MP who also happens to 12 be a barrister. He is threatening to bring an Early Day 13 Motion in the House. Dave is meeting him in Parliament 14 on 10 December but at this stage [he] didn't feel he 15 needed any [assistance]. However, he was alerting us to 16 the fact that this has reappeared. 17 "... our body of evidence is mature, so we're well 18 placed to help Post Office today. But when 'Horizon 19 Online' is installed in 2010, much of our existing 20 material will become irrelevant", et cetera. 21 You were copied in on this, weren't you? 22 A. I was. 23 Q. So I think you knew due that the data integrity issue 24 had not gone away as a result of your paper and there 25 was a possibility that it was going to be raised with 108 1 an MP in the future? 2 A. I must have read that in the report. I can't remember 3 that now but, yes, I accept that I was copied on this 4 ale. 5 Q. Yes, and then if we scroll up, please -- and up the 6 page, just stop there -- some internal emails not 7 including you. Then scroll up, please, keep going, and 8 keep going, and then stop there. Ann Sinclair forwards 9 that chain to Gavin Bounds. What did Ann Sinclair do? 10 A. I've no idea. I've not come across the name before. 11 Q. Ann says: 12 "... following discussions with our press folks and 13 Suzie on Monday [night] we have agreed that: 14 "We need to downplay [the issue] as much as, 15 possible, we don't want to guilty drawn into the debate 16 as we are only a supplier. 17 "There is a document that addresses all the 18 technical questions that was produced for [Post Office] 19 for a meeting with MPs last week ..." 20 Is it right that there was a list of hardware issues 21 that was produced that you were required to address? 22 A. That is my understanding and that is, effectively, what 23 section 3 of the document was all about. Someone gave 24 me that list. 25 Q. So that reference there, there is a document that 109 1 addresses all the technical questions. That's 2 a reference back to your report of 2 October? 3 A. I'm assuming so but I don't know for a definite fact. 4 Q. I mean, it says that it was "produced for [Post Office] 5 for a meeting with MPs last week"; do you think that was 6 a different document? 7 A. It could be. As I say, I produced on it 2 October and 8 I think now we're in December, so it could be something 9 different. I don't know. 10 Q. Did you prepare a second or supplemental document -- 11 A. No. 12 Q. -- for the purposes of a meeting with MPs? 13 A. No. 14 Q. Thank you. That can come down. 15 In short, then, you say that your paper was not 16 a report about or a survey of all issues with Legacy 17 Horizon. 18 A. Correct. 19 Q. Again, did anyone explain to you why, if questions were 20 being raised about the integrity of data produced by 21 Horizon, you were not asked to produce a report, which 22 was a survey of all of the issues which had affected 23 Legacy Horizon? 24 A. I was asked to produce a report on the fairly narrow 25 basis of those hardware type issues and, as I say, 110 1 I believe that someone actually gave me a list of the 2 ones to cover and those were the ones that I covered in 3 section 3. 4 Q. So it was sought for a limited purpose, its contents 5 fulfil that limited purpose and it shouldn't be used for 6 any other purpose? 7 A. That was the situation at the time, yes. 8 Q. Can we turn to the data integrity report for Horizon 9 Online, please. FUJ00080534. We see the title again, 10 "Horizon Online Data Integrity". We see again you're 11 the author? 12 A. Yes. 13 Q. If we go to the foot of the page, we can see the date of 14 this version, 25 November 2011, version 4. 15 A. Yes. 16 Q. If we go over the page, please, and, again, past the 17 index, sorry, to "Document History", we can see that the 18 first draft was created on 18 January that year and then 19 if we scroll down -- I'm not going to go through the 20 summary of changes -- version 4, which I think is the 21 one approved for distribution is 25 November 2011. 22 A. Yes. 23 Q. Did the purpose of the report change in those ten months 24 or so? 25 A. Yes, it got -- one of the versions, I can't remember 111 1 which of the versions, was -- there was talk about 2 having a formal review of the integrity of the audit 3 trail of Horizon Online by an external firm of 4 consultants. 5 Q. KPMG? 6 A. KPMG, and the report was amended to actually effectively 7 be a scoping document for what they were being asked to 8 provide. 9 Q. What was its original purpose? 10 A. I think its original purpose was to be a companion 11 document with a similar sort of scope to the Legacy 12 Horizon version. 13 Q. Why did you need a companion document about hardware 14 failures in relation to Horizon Online? 15 A. Well, it wasn't just the hardware failures; it was also 16 talking about the way that the audit trail worked and so 17 this was talking about how the audit trail worked on 18 Horizon Online because that was very different from 19 Legacy Horizon. 20 Q. Why did you need a companion document about how the 21 audit trail worked? 22 A. I -- 23 Q. Why did Fujitsu need one? 24 A. I can't remember exactly why I was asked to write this 25 at the time now. 112 1 Q. I was about to ask: who asked you to write it and why? 2 A. I can't remember, I'm sorry. There maybe emails about 3 it but I've not seen them as part of the Inquiry. 4 Q. Can we go to page 7, please, terms of reference in bold 5 at the top: 6 "Fujitsu would like to instigate an independent 7 audit of the [Horizon Online] environment currently 8 delivered to Post Office to provide confidence that the 9 solution has intrinsic security controls commensurate 10 with the requirement for legal admissibility. This 11 would enable a legal review of contract compliance." 12 Just trying to decode what that says, is that saying 13 that this report will show that Horizon Online has 14 intrinsic security commensurate with the requirement for 15 legal admissibility or the report that's going to be 16 prepared by KPMG. 17 A. The latter. So this was effectively the terms of 18 reference for KPMG to produce their report, which never 19 actually happened. 20 Q. Whilst you've said that, why didn't it happen, to your 21 knowledge? 22 A. I don't know why it didn't happen. I think I've 23 understood since then that the Second Sight review 24 subsumed it but I think the scope of the Second Sight 25 review was actually very different from what this was 113 1 intended to be. 2 Q. Who gave you to understand that the Second Sight piece 3 of work was the surrogate for a Fujitsu commissioned 4 review that would provide confidence that Horizon Online 5 had security controls commensurate with the requirement 6 for legal admissibility? 7 A. It probably -- I'm not sure if I was told explicitly, it 8 could well be an assumption on my part, that this all 9 seemed to go away at the same time that the Second Sight 10 review started up. 11 Q. What about the suggestion that this will enable a legal 12 review of contract compliance? To your understanding, 13 did Fujitsu engage in that process at all, a legal 14 review of contract compliance? 15 A. My understanding is that that was what KPMG were being 16 asked to do but that never happened. 17 Q. Did you get told at the time that neither of those 18 things happened because of Second Sight? 19 A. I don't think I was told that explicitly but I can't 20 remember. 21 Q. All you knew is that they didn't happen? 22 A. Correct. 23 Q. At paragraph 1.1 "Objective": 24 "Now that Horizon Online has been operational for 25 12 months, Fujitsu is undertaking a legal review of its 114 1 compliance with its contract obligations and in order to 2 enable that, would like to undertake an independent 3 assessment to demonstrate the adequacy of the security 4 controls that have been designed into the system to 5 provide assurance in the robustness of the audit of the 6 transactional data that may be used as evidence in 7 court." 8 So, just stopping there, did you understand that the 9 review to be undertaken by the third party, I'm going to 10 call them KPMG -- 11 A. Yeah. 12 Q. -- to be undertaken by KPMG, had as its focus assurance 13 in the robustness of data used as evidence in court? 14 A. That is what I understood was the purpose of that. So 15 there were a number of people involved in the 16 discussions with KPMG, and this document was just 17 something that I was the author of but a lot of the 18 input was from the team who were working on the 19 discussions with KPMG. I remember going to two or three 20 briefing meetings with KPMG with other people within 21 Fujitsu as well. 22 Q. Would you agree that the purpose of providing assurance 23 in the robustness of the audit of transactional data 24 that that may be used as evidence in court, is quite 25 a broad Statement of Purpose? 115 1 A. Yes, I've probably not really thought that through 2 particularly but, as I say, those are not necessarily my 3 words but I accept that they're included in my document. 4 Q. "The purpose of this document [ie the one you were 5 writing, to continue] is to define the terms of 6 reference for the project and to provide a technical 7 description of measures that are built into Horizon 8 Online to ensure data integrity." 9 Those are your words; is that right? 10 A. Those are my words, yes. 11 Q. Would you agree that that purpose, as described, is 12 a broad Statement of Purpose? 13 A. It is a fairly broad statement, yes. 14 Q. Would you agree that the existence of bugs, errors and 15 defects in Horizon may affect data integrity? 16 A. I don't think -- at the time, I wasn't aware of 17 outstanding bugs, errors and defects in Horizon Online. 18 Q. That's a different issue. 19 A. Right. 20 Q. Would you accept that the existence of bugs, errors and 21 defects in Horizon Online may affect data integrity? 22 A. I think what we were trying to show here was the 23 integrity of the audit trail, therefore, if there were 24 any bugs, errors or defects in the -- in how the system 25 operated, then it could all go back to looking at the 116 1 integrity of the audit trail from which any reports 2 produced by the system could then be traced back. 3 Q. I will try a third time: would you agree that the 4 existence of bugs, errors and defects in Horizon Online 5 may affect data integrity? 6 A. I don't think they did affect the data integrity in 7 terms of the integrity of the audit trail, which was the 8 focus of this document. 9 Q. Is that why the document does not address whether or not 10 there are bugs, errors or defects in Horizon Online? 11 A. Yes, because it was thought that by actually showing the 12 integrity of the audit trail, then that bypassed the 13 need to consider bugs, errors and defects in the other 14 parts of the system because everything goes back to the 15 data that's in the audit trail. 16 Q. Was explicit thought and consideration given to that, 17 that, if we focus on the audit trail, we can bypass the 18 existence of bugs, errors and defects? 19 A. "Bypass" might be a bad choice of words. The point was 20 that, having the limited focus of showing that the audit 21 trail had integrity was something that was practical to 22 actually be able to prove, as was -- which is what we 23 were effectively asking KPMG to do. 24 Q. Is that reflected on page 9, if we go forwards, please, 25 under "Purpose"? In the last paragraph there under 117 1 "Purpose": 2 "The scope of this paper is restricted to showing 3 the integrity of the audit trail and that it accurately 4 reflects the transactions entered at the counter." 5 A. Yes, that's effectively what I was trying to say before. 6 Q. You tell us in your witness statement, no need to turn 7 it up it's paragraphs 164 and 165, that: 8 "My report was intended to be included in briefing 9 materials for KPMG so that they had a high-level 10 technical overview as to how data was recorded in the 11 audit trail." 12 Is that right? 13 A. Yes. 14 Q. This report didn't consider when and how the audit 15 information was interrogated, did it? 16 A. No. 17 Q. It provided only a description of security controls? 18 A. I'm sorry, I'm not sure what you mean by "security 19 controls". 20 Q. Well, were you asked to conduct any research into the 21 operation of the controls that are mentioned in the 22 report, and whether they were in fact functioning 23 adequately on the ground? 24 A. This is what KPMG were being asked to look into. 25 Q. I see. So you didn't do that. You were describing how 118 1 the system was supposed to work -- 2 A. So that they could then prove that it did. 3 Q. It was down to them to prove or disprove, presumably, 4 whether, in fact, it was fit for purpose? 5 A. Correct. 6 Q. But that never happened? 7 A. Correct. 8 Q. So, again, nobody should take your report as being the 9 description of Horizon Online's overall integrity? 10 A. It was showing our belief of how the audit trail was 11 a way of showing that the system had integrity, yes. 12 Q. How it's supposed to work, in theory? 13 A. Yes. 14 Q. So the scope of the report was limited by those 15 commissioning it, correct? 16 A. Yes, I guess so, yeah. 17 Q. In the light of your knowledge of the system by this 18 time -- it's two years or so in, so November 2011 -- 19 A. Yes. 20 Q. -- and the bugs that had arisen in particular in the 21 course of development, did you ever question the limited 22 scope of the commission? 23 A. No, because I felt that, once the pilot -- I accept the 24 fact that there were some issues during the pilot phase 25 of Horizon Online but, once the pilot was over, 119 1 I believed that the audit trail of Horizon Online did 2 indeed the integrity that I was saying should be proved 3 as part of this report. 4 Q. But you had no evidence for that because that was what 5 KPMG were going to find out? 6 A. I had no independent evidence but we had found 7 empirically that it had the right -- that it had the 8 information there -- there for doing so. That we hadn't 9 had -- we hadn't found any outstanding problems in the 10 way that the audit trail operated and the audit trail 11 was seen to be accurately recording what had actually 12 happened within the branches, and we had no evidence to 13 the contrary. So, no, we hadn't proved it positively, 14 but we had nothing to actually say that it wasn't 15 correct. 16 Q. Why was it necessary for KPMG to investigate at all, 17 then? 18 A. I think the senior management within Fujitsu at the time 19 felt that it would be helpful to be able to show -- have 20 an independent proof, not just an internal assessment of 21 what was happening. 22 Q. If we go back to page 7, please, the top line of the 23 terms of reference includes the fact that the 24 independent report would seek to provide confidence that 25 Horizon Online has intrinsic security controls 120 1 commensurate with the requirement for legal 2 admissibility. 3 By this stage, you were aware that the data being 4 produced by Horizon, both in fact Legacy Horizon and 5 Horizon Online, was being used to support criminal 6 prosecutions? 7 A. Yes. 8 Q. You were, I suspect, in any event, aware that the data 9 being produced by both systems formed the core 10 accounting information for any subpostmaster's 11 business -- 12 A. Yes. 13 Q. -- and, indeed, the business of the Post Office itself? 14 A. Yes. 15 Q. So integrity of the data was absolutely critical, for 16 all three reasons, or all three purposes? 17 A. Yes. 18 Q. From their titles, these reports address Horizon 19 integrity in the Legacy system and Horizon Online data 20 integrity, but did not address everything that you knew 21 about Horizon data integrity either in the online system 22 or on -- in Legacy, did they? 23 A. I think they were trying to show that the audit 24 information did indeed have integrity and I believed at 25 the time and I still believe now that it does. 121 1 Q. For example, in relation to Legacy Horizon, they didn't 2 address faults in development or in the operation of the 3 Riposte system? 4 A. No, but I didn't think that was relevant in terms of the 5 integrity of the actual audit trail, which is what was 6 the basic subject that it was about. 7 Q. They didn't address any detail of the Riposte lock 8 problem, for example? 9 A. It doesn't do that. That wasn't the scope of the 10 document at the time. 11 Q. It didn't address all of the circumstances in which 12 a discrepancy might result and require reconciliation? 13 A. I'm not quite sure what you mean by that. 14 Q. Well, the raising of the issue by a subpostmaster, 15 reporting it, the issue of a transaction correction by 16 Fujitsu and the Post Office, it didn't address all of 17 that process? 18 A. I didn't see that as being the purpose of the documents. 19 The purposes of the documents were quite narrow, as 20 we've already discussed. 21 Q. If the work that KPMG was supposed to undertake was to 22 be presented to a court, would you agree that the 23 purpose would have been to satisfy the court as to both 24 the integrity and reliability of the Horizon system and 25 therefore the data that was being produced in the case 122 1 before it? 2 A. Yes, I think so. 3 Q. An inadequate substitute would be to produce to a court 4 your two reports? 5 A. I believe that that was covering the integrity of the 6 audit trail and it was based on information in the audit 7 trail that was the data that was being presented to the 8 courts. 9 Q. So you disagree with my suggestion. You think that 10 an adequate and proper substitute was to present only 11 your limited, constrained and narrow reports to a court, 12 as showing Horizon integrity? 13 A. That was how I understood things at the time. 14 Q. And now? 15 A. I think -- I think they still show the integrity of the 16 overall audit trail, which is the data that's being 17 presented to the courts. But I accept now that I don't 18 understand all the legal niceties, which -- and 19 I clearly didn't at the time either. But I wasn't aware 20 of my ignorance then. 21 Q. Can we, on that note, turn to the second part of my 22 questions to you, which concerns some of the case 23 studies -- 24 A. Okay. 25 Q. -- and can we start with the case brought against Hughie 123 1 Thomas, who ran a Post Office in Wales. You explain in 2 a witness statement to us your involvement in the case 3 was some months after the Post Office had concluded its 4 investigation of Mr Thomas and, in fact, he'd been 5 charged. Correct? 6 A. That is what I now understand, yes. I didn't at the 7 time, necessarily. 8 Q. Can we start by tracing your earliest involvement in the 9 case, that I can see, by looking at FUJ00122203. If we 10 look at page 6, please. I think we can see Mr Ward, at 11 12.43 on 10 March 2006 emailing a generic Fujitsu email 12 address; can you see that? It just says, "To Fujitsu" 13 on the right? 14 A. Yes, I see that. I've no idea what email address that 15 would have gone to. 16 Q. No. The subject is "ARQs, statement request and 17 assistance". Then if we scroll down, please. He says: 18 "... sorry for the length of this email ..." 19 I'm going to ignore the next two paragraphs, the 20 ones starting "Both of the above" and "Brian". Then if 21 we scroll down over the page, the third main paragraph, 22 the one starting "On a separate matter"? 23 A. Yes, I see that. 24 Q. He says: 25 "On a separate matter, I also require a witness 124 1 statement in respect of the following ARQs ..." 2 For some reason, they've been redacted. We know the 3 numbers -- 4 A. Yeah. 5 Q. -- and they appear in other emails: 6 "... all of which relate to [Mr Thomas' Post 7 Office]. We need the usual (leave out paragraphs H(b) 8 and J, but do need para K (call logs) covering 9 an analysis over the period 01/11/04 to 30/11/05. 10 Penny -- you may recall this one which relates to nil 11 transactions, my previous emails ... refer. Can you add 12 an extra paragraph in your statement explaining how 13 online banking transactions are processed and the data 14 downloaded and how nil transactions can occur." 15 I think that's all that there is about Mr Thomas' 16 sub post office. Okay? 17 A. Yes. 18 Q. So a request to either Penny Thomas or Neneh Lowther for 19 a request for a witness statement that addresses some 20 ARQs about Mr Thomas' Post Office and explains how nil 21 transactions can occur. 22 A. Yes. 23 Q. Can we go to FUJ00152582, page 2, please, and scroll 24 down a little bit. Mr Pinder emails you on 21 March, so 25 that's 11 days after the email we've just looked at -- 125 1 A. Yes. 2 Q. -- and says: 3 "Gareth 4 "Re Tel-con 5 "As discussed, please see extract from a recent 6 email below in italics from Graham Ward ... regarding 7 providing a statement about nil transactions and online 8 banking. If you are able to put something together for 9 us I would be very grateful. If you send it back I will 10 arrange for Neneh or Penny to write a statement for your 11 signature." 12 Then the part of the email we've just read out, 13 which does include the three ARQ numbers, is set out. 14 The last part of which is in italics. 15 A. Yes. 16 Q. "Can you add an extra paragraph in your statement 17 explaining how online banking transactions are processed 18 and the data downloaded and the nil transactions can 19 occur." 20 So you are asked to "put something together for us." 21 A. Yes. 22 Q. Addressing nil transactions and online banking? 23 A. Yes. 24 Q. His email seems to refer to a conversation with you. It 25 says, "Re Tel-con", at the top? 126 1 A. Yes, I see that. 2 Q. I don't suppose you remember what Mr Pinder told you of 3 Post Office's requests? 4 A. No, I mean, clearly I can understand what he's saying in 5 the email but what he actually said on the phone I've no 6 idea, I'm afraid. 7 Q. Did you have anything explained to you, by Mr Pinder or 8 otherwise, as to what was expected of you when creating 9 a witness statement or other document in response to 10 this request? 11 A. I can't remember. What I believe, from looking at the 12 documentation, I was asked to do was purely look at zero 13 value banking transactions and see why they came up as 14 being zero value banking transactions and that is what 15 I actually provided a report about. 16 Q. You'll see that this seems to ask you to provide 17 a witness statement about those things, doesn't it? 18 A. Yes, it does, and I think what I actually did was 19 provided a report to say "Is this what you want the 20 scope to be?" 21 Q. In fact, I think we're going to see that you reply by 22 email, not providing a witness statement, providing 23 an answer? 24 A. That is my recollection from the emails I've seen 25 lately, yes. 127 1 Q. So he's asking you for a witness statement, addressing 2 those two things. If we go up, please, to page 1, thank 3 you, same day, 21 March, you reply: 4 "I've had a look at the ARQs and I think there is 5 sufficient info there to explain in most cases why there 6 are zero value transactions. I suggest the following as 7 a brief explanation: 8 "There are three main reasons why a zero value 9 transaction may be generated as part of the banking 10 system: 11 "1 the transaction has no financial effect ... 12 "2. The transaction has been declined by the Bank 13 "3. There has been some sort of System Failure." 14 Then you explain response codes. The third 15 explanation is that a response code with a value greater 16 than 10 implies some sort of system failure. You say: 17 "Does that supply you with enough detail? 18 Identifying exactly why a given response code was 19 generated in each case and the exact circumstances under 20 which each one can be generated is much harder, but 21 I suspect that is unnecessary ..." 22 Then scroll down, you give some examples of some 23 response codes. So you respond not by providing 24 a witness statement: you say, "Here is a brief 25 explanation"? 128 1 A. Yes. 2 Q. Were you clearing with Mr Pinder what you said in 3 advance of committing it to a witness statement? 4 A. I was just providing the information. I don't think at 5 that stage I even knew what a witness statement was, 6 though I -- oh, I might have -- 7 Q. I think you'd already provide a draft -- 8 A. Yes, I'm trying to think but I don't think I'd actually 9 drafted one myself. It had been drafted for me by 10 Mr Pinder's team, in terms of getting the formatting 11 sorted out and -- 12 Q. That was in another case, in the Teja case? 13 A. That's the one, yes. 14 Q. So you did know what a witness statement was by then? 15 A. Yes, I must have done. 16 Q. Just scroll down to the top of the page and, sorry, just 17 scroll down to the email. You say: 18 "I suggest the following as a brief explanation ..." 19 Again, were you clearing with Mr Pinder what to say 20 in a witness statement before you made a witness 21 statement? 22 A. I'm not sure if it -- I think it was a case of: this is 23 the basic information, how do you want it presented? 24 I don't think I was sort of clearing -- I was happy that 25 technically that was a reasonable description of what 129 1 was happening and what the circumstances were, and that 2 was the basis of the information that I felt that he 3 needed for whatever witness statement was to be 4 produced. 5 Q. By somebody else? 6 A. For someone else to actually produce the statement, 7 though, it may be one that I would need to own and sign 8 up to at some point. 9 Q. What do you mean, you might need to own and sign up to 10 a witness statement? 11 A. That it would -- that it would need to be my witness 12 statement but, basically, at this stage, I was trying to 13 make sure that we had the scope correct. 14 Q. So you were clearing the scope for a proposed witness 15 statement by you? 16 A. Yes, okay. Yeah. 17 Q. If we scroll up to the top of the page. He's grateful 18 for your speedy response, he will ask Neneh to put if it 19 into statement form today for your perusal and signature 20 on Thursday if that's okay. 21 Can we go back then, please, to -- sorry, before we 22 do, was that a usual or an unusual approach in your 23 dealings with Fujitsu's litigation or Prosecution 24 Support Team, that you would draft your own witness 25 statements or somebody else would draft them for you? 130 1 A. I'm not sure that we can actually talk about the term 2 "usual", but this was, it was the first case that 3 I remember being involved in but, clearly, the case of 4 Teja has been drawn to my attention as part of the 5 Inquiry, which I had no recollection of whatsoever. 6 But -- so, in that way, this was the second time that 7 I'd been involved in doing anything with prosecution 8 support, so it was a bit difficult to find what is 9 usual. 10 Q. So too early to say what is usual? 11 A. Yes. 12 Q. Can we go back to FUJ00122203, which was the chain that 13 we were in before, and go to page 4, please. We can see 14 that at 1.01, Neneh Lowther sends Graham Ward a draft 15 witness statement "for the above re 'Nil Transactions'", 16 and asks him, Mr Ward, whether it meets his 17 requirements. Now, you tell us in your witness 18 statement -- no need to turn it up, it's paragraph 294 19 at page 98 of your third witness statement -- that you 20 cannot find a copy of the witness statement that was 21 attached to this email, and you're in good company: nor 22 can we. But that may not matter for the moment. 23 Can we go up to page 3, please. Thank you. We can 24 see that Mr Ward replies to Ms Lowther and Mr Pinder and 25 says, in relation to that witness statement that was 131 1 sent: 2 "The layout presumably is unfinished, paragraph, 3 spacings, etc. 4 "As per my earlier email, and more importantly the 5 3 spreadsheets sent with ARQ data need to be produced as 6 individual exhibits (GIJ/1, 2 and 3) and each column 7 header must be explained as you do in your usual 8 statements ... 9 "Also the line which begins at the foot of [the] 10 page, "[Response Code] with a value greater than 10 ...' 11 appears unfinished?" 12 This is perhaps, for our purposes, the most 13 important part: 14 "... I'm concerned with the words 'system failure' 15 which is also in an earlier line ... 'There has been 16 some sort of a system failure' -- What does this mean 17 exactly and is there any indication of a system failure 18 at this office during the period in question?" 19 Then if we go further up to page 3 we see that later 20 that day, at just after 3.00, Ms Lowther forwards that 21 to you; can you see that? 22 A. Yes, I see that. 23 Q. "Hi, Gareth, 24 "Please see [below] from Graham ... 25 "... ignore the ... bit about exhibits. I will 132 1 print out the exhibits and update your statement with 2 this. 3 "I understand that you have copies of the 4 spreadsheets ... 5 "Please could you look at his second paragraph 6 [that's the bit which explained concern about the 7 mention of system failures] and advise with your 8 comments again." 9 She sent your draft statement back and I think the 10 same comment as earlier applies, that there isn't a copy 11 of that available that you can find nor we can find. 12 Can we go up, please, to page 2, and keep going to 13 see the beginning part of Mr Jenkins' email. You reply 14 still later that day, 5.40: 15 "Neneh, 16 "I'm not quite sure what his problem is with what 17 I've said. 18 "Basically, any value of Response Code that is 19 greater than 10 does imply some end-to-end system 20 failure. The actual value makes it clearer what exactly 21 the failure is and where it has been detected. 22 "In the example there are two such codes ..." 23 You'll see what you say about those. You say: 24 "How do you want to play this? Do you want to add 25 in specific text to the witness statement to cover these 133 1 two codes or persuade [Post Office] that the generic 2 statement is okay (perhaps with some clearer words)?" 3 So what did you mean by "How do you want to play" 4 it? 5 A. What exactly was she asking me to do? Now I had this 6 information, what does she want to do next? And I was 7 giving her a couple of options as to what we could do 8 moving forward. 9 Q. Can we go to page 1, please, and scroll down to her 10 reply, the next day: 11 "Gareth, 12 "I have updated your [witness statement] with all 13 the column headings you explained to me earlier. 14 "I have not included your response below as I am not 15 quite sure where to fit it in. 16 "Can you help out with this. 17 "... Graham Ward is thinking that 'system failures' 18 are drastic events. Could you please explain a little 19 further on this like you said they are normal system 20 occurrences. 21 "Please sent back to me when you've finished so 22 I can send to [Graham Ward] if possible." 23 Then, on this chain, the last email -- scroll up -- 24 you attach a witness statement and say you've annotated 25 it with revisions: 134 1 "In particular, I don't feel I can include the last 2 two paragraphs, which may make the statement useless." 3 Can we look at the attachment to this email. 4 FUJ00122204. If we just scroll down, we can see it, it 5 is dated 23 March. First paragraph: is that something 6 you had drafted in the original or had been drafted for 7 you? 8 A. I'm not 100 per cent sure. I suspect that this was what 9 Neneh had produced in the earlier emails in the chain 10 and that bit of red that you see there is what I've 11 added in and this latest version is how I'm interpreting 12 it but I can't be 100 per cent certain. 13 Q. More generally, before we come to the red parts, if we 14 look at the black text, which is indented underneath 15 that first paragraph, we can see, effectively, that your 16 earlier email has been cut into this witness statement, 17 hasn't it? 18 A. Yes. 19 Q. Verbatim? 20 A. Yes, and perhaps this is what Graham Ward was referring 21 to, in terms of formatting problems. I'm not sure. 22 Q. To address the email that you'd received, you've added 23 in the words underlined, "Such failures are normal 24 occurrences", is that right? 25 A. I thought that may be sufficient to address the fact 135 1 that I could clearly see that people were concerned 2 about the use of the term "system failure", and I'm sure 3 we're going to discuss that further this afternoon but 4 I felt as a first cut, let's just say, well, they are 5 normal occurrences. 6 Q. What do you mean, "let's say as a first cut"? 7 A. Well, by "system failure" what this was talking about 8 was an end-to-end failure within the overall banking 9 system and so I thought that, just by saying this was 10 a normal occurrence, then that would be something that 11 would be sufficient to not make it considered emotive, 12 and I can clearly see that Graham Ward was concerned 13 about having the term "system failure" in the witness 14 statement because that could give the wrong impression 15 because that could -- I now understand that he was 16 thinking that meant that was a failure in Horizon. That 17 is not what I meant by a system failure in this 18 particular case. 19 Q. So you thought that your original language had sort of 20 scared the horses and this was a way of calming -- 21 A. Calming it town a bit, yes. 22 Q. Moving on, please, if we scroll on. You've added in, in 23 red, at the bottom there: 24 "The actual value provides further information as to 25 the nature of the failure within the overall system." 136 1 A. Again, as a way of trying to calm things down and make 2 it less emotive. 3 Q. I'm going to skip over, if we go to page 2, the other 4 amendments, and if we scroll through page 2 and go to 5 page 3 and scroll down. Thank you. You remember in 6 your email you said that you didn't feel that you could 7 include the last two paragraphs. Are they the 8 paragraphs beginning "There is no reason to believe", 9 and "Any records to which I refer"? 10 A. Yes. 11 Q. You said that you don't feel that you could include them 12 and that might make the statement useless; is that 13 right -- 14 A. Yes. 15 Q. -- in your email? You say, by way of comment, 16 underneath: 17 "I am not sure that the yellow bit is true. Can 18 this be deleted? All I've done is interpret the data in 19 spreadsheets that you've emailed to me." 20 A. And, in particular, I was referring to the second of 21 those two paragraphs, which was saying that I'd actually 22 got the records out of the audit system, because 23 I hadn't. 24 Q. Does the second paragraph actually say that you had got 25 the records out of the audit system? I don't think it 137 1 does? 2 A. It doesn't actually say that but it does imply that 3 I had actually got hold of the information and I hadn't 4 got hold of the information; I had just been sent 5 a spreadsheet by Neneh to say look to see what it says 6 in this spreadsheet, and I had no reason -- no 7 information as to where that spreadsheet had come from. 8 Q. I want to concentrate on the first paragraph instead, 9 please. That first paragraph in the second sentence 10 says: 11 "To the best of my knowledge and belief at all 12 material times the computer was operating, or if not, 13 any respect in which it was not operating properly, or 14 was out of operation was not such as to affect the 15 information held on it." 16 You said that you weren't sure that was true and you 17 asked for it to be deleted, correct? 18 A. Yes. 19 Q. Now, in terms of the guidance that you'd received about 20 writing this statement, you tell us in your witness 21 statement -- no need to turn it up, it's paragraph 313 22 of your third witness statement -- that you do not 23 recall receiving any briefing from the Post Office or 24 anyone else setting out the factual history of the case 25 or summarising the prosecution evidence, or the issues 138 1 raised by Mr Thomas? 2 A. Correct. 3 Q. Did you ask for any? 4 A. I didn't think I needed to because I was being asked 5 a very simple question: what do these zero value 6 transactions mean? 7 Q. So the answer is you didn't ask for any guidance because 8 you didn't think it was necessary? 9 A. Correct. 10 Q. You tell us in your witness statement -- it's 11 paragraph 316 of your third witness statement -- that 12 you do not have a clear memory of anyone mention anyone 13 the word "expert" at any point and that you understood 14 that you were giving evidence about Horizon because of 15 your knowledge and experience of it. It was never 16 explained to you that there might be a difference 17 between giving you giving factual evidence about Horizon 18 and expert evidence, in this context? 19 A. Correct. 20 Q. This witness statement and the drafts of it do not 21 contain any expert declaration, do they? 22 A. No. 23 Q. Why do you think you were chosen to prepare a witness 24 statement regarding nil transactions in Mr Thomas' 25 branch? 139 1 A. Because I understood the way that nil transactions could 2 be generated. I'd been involved in the design of the 3 banking interfaces, so I actually understood how you 4 could get zero value banking transactions in the 5 circumstances and how to interpret them in the ARQ data. 6 Q. Do you know whether anyone else within Fujitsu was 7 considered as a provider of this statement? 8 A. I don't know that. 9 Q. Did you encounter any challenges in preparing the 10 statement requested by Graham Ward from the Post Office, 11 without the factual history of Mr Thomas' case, or even 12 a summary of the prosecution case? 13 A. I believed that I was being asked a very simple 14 question: why are there zero value transactions. And 15 that is all I was trying to address here. 16 Q. When you received Mr Ward's request via Mr Pinder and 17 Ms Lowther, was your understanding that you were 18 providing a factual analysis of Horizon or giving any 19 opinion evidence? 20 A. I think that -- I saw that as being factual information. 21 Q. Did you understand the legal significance and full 22 implications of preparing a witness statement for the 23 purposes of criminal proceedings against the 24 subpostmaster in a prosecution brought by the Post 25 Office? 140 1 A. No. 2 Q. Did you ask for guidance? 3 A. No, I thought I was just doing a simple bit of analysis 4 identifying the reason why there was zero value 5 transactions, and I thought that was a relatively 6 straightforward bit of factual information that I could 7 provide. 8 Q. Did you have adequate time to prepare this witness 9 statement? 10 A. I can't remember how much time I had. I clearly had 11 looked at the data that I had been presented with and 12 had had time to analyse all the zero value transactions 13 in the data that I'd been provided with. I can't 14 remember now how much data that was. 15 Q. Can we try and display, before the break, two documents 16 side by side. The one that's on the screen at the 17 moment, please, and the second document, your third 18 witness statement, at page 33. Perfect. You say in 19 your witness statement at paragraph 102: 20 "I am aware that there is a question in the Inquiry 21 as to what the two 'boilerplate' or 'standard' 22 paragraphs that appear at the very end of the standard 23 Fujitsu witness statement which exhibits the ARQ 24 spreadsheets actually meant. These two paragraphs are 25 usually read as follows ..." 141 1 You set them out and they broadly match the 2 paragraphs on the right-hand side, can you see? 3 A. Yes, yeah. 4 Q. Moving on to 103: 5 "As I understand it, some witnesses to the Inquiry 6 have suggested that these paragraphs were supposed to 7 show (or were [intended] to mean) that Horizon was 8 working properly at a given branch or even that Horizon 9 was working properly across the whole estate. That is 10 not what I thought these paragraphs were intended to 11 mean. 12 "As I have set out above [paragraph 104], Fujitsu's 13 operational manuals concerning prosecution support 14 appear to deal with the integrity of audit data, rather 15 than the integrity of the Horizon system." 16 Just stopping there, earlier in the statement you 17 rehearsed those. You hadn't seen them at the time, in 18 fact, those manuals? 19 A. Correct. 20 Q. So you're referring here, in support of your position, 21 to documents that you hadn't seen at the time? 22 A. Yes. 23 Q. "As I have set out above, Fujitsu's operational manuals 24 concerning prosecution support appear to deal with the 25 integrity of audit data rather than the integrity of the 142 1 Horizon system. They do not suggest the standard 2 Fujitsu witness statement would address the integrity of 3 the overall Horizon system at a given branch. Indeed, 4 as I have already noted, paragraph L of this standard 5 witness statement referred to the integrity of the audit 6 data exhibited to the witness statement in terms of its 7 origination, storage and retrieval. 8 "105. The individuals who signed this standard 9 witness statement (such as Penny and Neneh) could not 10 have given an opinion about the integrity of the Horizon 11 system, whether at the particular branch or generally. 12 They did not have the technical expertise and would not 13 have been qualified to do this. They could comment on 14 the integrity of the ARQ extraction process that they 15 had personally undertaken. 16 "106. All of this explains why it makes little 17 sense to me that the standard closing paragraphs in the 18 Fujitsu witness statement could have been intended to 19 have some broader meaning about the integrity of 20 Horizon. This witness statement was only concerned with 21 the production of time limited ARQ data and did not (for 22 example) set out an analysis of the data. I think there 23 is a clear distinction between a statement that the 24 Horizon system was working (or a statement to its 25 overall health) at a given branch during a specific 143 1 period and a statement about the way in which the ARQ 2 data for that branch had been produced (so that it had 3 integrity). 4 "There is no explanation (that I have seen) in the 5 Fujitsu documents as to what these two standard 6 paragraphs were supposed to mean. I can see from the 7 communications in the case study of Mr Hughie Thomas 8 that I raised a question about the two paragraphs." 9 That's essentially what we're looking at on the 10 right-hand side. 11 A. Indeed. 12 Q. "Looking back now, I think my understanding was that the 13 first of these two paragraphs related to the proper 14 operation of the computers involved in the production of 15 the witness statement and that the second standard 16 paragraph related to the process by which any records 17 referred to in the witness statement had been obtained 18 and produced. In Mr Thomas' case, I think that my 19 concern was that I could not include these paragraphs 20 because I had not extracted the ARQ spreadsheets that my 21 draft statement was referring to. By this I mean that 22 I could not speak to the computer which had extracted 23 the spreadsheets as working properly." 24 You deal with the issue in more detail below. 25 Just going back to the beginning part of that 144 1 paragraph 107, thank you. Four lines in, you say, 2 "Looking back at it now, I think my understanding was 3 that the first of the two paragraphs related to the 4 proper operation of the computers involved in the 5 production of the witness statement ..." 6 Just stop there. So you're saying, by that, that 7 the Word Processor or other computer on which the 8 statement was being typed, or typed for you, was working 9 properly? 10 A. And whatever was being used for doing the analysis, and 11 so on, yes. 12 Q. So you're saying that the first paragraph we see on the 13 right-hand side in yellow relates to the computer on 14 which the statement was being typed? 15 A. That is the only way I can understand that as making any 16 sort of sense in terms of the people who are producing 17 that sort of statement. 18 Q. You ask on the right-hand side that that paragraph is 19 deleted, yes? 20 A. Because I -- 21 Q. Just take it in stages. You asked that it's deleted. 22 A. Yeah. 23 Q. Did you think the computer on which the statement was 24 being typed was not working properly? 25 A. No. 145 1 Q. You were presumably satisfied that your draft witness 2 statement was not inaccurate because of any improper use 3 of the computer on which it was being drafted? 4 A. Correct. 5 Q. Put another way, to the best of your knowledge and 6 belief, the computer on which you drafted the witness 7 statement was operating properly at the time? 8 A. Yes. 9 Q. If the computer on which you drafted the statement 10 wasn't operating properly, presumably any such problems 11 with the computer wouldn't affect the accuracy of the 12 content of the witness statement? 13 A. Yes, I think so, yeah. 14 Q. On what basis, therefore, could you have any doubts as 15 to the truthfulness of the first section highlighted on 16 the right, on the basis of the reason that you have 17 given -- you give, as is highlighted in yellow on the 18 left? 19 A. I was just unclear about the whole lot and maybe I must 20 have had some sort of discussion and then, when they 21 said, well, that needs to be in there, then I realised 22 what it must have meant, because that was the only 23 sensible explanation for what that paragraph could mean. 24 Q. No, you say; looking back now, your understanding is 25 that the first of the two paragraphs is, as I've 146 1 summarised it, about the operation, the proper 2 operation, of the computers being involved to type up 3 the witness statement? 4 A. Yeah. 5 Q. You can't have honestly have believed that? 6 A. Well, I couldn't see any other way that that paragraph 7 made any sense. 8 Q. Are you trying to explain away the first paragraph on 9 the right-hand side through an explanation on the 10 left-hand side which doesn't hold substance? 11 A. I think it does hold substance because I couldn't see it 12 having any sort of meaning in any other form. 13 Q. Did you have any reason to believe that the computers 14 involved in collating the ARQ data were not operating 15 properly? 16 A. I had no knowledge of whether they were operating 17 properly or not because I wasn't using them. I didn't 18 have access to them. 19 Q. So if the first paragraph related to the computers used 20 to obtain the ARQ data, that might be a proper reason 21 for the deletion of that paragraph? 22 A. Yes. 23 Q. But that's not the reason that you give in your witness 24 statement? 25 A. Well, that was part of the -- it was the computers 147 1 involved in the production of that, so not just the 2 typing up but, actually, the extraction of the ARQ data 3 as well, yes. 4 Q. Is the truth of the matter that what you explain in this 5 paragraph, 107, one that you've come up with after the 6 event to seek to explain away this comment in the draft 7 witness statement that you didn't think you could 8 include the last two paragraphs? 9 A. I just don't remember what it was. I certainly did 10 include that paragraph and the only way it can make any 11 sense is with the explanation I've come up with. 12 Q. Or is the truth of the matter that you know that the 13 first paragraph on the right refers, and was intended to 14 refer, to Horizon, and the truth of the matter is that 15 you didn't think you could include that first paragraph 16 because you knew that there was no way that you could 17 say that the Horizon system, ie the computer, was 18 operating properly or, if not, any respect in which it 19 was not operating properly wouldn't affect the 20 information held in it? 21 A. I wouldn't have referred to the Horizon system as "the 22 computer"; I would have referred to it as "the Horizon 23 system". 24 SIR WYN WILLIAMS: Well, you didn't write it though it, did 25 you? 148 1 A. Yes, but as I say -- 2 SIR WYN WILLIAMS: Well, no, you didn't write it, did you? 3 A. No. 4 SIR WYN WILLIAMS: No. 5 MR BEER: If we go back to page 1, please. On the 6 right-hand side, if we scroll down, paragraph 1, you 7 say, second line: 8 "I have been ... involved in many aspects of design 9 and implementation of the computer system known as 10 Horizon ..." 11 Isn't your reference in the last paragraph 12 a reference back to that computer system? 13 A. I didn't think so. 14 SIR WYN WILLIAMS: I'm sorry to be persistent but you didn't 15 write that either, did you? 16 A. I think I probably would have written the first 17 paragraph but I -- 18 SIR WYN WILLIAMS: I thought you told Mr Beer earlier that 19 that was a paragraph suggested to you, and I'm not 20 suggesting it was improper to suggest it to you, simply 21 that it wasn't you who had written it. 22 A. Right. Okay, I can't remember exactly who wrote which 23 bits at the time, sorry. 24 SIR WYN WILLIAMS: All right. 25 MR BEER: Can we go back to the last page of the statement 149 1 on the right -- sorry, one page up, the second paragraph 2 in yellow: 3 "Any records to which I refer in my statement form 4 part of records relating to the business of Fujitsu ..." 5 Why couldn't you say that? 6 A. Because I thought it was saying that I'd actually 7 extracted them -- 8 Q. It doesn't say that -- sorry to speak over you. It 9 doesn't say, "I've extracted them". 10 A. That is what I thought that they were saying. 11 Q. So you're saying that you thought that that last 12 paragraph said that you personally had exacted them and 13 that's why you didn't think you could properly say them? 14 A. Yes. 15 MR BEER: Thank you. 16 Sir, it's 3.00 now. I wonder whether that would be 17 an appropriate moment. 18 SIR WYN WILLIAMS: Yes, of course. 3.15, yes? 19 MR BEER: Thank you very much. 20 (3.00 pm) 21 (A short break) 22 (3.17 pm) 23 SIR WYN WILLIAMS: Yes, Mr Beer? 24 MR BEER: Thank you, sir. 25 Mr Jenkins, can we pick up the chronology with 150 1 24 March 2006, in Mr Thomas' case, by looking at 2 FUJ00122210. If we look at that first page and scroll 3 down, thank you, an email, from Mr Ward to Ms Lowther, 4 copied to Brian Pinder, Mr Baines, Mr Dawkins, and Diane 5 Matthews. So not you at this stage, although we'll see 6 from the email above in a moment that this was sent on 7 to you. 8 A. Yes. 9 Q. He, Mr Ward, says: 10 "This statement needs more work ... I've attached 11 a suggested draft with a number of comments (as 12 mentioned previously I think the 'system failure ... 13 normal occurrence' line is potentially very damaging). 14 It may be worth considering someone from our team taking 15 a statement directly from Gareth (where is he based?) 16 "Whilst there is some urgency with this, it is more 17 important to get it right and ensure we are not 18 embarrassed at court, which we certainly could be if we 19 produced a statement accepting 'system failures are 20 normal occurrences'." 21 Do you agree that Mr Ward was seeking here directly 22 to influence the contents of your witness statement in 23 criminal proceedings before the Crown Court? 24 A. I think that could be interpreted in that way, yes. 25 Q. Do you interpret it in that way? 151 1 A. I think he was just looking at getting my meaning made 2 clearer, in that I -- I can understand why he felt that 3 the term "system failure" was going to be potentially 4 causing embarrassment, and I can understand that the way 5 I was using it was not in the way that he was 6 interpreting it. So I felt that there was the scope for 7 improving the wording. 8 Q. He was seeking to influence, do you agree, the contents 9 of your witness statement, as it was presently drafted, 10 because it may be damaging to the prospects of 11 a prosecution? 12 A. Yes, I think you'd have to ... say that, yes. 13 Q. And that the Post Office could thereby be embarrassed at 14 court? 15 A. Yes. 16 Q. This was sent to you, if we just look up. What did you 17 think by this attempt by the Post Office to attempt to 18 alter your evidence? 19 A. I don't think I interpreted it in quite that way at the 20 time, I just saw it as being a way of improving the 21 wording of what I had in the statements, rather than 22 being -- just trying to alter what I was saying. I just 23 saw it as a way of improving things. 24 Q. Let's look at how he tried to improve it by looking at 25 the attachment, FUJ00122211. Can we go to the second 152 1 page, please, and scroll down. Stop there. 2 You had typed in your original email, and it had 3 been cut into a witness statement for you: 4 "There are three main reasons why a zero value 5 transaction may be generated as part of the banking 6 system ... 7 "The transaction has no financial effect ... 8 "The transaction has been declined by the Bank 9 "There has been some sort of System Failure. Such 10 failures are normal occurrences." 11 If we stop there, he has struck through, either by 12 deletion or by typing over -- and the Word system has 13 thereby struck through because its Track Changes -- the 14 word "main", hasn't he? 15 A. Yes. 16 Q. He has added a question: 17 "If these are the main reasons, what are the rest?" 18 He has struck through again, in the same way that 19 I've just described, either by direct deletion or by 20 highlighting and then typing over: 21 "There has been some sort of system failure. Such 22 failures are normal occurrences." 23 A. Yes. 24 Q. He has added some comments in brackets afterwards. This 25 is comments are addressed to you, aren't they? 153 1 A. Yes, I assume so. 2 Q. "This a really poor choice of words which seems to 3 accept that failures in the system are normal and 4 therefore may well support the [subpostmaster's] claim 5 that the system is to blame for the losses!!!!" 6 You would agree, would you, that the words in 7 brackets there that are highlighted seek to explain the 8 reason for the deletion of the "system failure" reason? 9 A. Yes. 10 Q. That's how it reads and it's how you would read it. 11 A. Yes. 12 Q. It's how you would have read it at the time, do you 13 agree? 14 A. I think so, yes. 15 Q. What did you think of this? Did you think this was 16 an attempt to alter your evidence? If you didn't think 17 the email was, what about the striking through of 18 passages of your witness statement? 19 A. I saw that as seeking to see whether I was happy to 20 reword it in using less emotive words, and he's made it 21 very clear that he doesn't like the words that I've used 22 there. 23 Q. Does he say, "Can you explain this by using less emotive 24 words"? 25 A. No, he doesn't. 154 1 Q. Why did you read it or see it as for him seeking for you 2 to use less emotive words. 3 A. Because clearly he didn't like the words "system 4 failure" and he had made that clear both in the covering 5 email and in the document. 6 Q. And by deleting it? 7 A. And by deleting it, yes. 8 Q. Did Mr Ward, to your knowledge, have any technical 9 knowledge? 10 A. I don't believe so but I don't know. 11 Q. At the time, did you consider his conduct to be 12 appropriate? 13 A. I saw him as trying to get -- to improve the way that 14 the statement was worded. 15 Q. So you thought this was appropriate? 16 A. I didn't think it was inappropriate at the time. 17 Q. You presumably realised that these were very serious 18 matters, these were proceedings in the Crown Court? 19 A. Oh, yes. 20 Q. This is your evidence and not his? 21 A. Yes. 22 Q. There's a declaration at the top of the statement, 23 which -- 24 A. Yes. 25 Q. -- I think you will have read? 155 1 A. Yes. 2 Q. If we go back to that at page 1, that statutory 3 declaration at the top there? 4 A. Yes, yes. 5 Q. It's a personal witness statement? 6 A. Yes. 7 Q. Did you approach anyone within Fujitsu to say, "Is this 8 appropriate, a Casework Manager suggesting changes to my 9 evidence?" 10 A. I may have discussed it with Neneh or Penny, I don't 11 know -- or Brian Pinder -- I don't think I would have 12 discussed it with anyone else. 13 Q. If we go back to page 3, and scroll down. We can see 14 that the last two paragraphs, the boilerplate 15 paragraphs, don't appear there, either as text or as 16 being deleted. 17 A. Yes. 18 Q. Do you think you had deleted those or you'd removed them 19 when you sent back your previous version? 20 A. I don't know. 21 Q. Can we go to FUJ00152587, top of page 2, please. You 22 reply to him, sending that marked-up version of the 23 witness statement to you: 24 "I've added some further annotations to your 25 annotations. Does this move us forward?" 156 1 Did you regard this as some kind of negotiation in 2 which you were required to haggle with the Post Office 3 about the way in which you described the Horizon system 4 and whether it had defects? 5 A. I suppose in a way I probably did, because that was how 6 I was used to -- when writing technical documents, then 7 I would receive comments them and then make up dates 8 based on the comments I'd received, and then update the 9 next draft of the document. So I saw this as being 10 a similar sort of process. 11 Q. Do I take it from that that you didn't regard it as 12 inappropriate to be haggling with the prosecutor over 13 the content of your personal witness statement? 14 A. Yes, I think you can take it that way. 15 Q. Can we look at the attachment that you sent back, asking 16 whether it moved us forwards. FUJ00122216. Now, 17 I think the tracked changes have been lost on this but 18 I think we can make sense of it. If we scroll down, 19 please, paragraph 1, no changes. Paragraph 2, I think 20 we can see a subtle difference of change in text in 21 line 4; can you see that? 22 A. Ah, yes. 23 Q. It looks like a smaller font size. Is that likely to be 24 your annotation? 25 A. Yes, the "I'm not sure about this, I've had nothing to 157 1 do with producing the spreadsheets", I think is my text 2 and, in fact, the rest of that paragraph is mine. So 3 I think the original must have ended at "GIJ/03" and 4 I think the "I'm not sure about this" is probably mine 5 as well. 6 Q. So the passage which reads: 7 "I'm not sure about this, I've had nothing to do 8 with producing the spreadsheets. All I've done is make 9 some statements based on what is in the spreadsheet. 10 I assume that Neneh or Penny produced the spreadsheets, 11 but I have no personal knowledge as to what was included 12 within them and what was excluded. For all I know, you 13 could have typed them up from scratch." 14 So you were saying, "I'm unhappy with the four lines 15 that appear beforehand. I don't want them included"? 16 A. Yes. 17 Q. "Because I haven't produced any of the spreadsheets, 18 I've got no personal of what was included or excluded 19 from the spreadsheets: you could have just made them 20 up"? 21 A. Yes. 22 Q. Can we look, please, at the same time as this document, 23 at FUJ00122237. That, on the left-hand side, is the 24 final version of the witness statement, dated 6 April 25 2006, which the Post Office in their disclosure tell us 158 1 was the one submitted in the prosecution to the court or 2 disclosed to the court in the prosecution of Mr Thomas, 3 okay? 4 A. Okay, but I'm not aware that you've actually signed it. 5 Q. No, in this version -- well, I might as well ask you 6 now: did you sign a version? 7 A. I can't remember is the simple answer to that. 8 Q. Do you remember them coming to see you? 9 A. I don't actually remember that, although I can see from 10 the email exchanges referring to "Meeting you a couple 11 of weeks ago" that it must have happened but I've no 12 recollection of that meeting. 13 Q. Okay, so you would agree that the evidence tends to show 14 that they did come to see you to take a witness 15 statement from you? 16 A. That appears to be what happened, it's just that I don't 17 remember the actual meeting. 18 Q. Then scroll down, please. If we look at the third 19 paragraph, you say: 20 "I extracted data from this system ..." 21 That's the PEAK system. 22 A. Yes. 23 Q. "... concerning [the] post office. The data extracted 24 was for all transactions for three specific time periods 25 ... From this data, I then extracted all the banking 159 1 transactions which showed a zero value." 2 Then it continues. 3 A. Yes. 4 Q. Is it right that you obtained the data from the PEAK? 5 A. That's what it says there. Therefore, I assume I must 6 have done so. 7 Q. The data from the PEAK was not created by the PEAK. It 8 had already been created, captured and attached to the 9 PEAK by someone else, hadn't it? 10 A. Yes, I realise that now and, clearly, the PEAK is not 11 a good repository of the data like the ARQ data but 12 I felt that at least what I was saying there was 13 correct, in that I had extracted the data from PEAK, 14 whilst I was unable to say that I had extracted the data 15 from the ARQ system. 16 Q. Well, just reading what you say on the left-hand side in 17 paragraphs 2 and 3, you say there was a fault management 18 the system called the PEAK system and then you say: 19 "I extracted data from this system concerning [the] 20 post office. The data extracted was for all 21 transactions for three specific time periods requested." 22 There was no data on the PEAK system for all 23 transactions for the three specific time periods, was 24 there? 25 A. I think there must have been at the time because, if 160 1 a message store was extracted and put onto the PEAK 2 system, it will have 35 days' worth of data, so -- 3 Q. We're talking at cross purposes. I'm sorry, it's me. 4 The PEAK system itself doesn't create that data. It's 5 not a means of capture of it. We've seen the PEAK 6 system used. Sometimes people just attach documents to 7 it -- 8 A. Yes, yeah. 9 Q. -- almost drag them into the PEAK? 10 A. Yeah. 11 Q. Is that what had happened here, the three sets of data 12 had been dragged into the PEAK by somebody else? 13 A. Yes, it had been. 14 Q. So when you say, "I extracted data from the system", the 15 PEAK system, and the data extracted was for all 16 transactions, what you're doing is not saying that 17 somebody else had performed the original extraction 18 function? 19 A. Correct. 20 Q. How was it any better from what you described on the 21 right-hand side, where you say, "For all I know, you 22 could have typed them up from scratch"? 23 A. Looking at it now I agree but, for some reason, I must 24 have decided that that was all right at the time but 25 I can't remember why. 161 1 Q. Well, can I press you on that. You clearly rightly, on 2 the right-hand side, say, "I can't exhibit three sets of 3 ARQs because I've had no role in obtaining the data in 4 those ARQs"? 5 A. Yes. 6 Q. On the left-hand side, in what appears to be your final 7 statement, you say, "I have extracted that data from the 8 PEAK". It was no better than on the right-hand side, 9 was it? 10 A. I accept that now but, clearly, I had not thought it 11 through in that way at the time. 12 Q. Can we look, please -- both those documents can come 13 down -- at FUJ00122230. In between those two drafts, 14 you sent this email to Brian Pinder and say: 15 "Brian, 16 "I've taken the data off the PEAK and carried out my 17 own analysis of it and presented the results in the 18 attached Word document." 19 Then there are the ARQs, 401, 459 and 460,which 20 I think are Excel documents, and then a Word document, 21 which is your own analysis; is that right? 22 A. I believe so, yes. 23 Q. So again, this not you extracting data, proprietary data 24 from the system, it's you dragging data captured by 25 somebody else from the PEAK system, correct? 162 1 A. Yes, I think the one thing that is slightly different 2 here is that, in the previous case, the spreadsheets had 3 been produced from the raw ARQ data by Neneh or Penny, 4 whilst, in this case, I'd actually extracted the 5 spreadsheets from the PEAKs myself. So I think that's 6 the subtle difference here but I accept that it's a very 7 subtle difference. 8 Q. Can we look at the first attachment to your email here, 9 where you say, "I've ... carried out my own analysis of 10 it and presented the results in [a] Word document", by 11 looking at that attachment. FUJ00122229, your document 12 of 30 March 2006: 13 "This note is provided as input to a witness 14 statement ... 15 "What I've Done 16 "Penny Thomas provided me with extracts from Audited 17 Data for 3 separate periods. 18 "I have taken this data and extracted details of all 19 banking transactions and analysed the zero value 20 transactions." 21 Here, you're referring to the data that Penny Thomas 22 has provided you, rather than data which you have 23 dragged from the PEAK system. 24 A. I accept that that's what that says and I just can't 25 remember what had actually happened at the time. 163 1 Q. Well, this is under a heading "What I've Done"? 2 A. Yes. 3 Q. So, presumably, that's what you did? 4 A. Yes. Sorry, I just can't remember the details of 5 exactly what was happening at the time. 6 Q. Well, you accept here what you're saying is you've in 7 fact used the data provided by Penny Thomas, not the 8 data extracted from the PEAK? 9 A. I can't explain the difference. 10 Q. So, in fact, the analysis on the basis of this document 11 was based on extracts provided for you by Penny Thomas, 12 which you said you were earlier unhappy to rely on, 13 because she could have typed them up from scratch, and 14 then you ended up making a witness statement, if we go 15 back to it, FUJ00122216, that's the one which said, "For 16 all I know, you could have typed this up from scratch", 17 yes? 18 A. Yes. 19 Q. Then you ended up -- if we can also display FUJ00122237. 20 So, based on your Word document, you'd, in fact, used 21 the data provided by Penny Thomas. If we scroll down on 22 the left-hand side, you tell us about the PEAK in the 23 second paragraph, and then you say: 24 "I extracted data from this system concerning [the] 25 post office. The data extracted was for all 164 1 transactions for three specific time periods ..." 2 That wasn't true, was it? 3 A. I'm not sure. Whether something had happened between 4 30 March and 6 April, I just can't remember. 5 Q. Was the wording you used in this final statement here 6 a loincloth to cover up what you had done? You rightly 7 realised that you couldn't say that you'd extracted the 8 ARQ data nor could you produce it yourself. You said in 9 your Word document that you didn't in fact analyse data 10 that you'd taken from the PEAK but, as this document 11 makes clear, you've used the phrase "I've extracted from 12 the PEAK" to paper over what, in fact, had happened? 13 A. If I said there that I extracted it from the PEAK then 14 I would have extracted it from the PEAK but I can't 15 remember exactly when I did that and how I did that. 16 Q. So, if that's correct, then there's something else that 17 has occurred after the 30 March document when your 18 analysis was based on Penny Thomas' data? 19 A. I think there must have been, yes. 20 Q. If we go back to just FUJ00122216, on the right-hand 21 side that's it, and if we can go to page 2, please -- 22 and scroll down, thank you -- in the paragraph that 23 begins "Should the spreadsheets be regenerated", three 24 lines in, the statement reads: 25 "There are three reasons why a zero value 165 1 transaction may be generated as part of the banking 2 system: 3 "[1] The transaction has no financial effect ... 4 "[2] The transaction hacks been declined by the bank 5 ..." 6 Then the third reason, "system failure", has 7 disappeared, hasn't it? 8 A. Yes, and that, I think, is due to the document we looked 9 at earlier. 10 Q. Ie it had been deleted in the tracked changes -- 11 A. Yes. 12 Q. -- by Mr Ward? 13 A. Yes. 14 Q. What remains is his addition "This is a really poor 15 choice of words", and you reply to that in the passage 16 after the brackets; is that right? 17 A. Yes, so the version that I would have seen would have 18 had the deleted thing and I'd not twigged that I hadn't 19 undeleted it or replaced it with something. 20 Q. Well, I'm not sure you would have undeleted it because 21 you say, "Can you please suggest something better, 22 then"? 23 A. Yes. 24 Q. Why were you asking a casework handler to suggest what 25 you should say in a witness statement for proceedings in 166 1 the Crown Court? 2 A. Because he clearly didn't like the wording that I was 3 using, so I was asking if he had a better suggestion for 4 me to use. 5 Q. Again, did you see it as appropriate to haggle with the 6 prosecutor over the content of your witness statement? 7 A. I didn't see it as inappropriate at that time but 8 I clearly see now that it was. 9 Q. You explain that: 10 "What we have here are genuine failures of the 11 end-to-end system which are not part of the normal 12 operation, but are anticipated, and, the system is 13 designed to cope with them. Some such failures could be 14 engineered as part of a malicious attack (but that 15 doesn't apply to those failures that appear in the 16 evidence presented). In all cases the system is 17 designed to identify such failures and handle them in 18 such a way that the Customer, the Postmaster, Post 19 Office and the FIs ..." 20 FIs? 21 A. Financial institutions. It's a code for banks, 22 basically. 23 Q. "... are all clear as to the status of the transaction 24 and any necessary financial reconciliation takes place. 25 I guess one option ..." 167 1 This is still you speaking, yes? 2 A. Yes. 3 Q. "I guess one option is to delete the paragraph since it 4 is purely an introduction to the following more detailed 5 description." 6 So, again, was this part of the continuing dialogue 7 that you thought it was appropriate to have as to how to 8 word your witness statement in proceedings before the 9 Crown Court? 10 A. Yes, I think it was. 11 Q. Can we look, please, at the final witness statement 12 alongside, FUJ00122237, and on the left-hand side scroll 13 down, please -- and stop. If we just scroll up a little 14 bit. Thank you. You explain response codes 3, 4, 6 and 15 8: the financial institution has declined the 16 transaction. 17 You explain codes 23 and 26, which indicate 18 an unusual response from the financial institution and, 19 again, no cash should change hands. 20 You say: 21 "... Response Code 32 indicates the Horizon's 22 central system has received no response to the financial 23 institution within a timeout period and so no Cash 24 should change hands." 25 You then add: 168 1 "All these are perfectly normal occurrences and 2 should result in clear instructions being passed to the 3 counter clerk to ensure that no Cash should change 4 hands." 5 Just stopping there, it would appear that Mr Ward 6 has been successful in pausing the alteration of your 7 evidence? 8 A. Yes. Well, I still stand by what it actually says there 9 as being an accurate representation of what had actually 10 happened within the system. 11 Q. Is that why you were a willing actor in this process? 12 A. Yes. 13 Q. Why you went along with his pushback? 14 A. Yes, because it was still accurately reflecting what had 15 actually happened, but using less emotive words, which 16 he clearly wasn't happy with. 17 Q. Then can we see underneath, on the left-hand side, the 18 objectionable last two paragraphs, the ones that, 19 remember, had been marked up in yellow, have been 20 replaced by: 21 "There is no reason to believe that the information 22 in this statement is inaccurate because of the improper 23 use of the computer. To the best of my knowledge and 24 belief at all material times the computer was operating 25 properly, or if not, any respect in which it was not 169 1 operating properly, or was out of operation was not such 2 as to affect the information held on it. I hold 3 a responsible position in relation to the working of the 4 computer." 5 You told us in your witness statement that that 6 paragraph, which is broadly the first of the two yellow 7 paragraphs in the earlier witness statement, was all 8 about the computer on which you were typing the 9 statement? 10 A. And doing the analysis. 11 Q. What do you mean by "and doing the analysis"? 12 A. I was doing some analysis in the Excel spreadsheets that 13 were then attached to the statement. 14 Q. What is "the computer" referring to, on two occasions in 15 that witness statement? 16 A. The computer on which I was actually doing those. In 17 other words, the PC on my desk. 18 Q. Really? You're saying your desktop was working 19 properly? 20 A. Yeah. 21 Q. So it was typing out the words -- 22 A. Yeah. 23 Q. -- when you typed "All these after perfectly normal 24 occurrences"? 25 A. Yes. 170 1 Q. "The computer" was intended to refer to, here, 2 a computer in the wider sense, wasn't it? It refers to 3 Horizon, doesn't it? 4 A. I don't think I was in a position to say that Horizon 5 was working correctly at the time, as far as this was 6 concerned. 7 Q. Why, given what you'd said in your comment on the yellow 8 highlighted paragraph, "These need to be deleted, 9 I can't say these", has it now re-emerged? 10 A. Because I must have then understood that it was 11 referring to what I was actually producing the statement 12 on. 13 Q. How did you obtain such an understanding? 14 A. I don't know. 15 Q. From whom did you obtain such an understanding? 16 A. I don't know. 17 Q. Can we go back to the email chain, FUJ00152587, and 18 Mr Ward's reply: 19 "I do not understand why this statement, which was 20 originally requested on 10 March is taking so long ... 21 I appreciate it is slightly unusual, but I do not 22 understand the confusion as I thought I'd made our 23 requirements clear. 24 "... Gareth's annotations do not take us forward at 25 all (and I'm sure this is not Gareth's fault). Gareth 171 1 has indicated in the attachment below that 2 3 spreadsheets produced by your team (which show the 3 'NIL' transactions and which were in addition to the 4 usual event and transaction logs) were not produced by 5 him, therefore, as he quite rightly points out, he is 6 not in a position to produce them in his statement." 7 In the event, that's what you ended up doing, 8 producing them in your statement, didn't you? 9 A. But I got it the information from a different source. 10 Q. Ie they were Excel spreadsheets given to you in an email 11 by Penny or Neneh, you didn't think that was 12 appropriate; they were attached to a PEAK, you took them 13 from the PEAK, and thought that was acceptable? 14 A. The spreadsheets weren't attached to the PEAK; it was 15 the message store that would have been attached to the 16 PEAK. 17 Q. Yes -- 18 A. So I had actually extracted the spreadsheets myself and 19 that's what I thought was the difference. 20 Q. He continues: 21 "As already stated, we urgently need a statement 22 producing these 3 additional spreadsheets explaining in 23 general terms under what circumstances 'Nil' 24 transactions occur and in particular how the 'Nil' 25 transactions at Gaerwen occurred ... The same statement 172 1 needs to include a paragraph which states there is no 2 evidence of a system failure ... in relation to the 3 'Nil' transactions at the office. We do not need to 4 mention 'system failures being normal occurrences' if 5 there is no evidence of such a problem at this office." 6 Was it because you were sent this email that you 7 modified your evidence? 8 A. I think that's why I went and extracted the data from 9 PEAK, yes. 10 Q. And you modified the contents of your witness statement? 11 A. Well, to reflect that. 12 Q. Well, and to delete the "system failures are normal 13 occurrences" line? 14 A. Yes. 15 Q. So Mr Ward expressed his concerns on numerous occasions 16 over your wording, including, in your witness statement, 17 "system failures". He told you that system failures 18 were potentially very damaging to mention. He told you 19 that you should not mention them -- "We do not need to 20 mention them" -- and that they were a really poor choice 21 of words. 22 Did all of that pile pressure upon you to remove the 23 reference to "system failures" from your witness 24 statement? 25 A. Clearly, I did that as a result of that, but I was happy 173 1 with the wording that I ended up with. 2 Q. Okay. Did you ever go back to any lawyer within Fujitsu 3 to say, "Look, I'm being embroiled in a criminal 4 prosecution here. I'm being asked to sign witness 5 statements off and there's this man from the Post Office 6 who is applying pressure on me to change the substance 7 of my evidence"? 8 A. No, I didn't, but I clearly understand now that I should 9 have done. 10 MR BEER: Thank you. 11 Sir, we would now turn to the Seema Misra case 12 study. It is probably an appropriate moment to take 13 a break. I will be able, comfortably, to finish my 14 questions tomorrow of Mr Jenkins. 15 SIR WYN WILLIAMS: By "a break", you mean adjourn for the 16 day, Mr Beer? 17 MR BEER: Yes, a long break. 18 SIR WYN WILLIAMS: A long break. Right. 19 All right, that's what we'll do, and we'll start 20 again at 9.45 tomorrow? 21 MR BEER: It is 9.45, please, yes. 22 SIR WYN WILLIAMS: Yes, fine. 23 MR BEER: Thank you. 24 SIR WYN WILLIAMS: Oh, I mentioned last evening, Mr Jenkins, 25 not to talk about your evidence. There shouldn't be any 174 1 secret about it: because there was a late disclosure of 2 documents to you, I gave permission for your lawyer to 3 speak to you about it. 4 THE WITNESS: Thank you. 5 SIR WYN WILLIAMS: But we're now back to where we were 6 before that late disclosure, so don't talk to anyone 7 about your evidence. All right? 8 THE WITNESS: I understand that. Thank you, sir. 9 SIR WYN WILLIAMS: All right. Fine. 10 (3.55 pm) 11 (The hearing adjourned until 9.45 am the following day) 12 13 14 15 16 17 18 19 20 21 22 23 24 25 175 I N D E X GARETH IDRIS JENKINS (continued) ..............1 Questioned by MR BEER (continued) .............1 176