1 Tuesday, 16 May 2023 2 (10.00 am) 3 MR BEER: Good morning, sir, can you see and hear me? 4 SIR WYN WILLIAMS: Yes. 5 MR BEER: May I call Michael Peach, please. 6 MICHAEL EDWARD PRYOR PEACH (affirmed) 7 Questioned by MR BEER 8 MR BEER: Good morning, Mr Peach. My name is Jason Beer and 9 I ask questions on behalf of the Inquiry. Can you give 10 us your full name, please. 11 A. Michael Edward Pryor Peach. 12 Q. I think in the documents you are certainly known as Mik 13 Peach? 14 A. Yes. 15 Q. With Mik being spelt M-I-K; is that right? 16 A. Yes. 17 Q. Thank you very much for coming to the Inquiry to assist 18 it in its work, for previously providing a witness 19 statement to the Inquiry. We're very grateful to you. 20 You should have in front of you a hard copy of the 21 witness statement. It's dated 3 March 2023 and, if you 22 turn to the last page of it, which is page 52, there 23 should be a signature? 24 A. Yes. 25 Q. Is that your signature? 1 1 A. Yes, it is. 2 Q. Before I ask you whether the contents of it are true, 3 can you turn up paragraphs 1 and 7, please. In 4 paragraph 1, second sentence, you say: 5 "I left Fujitsu in September 2009." 6 In paragraph 7, second sentence, you say: 7 "I held this role until I left Fujitsu on 8 30 September 2009." 9 Is there a correction that you would like to make 10 to those two sentences? 11 A. Yes, there is. I now, having reviewed documents from 12 last week, believe that I did not actually leave until 13 December 2009. 14 Q. Is that because we have shown you, amongst other things, 15 some emails with your name on it that post-date 16 September 2009? 17 A. That's correct. 18 Q. I think you heard the evidence of Mr Parker the other 19 day? 20 A. I did. 21 Q. Thank you. 22 I think there's another issue of emphasis that 23 you've raised later in the witness statement -- I won't 24 get you to make the correction now -- we'll deal with 25 that in due course. 2 1 A. Understood. 2 Q. But, subject to those points, are the contents of this 3 witness statement true to the best of your knowledge? 4 A. Yes, they are. 5 Q. Thank you very much. For the purpose of the transcript 6 the URN is WITN04510100. 7 Can I start, please, with your professional 8 experience, career and qualifications. Do you have any 9 professional qualifications that are relevant to the 10 issues that we're to discuss today in your evidence? 11 A. No. 12 Q. I think you joined ICL, as it was then known, in 1980; 13 is that right? 14 A. That's correct. 15 Q. Having worked there for 17 years, you joined the Pathway 16 project or Pathway, as it was then known, in 1997; is 17 that right? 18 A. Yes, I did. 19 Q. You joined as manager of the SSC. 20 A. Yes. 21 Q. A role which you occupied for 12 or so years until you 22 left in December 2009? 23 A. Yes. 24 Q. When you were manager of the SSC, did the person to whom 25 you reported remain the same over that 12-year period? 3 1 A. No. 2 Q. Was the identity of their role the same, ie their job 3 function -- 4 A. No. 5 Q. -- or did that change? 6 A. That changed as well. 7 Q. Can you tell us first by job function and then by name, 8 if you can remember it, who the relevant report was? 9 A. Initially I reported to Stephen Muchow, who was Customer 10 Service Director. At later times, I reported to the 11 Support Services Manager, who -- they reported to the 12 Customer Service Director. 13 Q. Who was the Support Services Manager? 14 A. There were a number: Peter Burden, Carl Marx, Andy Hall, 15 Naomi Elliot, at which point I've run out. There were 16 more. 17 Q. Okay. Can you remember roughly in the 12-year period 18 when the change over occurred, from when you were 19 reporting straight into a director and then when there 20 was somebody who was between you and a director? 21 A. The first change occurred under Stephen Muchow, so it 22 would have been round about 1999. There were times at 23 later dates when the structure changed and I reported to 24 the CS Director, other times when I reported to the 25 Support Services Manager. So it wasn't a consistent 4 1 move down the organisation: it was move down, move up, 2 move down, move up. 3 Q. I see. I understand. How many staff did you manage in 4 the SSC? 5 A. Initially six, later moving to between 25 and 30. 6 Q. We've heard that there was a flat reporting structure 7 with everyone reporting to you; is that accurate? 8 A. That's correct. 9 Q. I think, however, Mr Parker, Steve Parker, was nominally 10 your deputy and, in particular, he deputised for you 11 when you were away; is that also correct? 12 A. That's correct. 13 Q. Did he perform any other roles as deputy manager? 14 A. Not that I can think of. 15 Q. Can we look at the role of the SSC, please. We've heard 16 a lot of evidence about this already, so I'm going to 17 take things relatively briefly. Can we do so through 18 a document, FUJ00119994. 19 You should have in front of you a document called 20 "End to End Support Process, Operational Level 21 Agreement", dated 10 October 1999 as version 1 and, if 22 we just scroll down a little bit, please, we can see 23 that the author of it is you. 24 A. That's correct. 25 Q. This version is marked as a draft. If we go over the 5 1 page, please, to "Document control", you can see the 2 provenance of it, when it was first drafted, moving to 3 version 1. Do you know why it would still be marked on 4 the front page as a draft when it seems to have achieved 5 the status of version 1 in document control? 6 A. The version number that I have at the top of the page is 7 1. 8 Q. Yes. If we just go back to the first page, you can see 9 under "Status" it says "Draft"? 10 A. An oversight -- probably mine. 11 Q. Okay. Can we go to page 7, please. This section of the 12 document sets out the responsibilities, I think, of the 13 first and second line support up to the third line 14 support; is that right? 15 A. That's correct. 16 Q. So HSH and SMC obligations up to SSC, third line 17 obligations? 18 A. Yes. 19 Q. If we just go down, please, to (d), the responsibility 20 is said to be, for those two lines of support: 21 "To 'filter' all calls for which the problem is 22 already known to the support community and for which 23 a resolution is already known or has been generated. In 24 this case context the term 'resolution' can take 25 a number of forms, including: 6 1 "The statement that the problem is resolved in 2 release xxx of the Horizon solution. 3 "There is a documented workaround for the problem. 4 "The documentation relating to that part of the 5 system is under review of being changed." 6 Then in bold and italics: 7 "No calls passed to the SSC which are subsequently 8 resolved as known errors, except in cases where the 9 symptoms reported by the customer did not match the 10 symptoms recorded in the known error documentation, and 11 which therefore the HSH/SMC could not reasonably have 12 been expected to find." 13 Could you explain what this direction is to the 14 first and second lines of support, please? 15 A. The structure of the SSC and its function meant that we 16 were supposed to receive from second line only the first 17 instance of a new software problem. The targets 18 throughout this document were aimed at the HSH and SMC 19 to ensure that they did not overload the SSC with calls 20 that they could have filtered themselves. 21 Q. So was this a direction given to them from the very 22 start, to reduce calls related to so-called unknown 23 errors from being diverted and escalated to the SSC and 24 then through to fourth line support? 25 A. That was the intention of the document. I don't think 7 1 I would use the term "from the very start" because HSH 2 and SMC existed before I joined, but the exact 3 relationship and targets placed on them were not there 4 until I wrote this document. 5 Q. I see. So why did you introduce this? 6 A. I think it was -- the first job that Stephen Muchow gave 7 me to do when I arrived was "You need to sort out the 8 relationship between the four lines of support". 9 I based this document on previous experience of 10 supporting VME systems in order to make sure that the 11 SSC weren't overloaded. 12 MR BEER: Before we proceed can I just check, sir, that your 13 camera is working? You appear to have disappeared from 14 our screen -- and your microphone. 15 SIR WYN WILLIAMS: I think I probably mute myself generally 16 and inadvertently stopped the video. Sorry about that. 17 MR BEER: Yes, I think that's what happened, sir. Thank 18 you. 19 Mr Peach, what was the reason, as you understood 20 it, that Mr Muchow said you needed to sort out the 21 relationship between the four tiers of support? 22 A. Because although the four tiers of support were there, 23 the relationship between them had not been adequately 24 documented. 25 Q. Was it working adequately? 8 1 A. There was no live system at that time and my impression 2 was clearly not. 3 Q. Why was it your impression that it wasn't working 4 adequately? 5 A. Because there was no document such as this that defined 6 the relationship between the lines of support. 7 Q. Was this issue -- the passing of calls inappropriately 8 from lines 1 and 2 to line 3 -- an issue that remained 9 over the duration of your time as manager of the SSC? 10 A. No, it improved greatly as first and second line became 11 better trained, properly staffed and as the SSC got more 12 experience with the system. 13 Q. In that period of 12 or so years, were you aware of any 14 inappropriate pressure being placed on first and second 15 line support not to pass calls on to third line support? 16 A. No, I was not aware of such pressure. 17 Q. Can we move on a little bit, please, to FUJ00120446. 18 You will see this is dated 29 January 2001. It's 19 described as the "Customer Support Services Operations 20 Manual". The owner of it is Peter Burden. The author 21 is "Richard Burton, A&TC". Can you recall what that 22 stands for? 23 A. No, sorry. 24 Q. Then it says "Technical Authors" and then Peter Burden. 25 At this time, what function do you think Peter Burden 9 1 would have been performing? 2 A. I think he was the Support Services Manager. 3 Q. So somebody to whom you reported? 4 A. Correct. 5 Q. The distribution at the bottom of the page, second from 6 the bottom "SSC Manager". That's you? 7 A. That's correct. 8 Q. Now, the role of the SSC is set out in this policy 9 document. Can we turn to, please, page 8 and look at 10 paragraph 4.1: 11 "The principles by which the SSC operates are 12 documented in "End-to-End Support Process Operational 13 Level Agreement ..." 14 I think that's the document we just looked at? 15 A. Yes. 16 Q. The reference CS/FSP/006 is the document we looked at: 17 "... which defines the responsibilities of the 18 four levels of support towards each other. This 19 document is effectively a service level agreement 20 between the support units, outlining specific tasks and 21 measures of success. 22 "The aim of the SSC is to provide a support 23 capability to Pathway that resolves technical problems 24 in the minimum time and with the minimum amount of 25 disruption to the service. The SSC aims to provide 10 1 a centre of technical expertise for Customer Service, 2 providing technical advice, guidance and expertise 3 relating to all parts of the Pathway system. 4 "... specifically the SSC has responsibilities to: 5 "First and second ... 6 "Fourth line support." 7 Then 4.1.1: 8 "SSC responsibilities to first and second line 9 support." 10 If we can expand that to shown all 13 obligations, 11 ie look at the next page as well, if possible. Thank 12 you very much. You can see that there are 13 or so 13 obligations set out imposed on the SSC down to first and 14 second line support. What did you understand the idea 15 of this document as opposed to your document was in 16 setting these out in this way? 17 A. My recollection is that this document was a services 18 manual for the whole of Customer Service and that the 19 CS/FSP/006, so the previous document that we looked at, 20 formed the basis of the SSC part of the CS operations 21 manual. 22 Q. So this is looking at all four lines of support? 23 A. This document is the services manual for Customer 24 Service. It's not just the SSC, it's the other units 25 within Customer Service who had their own obligations to 11 1 other people. So this is basically collating all of 2 those into one document. 3 Q. You see obligation number 5 is to: 4 "Ensure that the incident is resolved within the 5 total time allowed by the contract between the customer 6 and Pathway." 7 A. Yes. 8 Q. Were there written service level agreements regulating 9 the work of the SSC setting out times, volumes and other 10 metrics? 11 A. No. As far as I'm aware, there were no SLAs or SLTs in 12 the contract that related to the resolution of software 13 problems. Most of the SLAs and SLTs related to hardware 14 issues and network. 15 Q. So what does this obligation mean then? 16 A. For me, it meant try and keep the SSC on track with the 17 obligations which were stated in the previous document. 18 But, in terms of obligations to the customer in the 19 contract, it has no meaning. 20 Q. Do you know why it's there, if it has no meaning? 21 A. Only because I believe it was extracted from the 22 previous document as one of the SSC's obligations. 23 Q. The previous document being the one we looked at -- 24 A. Yes. 25 Q. -- the one that you drafted? 12 1 A. Yes. 2 Q. Why did you, therefore, include something in a document 3 that had no meaning? 4 A. I didn't write this document. I will have reviewed it 5 but I didn't write it. I think it's probably a common 6 clause that would have included the other units that did 7 have SLAs and SLTs attached to them. 8 Q. This is taking the reader and, therefore, taking the SSC 9 back to the contract as a measure of progress, 10 performance or success, isn't it? 11 A. It is. 12 Q. But you're saying, in fact, to your knowledge, the 13 contract didn't contain such a measure? 14 A. Not for the resolution of software calls, no. 15 Q. Were there SLAs in respect of the responsiveness of the 16 HSH or the SMC? 17 A. Yes, there were. 18 Q. Do you know why there wasn't an equivalent for the SSC? 19 A. No, for certain, no. I believe, however, that most of 20 the SLAs and SLTs related to hardware so there were 21 specific times for engineers to visit post offices to 22 replace counters, et cetera, but I think it was always 23 accepted that, when it came to software problems, any 24 code fix would require extensive testing before it was 25 released to the live estate and would generally be 13 1 included in either a maintenance or a major release. 2 Q. Can we look at obligation 7, please: 3 "[To] Create and maintain a register of known 4 deficiencies within the Pathway system and the solution 5 to these problems, where known." 6 8: 7 "Allow the HSH and SMC access to this register so 8 that they can fulfil their function of filtering out 9 known errors." 10 Does this essentially describe the KEL system? 11 A. Yes, it does. 12 Q. Was the design and creation of KEL the response to 13 obligation 7, essentially -- 14 A. Yes. 15 Q. -- or the manifestation of obligation 7? 16 A. Yes. 17 Q. Did both first and second line support have access to 18 the KEL system? 19 A. Yes, they did. 20 Q. Can we look, please, at 4.1.2 further down the page -- 21 thank you -- which sets out SSC responsibilities to 22 fourth line support. Again, at item 2 there is recorded 23 an obligation to: 24 "Filter out all calls for which the problem is 25 already known to the support community and for which 14 1 a solution is already known or has been generated 2 [including] problems for which the SSC knows 3 a resolution but has not yet incorporated the resolution 4 into the Known Error Log." 5 Is this a common feature of support services; 6 namely, the filtering out at every stage of calls before 7 passage to the next stage? 8 A. Yes, I would describe it as that. 9 Q. Was the SSC, to your knowledge, ever under any pressure 10 to avoid passing problems up to the fourth line of 11 support? 12 A. No. 13 Q. If an issue was resolved under existing KEL guidance, or 14 an existing KEL, or if a problem was referred to the SSC 15 with insufficient evidence, would that be sent to fourth 16 line support for investigation? 17 A. Sorry, could you repeat the two conditions there? 18 Q. Yes. If an issue was thought to be resolved under 19 existing KEL guidance -- it's been referred in to SSC -- 20 A. Right. 21 Q. -- or if the issue had insufficient evidence of a system 22 problem, would that be referred to fourth line support 23 for further investigation? 24 A. It certainly could be if the KEL was believed to not 25 actually fix the problem. There was no restrictions 15 1 placed on any call that we could send to fourth line. 2 I mean, if we believed there was still an underlying 3 issue and it was a code problem, then we would send it 4 to fourth line, regardless of what documentation was 5 there. 6 Q. That document can come down now. 7 You said that "if we believed that it was a code 8 problem". 9 A. Yes. 10 Q. If there was no evidence of a code problem, what would 11 happen then? 12 A. The SSC person who was handling the call would make 13 a judgement about where they thought the problem existed 14 in the system. If it was possible that it was a code 15 problem, then it would still go to fourth line. If it 16 was likely to have been a hardware problem, it would 17 have gone back to the SMC, and so on. 18 Q. Would it be dismissed as a user error or possible user 19 error? 20 A. It would not be dismissed as a user error but it's 21 certainly possible that the SSC staff member could have 22 said, "I believe on the balance of probability that this 23 is most likely to be a user error". Actually, the term 24 we tended to use was "possible user error" not "this is 25 a user error". 16 1 Q. In what you have just said there you've used "on the 2 balance of probabilities it is a user error" -- 3 A. Yes. 4 Q. -- and "it is possible that it is a user error". 5 A. Yes. 6 Q. Do you recognise that there's a difference between 7 those? 8 A. Yes, I do. 9 Q. To what level of satisfaction did an SSC diagnostician 10 need to be satisfied in order to attribute the code 11 "user error" to a problem? 12 A. I could not quote you a percentage on that. I mean, 13 they would need to be fairly certain before passing it 14 back as a user error, as they would need to be fairly 15 certain that something was a code error to pass it 16 through to fourth line. If they were uncertain, they 17 would gather more evidence and diagnose it properly. 18 Q. If there was no evidence that it was a code problem -- 19 A. Right. 20 Q. -- would that cause them to say "Possible user error, 21 refer back to the subpostmaster for more information"? 22 A. Yes, that's certainly possible. 23 Q. What further information -- I realise we're talking at 24 a theoretical level at the moment, without a practical 25 example -- what kind of further information would you 17 1 expect a subpostmaster to provide? 2 A. Recollection of what it was that they had done prior to 3 reporting the error. It's a very difficult area because 4 there was not sufficient diagnostic capability on the 5 counters to examine exactly what the postmaster had 6 done. So, whilst the SSC could take all of the evidence 7 and put it through code or utilities that SSC staff had 8 produced, in order to check the code, what we could 9 never do was find out precisely what the postmaster had 10 done on the counter. 11 Q. You said that there wasn't a sufficient diagnostic 12 facility at the counter level. 13 A. That's correct. 14 Q. Can you just explain what you mean by that? 15 A. A log of keystrokes performed on the counter would have 16 been useful in a number of cases. 17 Q. Can you explain to us what you mean, because we've heard 18 different descriptions of what a keystroke log means, 19 what you mean by a keystroke log? 20 A. A log of every key depression or screen touch that had 21 taken place on the counter. 22 Q. Was it your understanding that that did not exist at 23 all? 24 A. During the time that I was SSC manager, I don't believe 25 it existed. 18 1 Q. What was the greatest level of scrutiny you could give 2 to what had occurred at a counter level? 3 A. The Riposte message store and there were three or four 4 log files that were kept on the counter. Their exact 5 contents I couldn't tell you. 6 Q. I described those, in the past, as recording when 7 a transaction occurred or when you committed something 8 to a stack. 9 A. Yes. 10 Q. That may be imprecise language. Using your language, 11 what would you say those message stores and files 12 recorded? 13 A. The message stores recorded all of the transactions done 14 by the Riposte software and there may well have been 15 a number of other things that I probably never knew. 16 There were also the NT event logs, which is when 17 an application or, indeed, the Microsoft software writes 18 to a log. There were, I believe, at least one, possibly 19 two others, PS Standard Log rings a bell but the 20 contents I couldn't tell you. 21 Q. Do you remember something called there POC log? 22 A. Only because I heard it mentioned when Anne Chambers 23 gave evidence to this Inquiry. 24 Q. Have you got no greater recollection than that? 25 A. No, and it's way too detailed technically for my 19 1 knowledge. 2 Q. So referencing an issue back to -- or referring an issue 3 back to a subpostmaster for the provision of more 4 information and evidence, that was difficult for the 5 subpostmaster -- would that be right -- because they 6 couldn't look at their system and themselves say, "The 7 system shows that I did X, Y and Z"? 8 A. That is correct. 9 Q. If the subpostmaster couldn't produce any more evidence 10 or information as to what had occurred, would the matter 11 then be -- would the PEAK be closed? 12 A. The PEAK would have been closed at the moment that the 13 call went back to SMC and HSH to ask for the further 14 evidence. It would then, if they managed to get the 15 further evidence, would be reopened. 16 Q. I see. So the closure of the call occurred upon 17 reference down. If the subpostmaster didn't come back 18 to first or second line support, was there any 19 obligation on the SSC to follow the call up? 20 A. On the SSC, no. 21 Q. Was there any obligation on first or second line support 22 to follow the call up? 23 A. I don't know HSH and SMC's processes, so I couldn't 24 comment on that. 25 Q. Can I examine, please, moving on a year still further 20 1 into June 2002 now, and look at POL00000877. This is 2 an internal assessment prepared by Fujitsu on 3 11 June 2002 and I think we can see from the second page 4 there's a list of those who were involved in the 5 internal assessment conducted over two days, I think, at 6 Feltham and Bracknell -- sorry, three days at Feltham 7 and Bracknell and we see your name in the list in 8 Customer Services. 9 A. Correct. 10 Q. Going back to the first page, please, and just scrolling 11 down to assessment summary, can you recall what this 12 was, this three-day assessment, at Feltham and 13 Bracknell? 14 A. No, sorry. 15 Q. Just looking at the document now, can you recall what 16 its purpose or function was? 17 A. The format appears to be similar to a BSI audit. I can 18 only assume it was an audit done for compliance with 19 ISO 9001, done internally not through BSI. 20 Q. So an internal audit? 21 A. Oh, yes. 22 Q. Can we look, please, back to the second page and look at 23 the summary. Just scroll down, please. Thank you. 24 The last bullet point on that summary says: 25 "... the main findings, and recommendations ... 21 1 were as follows ... 2 "There is considerable challenge to the Pathway to 3 continue to operate profitably in the context of 4 a demanding customer facing considerable change and 5 costs-reduction in their own business." 6 Did you understand the customer, ie the Post 7 Office, to be a demanding customer? 8 A. Yes. 9 Q. In what way was POL, the Post Office, a demanding 10 customer? 11 A. I would draw that conclusion purely from the number of 12 SLTs in the contract. 13 Q. So it was demanding from the start, as a matter of 14 contract -- 15 A. Oh, yes. 16 Q. -- rather than in the way that it behaved in the course 17 of the extract; is that right? 18 A. The SLTs in the contract were monitored and reported on 19 frequently and, if Fujitsu failed them, then there were 20 financial penalties. From the SSC point of view, that 21 didn't impact us at all. All of that was done by the 22 MSU, the Management Support Unit. For a while, the 23 Management Support Unit and the SSC both reported to the 24 Support Services Manager, so I was aware that the 25 reviews were taking place because there were monthly 22 1 management meetings. 2 Q. Within the SSC, were the considerable challenges? 3 A. With regard to a challenging customer? 4 Q. Yes. 5 A. No, there were technical challenges associated with each 6 call as it came in. 7 Q. This records "a considerable challenge to Pathway to 8 continue to operate profitably", so that's the Horizon 9 System within Fujitsu -- 10 A. Yes. 11 Q. -- being a challenge for it to continue to operate 12 profitably? Did you feel that challenge within the SSC? 13 A. No. 14 Q. What did you understand this to refer to? 15 A. I would have taken this to refer to performance against 16 the SLTs in the contract. 17 Q. Were there any SLTs in the contract that impinged on the 18 work of the SSC? 19 A. The only ones that I can recall were the obligation to 20 pass all transactions through to Post Office within 21 a certain time period. As I said, there were no SLTs 22 that I was aware of relating to the fixing of software 23 calls. 24 Q. Thank you. That can come down. 25 You were in the SSC from 1997 onwards -- 23 1 A. Correct. 2 Q. -- and, therefore, provided third line support whilst 3 the product was being tested and rolled out? 4 A. Yes. 5 Q. Had you been involved in the testing or the provision of 6 line support when other projects had been tested and 7 rolled out? 8 A. Are you referring to my previous time in ICL? 9 Q. Yes. 10 A. Only releases of VME and, when I was managing a Rapid 11 Application Development unit, then we were obviously 12 releasing applications. But those tended to be very 13 small, ten or more users -- certainly not 37,000 users. 14 Q. So nothing of this scale? 15 A. No. 16 Q. So had you got a reference point against which to 17 compare how easy or problematic the provision of 18 a support service was when you were engaged in the 19 provision of such services whilst Horizon was tested and 20 rolled out? 21 A. Only experiences in relation to releases of the VME 22 operating system. 23 Q. What was your experience, speaking in general terms, of 24 the provision of third line support when Horizon was 25 being tested and rolled out, so speaking between '97 and 24 1 mid-2000? 2 A. Sorry, can you explain what you meant by that? 3 Q. Yes. Looking at it in general terms -- 4 A. Right. 5 Q. -- what was your feeling, your impression, your 6 judgement, on how easy or difficult it was to provide 7 third line support between, say, 1997 and mid-2000? 8 A. Initially hard. Lots of inexperience in first and 9 second line of support and, indeed, with postmasters 10 using a completely new system, becoming progressively 11 easier as the different lines of support became more 12 experienced and the KEL system was populated. 13 Q. By mid-2000, was everything running smoothly? 14 A. I think, mid-2000, the rollout had not been completed. 15 I'm not certain when the rollout was completed. 16 Q. Take it by reference to the end of the rollout period 17 then. 18 A. At the end of the rollout period, it was already 19 beginning to become easier. 20 Q. Did you have a view as to the robustness and reliability 21 of the Horizon System by the end of rollout? 22 A. That's very difficult for someone in support to answer. 23 Nobody ever phones you to tell you the system is working 24 properly. 25 Q. Sorry, can you say that sentence again, please? 25 1 A. Nobody ever phones you to tell you the system were 2 working properly; you are constantly phoned when it's 3 not. So you obviously get a fairly jaundiced view. 4 Having said that, I would have described it as 5 generally working the way I would have expected it to 6 work. That sounds very vague, I know. I'm sorry. 7 I don't think I can be more precise. 8 Q. Do you remember Richard Roll? 9 A. Yes, I do. 10 Q. He worked in the SSC between January 2001 and August 11 2004. 12 A. Yes. 13 Q. So a period of about three and a half years and you 14 would have been his manager for the entirety of that 15 period? 16 A. I would. 17 Q. With Mr Parker acting as your deputy? 18 A. Correct. 19 Q. Mr Roll told the chair that: 20 "It was widely accepted that the underlying or 21 root cause [that was with problems with the system] were 22 that the system was crap, it needed rewriting but that 23 that was never going to happen because the money was not 24 available, the resources were not available to do so." 25 In that period, would you say that was a common 26 1 view: that the system was "crap"? 2 A. No. 3 Q. Was it widely accepted within the SSC that the system 4 was "crap" and "needed rewriting"? 5 A. No. I think what Richard failed to understand was that, 6 by the time that the code gets to the live estate, it 7 has already been through extensive testing and 8 acceptance formally by Post Office. So, essentially, 9 from the point of view of the support teams, that's the 10 code. There is no point in saying "I want this 11 completely rewritten", because it's already been through 12 a testing and acceptance process. 13 Q. What about if it's gone through the testing and 14 acceptance process and things have been papered over and 15 a decision has been made to proceed with a system that 16 is riddled with faults? 17 A. I would not use the term "riddled with faults". The 18 acceptance criteria, as specified between development 19 testing teams and the customer, would indicate that you 20 do not take the product to live based on a number of 21 criteria and those criteria would be things like no more 22 than "N" A priority calls outstanding, "X" B priority 23 calls, et cetera. So it would not, in my opinion, be 24 "crap" when it went out to the live estate. 25 Q. Are you saying that because a system has been accepted 27 1 it cannot have faults? 2 A. No, I'm saying that, because it has been accepted, the 3 number of acceptable faults, as agreed between Fujitsu 4 and the customer, would have been defined. Obviously 5 when it goes out to the live estate it has faults. 6 Every software has faults. 7 Q. What if the customer had decided to rewrite the 8 acceptance criteria a number of times to progressively 9 allow more and more faults to be present in the system 10 because there was pressure on the customer to move to 11 acceptance? 12 A. I have no knowledge of that taking place. 13 Q. Were you aware of variations to the contract between 14 Fujitsu and the Post Office -- 15 A. No. 16 Q. -- in 1999 and 2000 -- 17 A. No. 18 Q. -- where exactly that occurred? 19 A. No. 20 Q. Can we look please at WITN04600104. Thank you. 21 This is a document that you weren't copied in on 22 or were not an author nor a reviewer. It's dated 23 10 May 2000 and you'll see from the "Abstract" it 24 presents the observations and recommendations resulting 25 from an internal audit, along with agreed corrective 28 1 action, the action owner and the date by which the 2 action is to be complete. 3 If we go to page 9, please, you'll see in the top 4 left-hand box, against the reference 4.2.1, it is 5 recorded that: 6 "The audit identified that the EPOSS continues to 7 be unstable. PinICL evidence illustrated the numbers of 8 PinICLs raised since the 1998 Task Force and the rate of 9 their being raised. 10 "The EPOSS solutions report made specific 11 recommendations to consider the redesign and rewrite of 12 EPOSS, in part or in whole, to address the then known 13 shortcomings. In light of the continued evidence of 14 poor product quality these recommendations should be 15 reconsidered." 16 Did you know that there had been something called 17 the EPOSS Task Force? 18 A. No. 19 Q. Did you known that a report had been produced 20 recommending the consideration of the total or partial 21 rewrite and redesign of EPOSS? 22 A. No. 23 Q. Did you know that in May 2000 there had been 24 a recommendation by this internal audit that those 25 selfsame recommendations in the light of continued poor 29 1 product quality should be reconsidered? 2 A. No. 3 Q. Are those facts and matters of which you ought to have 4 known? 5 A. I don't think so. From what I can see from this 6 document, it's an internal discussion between 7 Development and Testing as to the state of the product 8 before it goes to the live estate. I would not have 9 been involved in any decisions that were taken on this 10 nor would I have expected to be. 11 Q. After this time, May 2000, you find within third line 12 support a preponderance of problems with EPOSS? 13 A. I can't recall and don't have the figures to tell you 14 how much was counter-based problems and how much was 15 central systems problems. 16 Q. Thank you. That can come down. 17 Going back to what Mr Roll told the Inquiry, he 18 said that, rather than a redesign and rewrite, which was 19 never going to happen on cost grounds, the SSC was left 20 to seek to patch up with the Development team the system 21 on an ad hoc basis. Is that accurate? 22 A. I don't agree with Richard's comments. I don't agree 23 with his initial premise. He didn't know -- I mean, 24 I didn't know the head count or the development budget. 25 I'm quite certain he didn't. So saying that it's all 30 1 due to lack of money or lack of resources, as far as I'm 2 concerned, is supposition on his behalf. When it comes 3 to the statement "The SSC were patching up things", 4 examining the cause of problems and fixing them is what 5 the support team did. 6 Q. He told the Inquiry: 7 "The software issues we were routinely 8 encountering could and did cause financial discrepancies 9 at branch level, including shortfalls being incorrectly 10 shown on the Horizon System." 11 Does that accord with your recollection? 12 A. There were certainly bugs in the system that could cause 13 those symptoms, yes. 14 Q. He told us that the Horizon cash accounts were, in his 15 words, "pretty ropey" and that he told you that, adding, 16 "Surely, these should be rewritten", and you agreed with 17 him and said: 18 "Yes, but it's never going to happen." 19 Is that accurate? 20 A. I don't recall that conversation at all. 21 Q. Are you saying, through the passage of time, it might 22 have occurred but you now do not recall or that, given 23 your view of the quality of the Horizon System, it is 24 something that is unlikely to have happened? 25 A. I am saying that an individual member of the SSC may 31 1 have expressed reservations of the code but I don't 2 recall the conversation, so I can't give you a reason 3 why I may have said what he believes I said. 4 Q. Mr Roll told us that: 5 "If we in the third line support were unable to 6 find the cause of a problem, this was reported up the 7 chain to fourth line but it was assumed that the 8 postmaster was to blame." 9 Was that a practice of which you were aware? 10 A. Absolutely not. 11 Q. Was it a common theme throughout the time that you were 12 the head of the SSC, that if positive evidence of 13 a software fault could not be found, it was assumed that 14 the subpostmaster was to blame and that's how it was 15 written up? 16 A. No, on two grounds. Firstly, whenever any call came in 17 I expected people to look at all the evidence and 18 diagnose it properly and that means you have no fixed 19 starting position. You don't assume from the beginning 20 that it's a user error, you don't assume it's a software 21 bug. 22 Secondly, we, to my knowledge, never used blame. 23 Even when calls were being returned as possible user 24 error, that could mean any number of things. It could 25 mean that documentation at the Post Office wasn't 32 1 accurate, hadn't been followed -- it's not a question of 2 blame. 3 Q. Mr Roll told the Inquiry that: 4 "Sometimes we were instructed not to let the 5 subpostmaster know that we had altered his system whilst 6 he was logged on. To my recollection, sometimes the 7 Post Office requested this, sometimes Fujitsu and 8 sometimes only our department knew of it." 9 Did you ever give any instructions not to inform 10 subpostmasters to tell them that their system had been 11 altered whilst they had been logged on? 12 A. No, I didn't give instructions of that sort. 13 Q. Were you aware of that practice? 14 A. I have become aware through a couple of documents that 15 I was sent to review by this Inquiry that somebody in 16 Post Office management had said "Don't tell the 17 subpostmaster about this". But, as far as I can see 18 from the documents that I've been supplied, there appear 19 to be two instances of it which were sent to my staff 20 and, without knowing the reason behind, I wouldn't like 21 to comment on that. 22 Q. Mr Roll told the Inquiry, and I'm afraid this is a long 23 quote: 24 "I recall one particular case where branch data 25 was not being replicated from a mobile Post Office 33 1 correctly and it appeared that the subpostmistress was 2 turning off the power mid-transaction. As we couldn't 3 fix the problem over the phone with the subpostmistress 4 she sent her laptop to Fujitsu for examination. Using 5 Post Office tests rigs on the 6th floor and comparing 6 the results with the laptop that had been returned to 7 Fujitsu, I discovered that the button which should have 8 put the laptop into standby mode was actually switching 9 off the power resulting in the disk crashing. 10 I disassembled the laptop to confirm this. Thus when 11 the posts mistress thought she was switching her counter 12 to standby mode, which would have initiated a controlled 13 shut down and allowed the data store to replicate the 14 servers, she was actually switching about power off, 15 which is what we were seeing in the SSC. 16 "When I raised this with my manager, Mik Peach, 17 who subsequently talked to the hardware team, I found 18 this was a known problem. One of the engineers had made 19 a mistake with a batch of laptops which had been sent 20 out to branches before the error was detected. No-one 21 outside the team responsible for building the laptops 22 had been informed of this. This meant I spent several 23 days investigating the problem. Whereas the 24 subpostmistress in this case was provided with 25 a replacement laptop, knowledge of this problem was kept 34 1 within the departments concerned and the batch of faulty 2 laptops was not recalled. It's my belief that Fujitsu 3 senior management and the Post Office was not informed." 4 Do you remember that incident? 5 A. From the time that it happened, no. From the Group 6 Litigation, yes, because I was called during that 7 trial -- not to go to the trial but I was telephoned and 8 asked if I remembered a specific hardware call from that 9 period. So "no" was the answer that I gave at that 10 time. I am aware of it now because I've read Richard 11 Roll's testimony in court and his appearance at this 12 Inquiry. 13 Q. What do you now recall then about the incident? 14 A. I've read through the original call and it's clear from 15 that -- I believe it says on it something like "This is 16 happening six minutes before POLO", which is Post Office 17 Log On". Since it's happening before the postmaster has 18 logged on, then no financial transactions can have been 19 impacted. 20 Secondly, he made comments that I had talked to 21 the hardware manager, which is certainly possible. 22 Q. He said "My manager, Mik Peach, knew. His friend who 23 ran the build team knew". Is that what you are 24 referring to? 25 A. That's what he said. To be clear, the lady that was 35 1 running or was our contact for hardware was based in 2 Stevenage. I was based in Bracknell. I don't think -- 3 I don't think we ever met face-to-face and we certainly 4 didn't meet socially until about five years after I'd 5 left Fujitsu. So to say I was doing her a favour as 6 a friend is his interpretation and, in my opinion, 7 nonsense. 8 Q. He said that it never got up the chain beyond the pair 9 of you, that he was told to hush it up. I asked him 10 "Who told you to hush it up" and he said you. Is that 11 accurate? 12 A. No. To be specific, if I had phoned the hardware 13 manager and was doing her the favour of hushing it up, 14 then the first person I would not have told would be 15 Richard Roll. I mean, if I would have wanted to hush it 16 up, I wouldn't have informed him of what had happened 17 and, in any case, as I've said in evidence to this 18 Inquiry, I told senior managers about that issue in my 19 monthly report that month. 20 Q. Why would he be the last person you'd tell? Was he 21 problematic? 22 A. No, just if I was going to hush it up, I just would not 23 have told him what had happened. 24 Q. Can we turn, please, to FUJ00087994. Can you see this 25 is a "Group Definitions" document for the secure NT 36 1 build release 2, dated 22 December 1998, yes? 2 A. Yes. 3 Q. If we just read the "Abstract": 4 "The ACP requires that access to Pathway systems 5 be controlled by the use of pre-defined roles to which 6 users can be assigned. Such roles will allow users to 7 access only those parts of the system, with associated 8 objects, they need in order to complete the tasks 9 associated with that particular role. This document 10 summarises this requirement and defines the roles, with 11 associated objects, domains and access requirements." 12 We can see that if we scroll down a little bit 13 you're amongst the distributees? 14 A. Right. 15 Q. Looking at this document, can you summarise, even having 16 read the abstract, what its purpose is? I think 17 I understand but can you help us to translate the 18 delightful language used? 19 A. Can you give me a moment to read that summary? 20 Q. Yes. 21 A. Okay, as I understand it, it's a way of setting up 22 Windows NT systems with defined roles each of which will 23 have defined access to the system and how the setup of 24 those roles should be achieved. 25 Q. So it's a means of writing into the system limitations 37 1 on the access rights of users? 2 A. Correct. 3 Q. Permissions, one might call it? 4 A. Indeed. 5 Q. This kind of document and this kind of approach is 6 natural in a system of this kind? 7 A. Yes. 8 Q. One might say essential? 9 A. I would say essential, yes. 10 Q. Why would you say essential? 11 A. Because you have to be clearly able to decide who is 12 accessing what and why. 13 Q. Why do you have to be able to identify who is accessing 14 what and why? 15 A. Partly because there will be contractual requirements, 16 partly because you have a need to establish an audit 17 trail for support people and what they're doing. 18 Q. Why would you need to establish an audit trail to see 19 what people are doing? 20 A. I would just regard that as being an essential part of 21 any system. Why -- I could not explain why. Just all 22 the systems I've ever worked on behave that way. It's 23 just natural. 24 Q. Just think about it a little more. Why in a system that 25 concerns financial data, for example -- 38 1 A. Yes. 2 Q. -- might you need a system of access limitations, 3 permissions and auditability after the event? 4 A. You would need them there in order to write an audit 5 trail. If you needed an audit trail, then I would 6 assume that it would be because of some form of possible 7 litigation after the event. 8 Q. What would you have in mind there, some litigation after 9 the event? 10 A. I really can't answer that. 11 Q. Speaking generally at your first couple of years, maybe 12 even further, maybe into 2000, 2001, 2002, were you 13 aware that the financial data produced by Horizon was 14 used as the basis for bringing civil and criminal 15 proceedings against subpostmasters? 16 A. No, I was not. 17 Q. When did you first become aware that the Horizon data 18 was used as the foundation for criminal proceedings or 19 civil proceedings? 20 A. When Anne Chambers went to court in what I subsequently 21 found was the Lee Castleton case. 22 Q. So about 2006? 23 A. Yes. 24 Q. Had anyone before then explained to you that one of the 25 reasons why audit or auditability of the system might be 39 1 essential was for that reason? 2 A. No. 3 Q. Can we look, please, at FUJ00088082. Can you see this 4 is a document dated 2003? 5 A. I can. 6 Q. So we've previously looked at Mr D'Alvarez's document of 7 December '98 saying this is what we need to do, these 8 are the access rights and permissions that need to be 9 written in, and these are the reasons why they need to 10 be written in. 11 A. Yes. 12 Q. Looking at again the abstract of this document, it 13 describes the support and use of OpenSSH. Can you now 14 recall what OpenSSH was? 15 A. It was a piece of software that provided secure access 16 to the system for the support teams which was both 17 secure and auditable. 18 Q. When was it introduced? 19 A. I'm not certain in terms of dates. I know it was 20 introduced with the Network Banking release of the 21 Horizon software because this product required software 22 on all of the counters as well as in the central 23 systems. 24 Q. Can we look please at page 15 and paragraph 7.1. This 25 is under "Permissions Problems": 40 1 "When attempting to diagnose problems with OpenSSH 2 ... it should be noted that permissions displayed by 3 OpenSSH don't necessarily reflect the full set of 4 permissions applied by Windows. This is because the 5 rich set of permissions supported by Windows with access 6 specified individually for multiple users and groups 7 cannot generally be mapped to the simple user group 8 other model offered by POSIX. Hence OpenSSH will 9 generally only display an approximation of the 10 permissions in POSIX form but will usually apply the 11 full set of Windows permissions. The permissions 12 displayed and applied are also affected by the setting 13 of the CYGWIN environment variable. As a result, you 14 should not rely on the permissions information displayed 15 in CYGWIN commands such as ..." and then an example is 16 given. 17 Can you translate what that means, please? 18 A. No. Most of those terms mean nothing to me at all. 19 That's way too technical for me. 20 Q. Does that reflect that you were a manager and, 21 therefore, managed people rather than carried out any 22 technical work yourself? 23 A. I didn't carry out technical work on the live Horizon 24 System at all and this sort of document, had I received 25 it for review, I would have passed to one of my 41 1 technical staff. 2 Q. We saw on the front page that it was distributed to you. 3 A. Right. 4 Q. And you were a mandatory review authority? 5 A. Yes. 6 Q. We can see that from the second page against your name. 7 Perhaps we should just look at page 2, please. 8 "Mandatory Review Authority, "Mik Peach", and then it's 9 got "by proxy". Does that reflect what you have just 10 said, that you would have got somebody else to do it? 11 A. Yes. 12 Q. Who amongst your team would you habitually pass these 13 things down to? 14 A. One of the five senior people. 15 Q. Who were they? 16 A. Steve Parker, Anne Chambers, Pat Carroll, Mark Wright, 17 John Simpkins. 18 Q. So if they did reply here -- and it looks like they did 19 because there's an asterisk against your name -- 20 A. Yes. 21 Q. -- it would have been one of those five? 22 A. It would have been one of those five. 23 MR BEER: Sir, that's an appropriate moment. I'm about to 24 move to a new topic. I wonder whether we could come 25 back at 11.25, please? 42 1 SIR WYN WILLIAMS: Yes, of course. That's fine. 2 MR BEER: Thank you very much, sir. 3 (11.10 am) 4 (A short break) 5 (11.25 am) 6 MR BEER: Sir, good morning. Can you see and hear me? 7 SIR WYN WILLIAMS: Yes, I can thank you. 8 MR BEER: Mr Peach, can we turn, please, to a passage in 9 your witness statement. It's paragraph 123, which is on 10 page 41. If we just blow up paragraph 123 -- thank 11 you -- you say: 12 "If a postmaster made a mistake, a transaction 13 could be 'reversed' (by inserting a 'reversal' or 14 'corrective' transaction) but it could not be deleted. 15 There were processes by which SSC staff could, under 16 instruction or approval from POL and with assistance 17 from the postmaster, insert corrective transactions and 18 I recall that there were processes in place to control 19 this rare occurrence, involving dual-person sign-off on 20 the PEAK and approved OCP requests for the SSC to do the 21 work, which I believe had been to be approved by POL as 22 well as Customer Service. An example of this process is 23 OCP 21918 ... dated 2 March", and you give the 24 reference: 25 "my recollection is that the process was 43 1 technically complex and could only be done in agreement 2 with the postmaster and was extremely rare." 3 So you are saying that it was very an extremely 4 rare occurrence which could only be done with the 5 agreement of the subpostmaster, with the knowledge and 6 approval of POL itself and Customer Services? 7 A. Correct. 8 Q. Can we just look then at the example that you give or 9 the reference you give there, which is FUJ00084131. We 10 can see the OCP number that you previously gave, 21918. 11 A. Yes. 12 Q. The title of the OCP "Insert corrective transactions at 13 branch 382137", and, if we can just read through this, 14 we haven't seen many of these before so I want to use 15 you to look at this. 16 A. Okay. 17 Q. "A set of unbalanced SC currency transactions were 18 written in error at branch [then the number is given] on 19 18 February. A set of equal but opposite transactions 20 is to be inserted to undo the effects. 21 "Justification: Enables the branch to balance 22 correctly, and data in POL FS will also be correct." 23 The date when it is going to be done by is set 24 out. 25 "Extra detail: Tested within SSC and proved to 44 1 generate a further [receipts and payments] mismatch 2 which negates the first, and also a gain to negate the 3 loss of just under £1,000 caused by the problem and 4 currently outstanding at the branch. The gain may not 5 be precisely the same as the original loss because of 6 variations in the exchange rates. 7 "POL (Julie Edgley) have already agreed to the 8 change, in an email attached to ..." and then the PEAK 9 number is given, yes? 10 A. Yes. 11 Q. "Regression: This change cannot be regressed." 12 Then further down, the email is in the comments 13 section at the bottom, I think: 14 "Anne, 15 "As discussed, POL are happy for you to make the 16 necessary system adjustments. 17 "From speaking to Wendy, the manager in the 18 branch, first thing on Tuesday morning (between 9 am and 19 10 am) is the quietest for them. 20 "I have advised Wendy that you will call her as 21 you are about to start and as you finish. 22 "Thank you." 23 So there is a record -- I mean, if we just look at 24 the second page of the document, POL approve this 25 change. Then scroll down to the foot of the page. 45 1 So there is a record in there that Julie -- that's 2 Julie Edgley, who was a live service assistant in POL 3 Service Delivery -- had spoken to the subpostmistress? 4 A. Yes. 5 Q. There is a record, therefore, that POL had agreed to the 6 change? 7 A. Correct. 8 Q. Was that always the case? 9 A. I can't recall any occurrences where it was not. 10 Q. Was it always supposed to be the case? 11 A. Absolutely. 12 Q. Who would write this document, the OCP? 13 A. This OCP -- I'm not sure what Gaby Reynolds' exact 14 position was at that time but she would be the liaison 15 between Post Office and the SSC. So she would be 16 acting, effectively, as a problem manager for this 17 instance. 18 Q. She was a Fujitsu employee? 19 A. Yes, she was. I'm not -- could we go back up? 20 Q. Yes, to page 1, please. 21 A. Yes. 22 Q. Look at the bottom half of the page. 23 A. Okay, I'm not aware of who actually raised the OCP 24 itself. Certainly at times, an OCR was used, rather 25 than an OCP. Has the Inquiry been told the difference? 46 1 Q. Yes, it has. 2 A. If it was to be an OCR to insert a transaction at 3 a branch, it would have been written by the SSC person 4 who was going to do the work because they would have 5 received the PEAK which highlighted the error. So they 6 would raise the OCR and it would then go to POL for 7 their approval and subsequently to me for sign off 8 before the work was done. 9 Q. How would you sign it off? 10 A. My recollection is there was -- electronically on the 11 form, just by sitting at my PC and putting my name in. 12 Q. Just go to the foot of page 2, please. You see 13 "Approval status" there. There appear to be some what 14 might, on a screen, be tiles to click on. 15 A. Yes. 16 Q. Am I right in thinking that they might be a printed 17 version of a tile to click on? 18 A. Yes. I think with this particular one it's an OCP and 19 the reason that we used OCRs more frequently was that 20 there are mandatory approvals on an OCP, which were not 21 relevant for an OCR: POA, Core Services SMC, for example 22 would not be required to approve a change of this sort. 23 Q. So this appears to be evidence of in support of what you 24 were saying in paragraph 123 of your witness statement; 25 namely, POL sign off and branch knowledge and agreement? 47 1 A. Yes. 2 Q. Could we look at FUJ00087194, please. Different OCP 3 17510: 4 "Write corrective bureau message for ..." and then 5 a branch code is given: 6 "A single SC message 183227 [et cetera] was 7 written in error on 26 November ... selling 1,000 US 8 dollars, with no corresponding settlement line. To 9 remove the effects of this message at both the branch 10 and on POLFS, we will insert a new message to negate the 11 effects of the original message. 12 "Justification: If the change is not made in the 13 counter messagestore (before the stock unit is balanced 14 on Wednesday), the branch will have an unexpected gain 15 of £484 (or thereabouts -- depends on exchange rate), 16 and a receipts and payments mismatch. This gain would 17 have to be resolved at the branch. There would also be 18 an inconsistency between the branch and POLFS to be 19 resolved. By correcting the problem locally, the branch 20 may not be aware of the problem, and there will be no 21 inconsistency between the branch and POLFS." 22 Then when it's planned for, some extra detail is 23 given. Then scroll down, please: 24 "The message will include a comment to show it has 25 been inserted to resolve this problem (this will not be 48 1 visible to the branch)." 2 Them there's some more detail. This appears to 3 suggest that a correction was to be made and made 4 deliberately, in a way that ensured that the branch was 5 not aware of the problem. 6 A. I'm not convinced that the wording of that means that 7 the branch were not aware there was a problem. 8 Certainly POL, as is stated there, were aware of the 9 problem. 10 Q. Yes, I'm focusing on the branch. 11 A. Okay. Okay, I don't think it's clear from the wording 12 whether the problem was not visible to the branch or 13 whether the comment that would be inserted into the 14 message would not be visible to the branch. 15 Q. Let's take it in stages. Do you agree that there's no 16 record on this document of the branch being informed of 17 the nature of the error -- 18 A. Yes, I agree -- 19 Q. -- the cause of the error and the way in which it's 20 going to be corrected? 21 A. I agree that there's no evidence in this document of 22 that. 23 Q. From what you said, there should be, shouldn't there? 24 A. I would have expected there to be, yes. 25 Q. So there should be? 49 1 A. Yes. 2 Q. Because, in your witness statement at paragraph 123, you 3 told us this was the system and you showed us exhibited 4 an OCP, which was evidence that the system was working? 5 A. Correct. 6 Q. So this is evidence of something different than that, 7 isn't it? 8 A. It's evidence that -- there is no evidence on this 9 document which suggests that the process was followed 10 fully. 11 Q. And, indeed, there's some evidence to suggest that, if 12 we just scroll up to "Justification", the last line of 13 "Justification": 14 "By correcting the problem locally, the branch may 15 not be aware of the problem ..." 16 A. I agree. 17 Q. Under "Extra detail", second paragraph: 18 "The message will include a comment to show it has 19 been inserted to resolve this problem (this will not be 20 visible to the branch)." 21 Can you think of reasons for recording the fact 22 that the branch will not be aware of the problem and the 23 message to correct the problem will not be visible to 24 the branch? Why would it be important to record those? 25 A. I don't have an explanation for that. 50 1 Q. Might it seem that they were positives: it's a good 2 thing that the branch won't be told and can't see? 3 A. In my opinion, whether or not the branch would know that 4 they had a problem is not a reason for not telling them. 5 Did that makes sense? Was that ... 6 Q. Yes, I understood it. But that's an answer to 7 a different question. I'm asking why somebody would 8 record in two places on this OCP? 9 A. I don't have an explanation for that. In the second 10 part where it's under the "Extra detail", the comment, 11 "The message will include a comment to show it has been 12 inserted", was part of the standard procedure from the 13 SSC. When inserting a message into a counter message 14 store there would be an addition made to the Riposte 15 message which said something along the lines of 16 "inserted by SSC to resolve PEAK thing". 17 So that second comment saying, "will include 18 a comment to show it has been inserted to resolve this 19 problem (this will not be visible to the branch)", 20 I would take to mean that that message inserted into the 21 Riposte message would not be visible to the branch. 22 Q. Yes. Can we turn back to paragraph 141 of your witness 23 statement, please, which is on page 45. It's the bottom 24 half of the page, thank you. You say: 25 "The purpose of the System Outline Design [this is 51 1 a document I took you to earlier] seems to be to specify 2 a tool set for different support units to enable them to 3 continue to support the systems, and to be fully 4 auditable. The System Outline Design resulted in the 5 use of SSH software, which was fully auditable -- 6 I believe via the audit servers, which were not 7 accessible by the SSC." 8 In that last sentence there, you say that the SSH 9 software was fully auditable? 10 A. Yes. 11 Q. In what respect or respects was it fully auditable? 12 A. The SSC would log in to the secure access servers in the 13 data centres and that was the sole route down to 14 counters. On the secure access servers, every keystroke 15 that was typed on the SSC work station was recorded in 16 a file and then I believe that file was sent down to the 17 audit servers. So, effectively, every keystroke on 18 every SSC work station was recorded. 19 Q. So there was a full keystroke record when members of the 20 SSC used the SSH software? 21 A. Yes, absolutely. 22 Q. You said that you believe it was sent down to audit; is 23 that right? 24 A. Yes. 25 Q. Automatically sent down to audit -- 52 1 A. Yes. 2 Q. -- or periodically? 3 A. I don't know the exact mechanism. I just remember 4 seeing a design document that said the files are held on 5 the SAS servers and then transferred to audit. 6 Q. Was that audit trail ever examined, to your knowledge, 7 ie to look at the keystrokes made by SSC staff? 8 A. I know that SSC did not support the audit server and did 9 not have direct access to it, so it would never have 10 been viewed by SSC staff. Whether or not it was viewed 11 by other staff, I have no knowledge. 12 Q. When you became aware that there were prosecutions and 13 civil proceedings based on Horizon data -- 14 A. Yes. 15 Q. -- were you aware of the SSH audit files ever being 16 accessed for those purposes? 17 A. My recollection is that I only ever knew of one case and 18 that was the one which involved Anne Chambers, and I was 19 not aware that the audit data was being used for 20 prosecutions at all. Does that answer the question or 21 is that ... 22 Q. Well, if you were only ever aware in your 12 years of 23 one case -- 24 A. One prosecution, yes. 25 Q. One civil proceeding? 53 1 A. Yes. 2 Q. Were you aware of an individual called Andrew Dunks? 3 A. Yes. 4 Q. Andy Dunks? 5 A. Yes. 6 Q. What's your recollection of where he worked? 7 A. My recollection is that he worked in the Security team 8 inside Customer Service. 9 Q. So the Customer Services POA Security Team? 10 A. Yes. 11 Q. CSPOA Security. What did you understand his job 12 function to be? 13 A. I don't recall knowing what his job function was. 14 Q. We understand that he was said to be the cryptographic 15 key manager. Does that ring any bells? 16 A. Yes. In addition to not having access to the audit 17 server, SSC did not have access to a key management 18 server, both of which, my understanding is, was 19 controlled by the Security team. So Andy would have 20 controlled the work of the key management applications 21 on that server. 22 Q. How frequent was your contact with Mr Dunks? 23 A. Difficult to say. My recollection says perhaps once 24 a month. 25 Q. Were you aware that Mr Dunks had contacts with members 54 1 of the SSC? 2 A. Yes. 3 Q. What was the nature and content of such contact, the 4 purpose of it? 5 A. I don't remember. 6 Q. How frequently would Mr Dunks be in contact with members 7 of your team? 8 A. That would be, to my recollection, once/twice a month. 9 Q. You didn't know what they were talking about or 10 exchanging emails or other communications about? 11 A. Not that I recall. I recall that the key management 12 server was kept in a locked room inside the secure area 13 in the SSC and, therefore, whenever Andy had to do some 14 work on that server somebody would have to let him into 15 the secure area. 16 Q. We've heard from Mr Dunks that he produced written 17 witness statements, so evidence -- 18 A. Right -- 19 Q. -- in written witness statements and exhibits to those 20 witness statements, for the purposes of taking criminal 21 proceedings against subpostmasters. Do you understand? 22 A. I understand. 23 Q. Did you know, in your decade or so of working in the 24 SSC, that that was part of his job? 25 A. I knew that there was a function inside the Security 55 1 team which was litigation support. I don't recall ever 2 associating that function with Andy Dunks. 3 Q. What did you understand litigation support did? 4 A. My understanding of that, which is very limited, was 5 that they were there to, in my mind, protect Customer 6 Service from possible litigation from outside. I was 7 not aware that they were acting in prosecutions of 8 postmasters. 9 Q. So you didn't know they were supporting litigation, 10 rather than defending against litigation? 11 A. I don't think I ever thought of it in those terms. 12 Q. In any event, we've heard from Mr Dunks that he produced 13 witness evidence and exhibits for the purposes of 14 criminal proceedings against many subpostmasters and 15 mistresses. You didn't know that that was his job or 16 part of his job? 17 A. Not that I can recall. 18 Q. I think it follows that you wouldn't know why he, the 19 Crypto Key Manager, had been selected to be the witness 20 that produced evidence against subpostmasters? 21 A. Him specifically, no, but he would have been one of the 22 few people that had access to the audit servers, so, as 23 a function of the Security team, I can understand it but 24 I would not have associated it with one individual. 25 Q. You referred to one of the few people that would have 56 1 had access to the audit servers. 2 A. Yes. 3 Q. What are you referring to as the audit servers there? 4 A. The audit servers were holding data from the system 5 which, I believe, included data from Riposte and from 6 all the SSC workstations. I didn't really get involved 7 with what the function of that server was because SSC 8 were not allowed to touch it and we didn't support it. 9 Q. I think it follows that you didn't know that Mr Dunks 10 was providing witness statements for the purposes of 11 prosecutions that made assertions, the witness 12 statements, that were, in part, based on conversations 13 that he was having with members of your team. 14 A. No, I don't recall anything of that sort. 15 Q. He told the Inquiry that when he received a request for 16 evidence, he would speak to a member of your team: 17 "... to get them to give a clear understanding so 18 I could make my judgement on that particular call." 19 So he was making a judgement on whether the 20 content of a call made by a subpostmaster or a mistress 21 could or could not explain the shortfall for which the 22 subpostmaster was being prosecuted. Do you understand? 23 A. I understand. 24 Q. He called this his due diligence exercise, that he was 25 speaking to members of your team to help to get their 57 1 help in explaining what calls meant and whether or not 2 the content of the call could explain away the shortfall 3 on which the subpostmaster was being prosecuted. 4 Understand? 5 A. I understand. I understand why Andy would have come to 6 members of the SSC for technical advice on a call and 7 what the Riposte messages meant. I don't recall ever 8 being aware that that was going to be used in any form 9 of litigation. 10 Q. Why would you known that he would be coming to members 11 of your team to ask for an explanation of what the 12 content of calls meant? 13 A. Because they were the technical expert on the contents 14 of the calls. 15 Q. Do you know why they weren't being asked to provide 16 evidence on the basis of the technical expertise that 17 they had of what had happened, rather than Mr Dunks who 18 performed a different function, Crypto Key Manager, 19 being asked to provide witness statements on the basis 20 of unrecorded and undocumented conversations with 21 members of your staff? 22 A. No. As I said earlier on, during my time as SSC manager 23 I was only aware of the one case. 24 Q. So this was going on below the surface without you ever 25 knowing about it? 58 1 A. I cannot recall ever knowing about it and I'm not 2 certain that the SSC staff members would have been aware 3 of why they were being asked about the calls. We were 4 completely open with anybody about what is the impact of 5 this PEAK, what's happening with it. So I'm -- no, I'm 6 in the dark as to much of this process. 7 Q. By that last answer, are you suggesting that Mr Dunks 8 may not have disclosed to members of your staff the 9 purpose of his call or the use to which the information 10 that he may be given might be put? 11 A. I am not certain that, whatever was being -- involved in 12 the discussion between Andy and SSC staff, that I was 13 ever aware of the use. I don't wish to ascribe 14 responsibility to that to Andy Dunks not telling SSC 15 staff or SSC staff not telling me. I just don't think 16 the subject came up. 17 Q. If you had been aware that Mr Dunks was conducting what 18 he described as a due diligence exercise, in deciding 19 whether or not the call or calls and the content of the 20 call or calls to the SSC could possibly explain away the 21 shortfall for which a subpostmaster was being 22 prosecuted, presumably you would have looked askance at 23 that? 24 A. I don't know is the honest answer to that. That's me 25 trying to predict emotions from a long time ago. 59 1 Q. Would you have been happy with your staff providing 2 evidence informally in this way that was being used to 3 prosecute subpostmasters? 4 A. I don't think that I would have been happy about it but 5 I can't be certain. 6 Q. Why do you think you probably would have been unhappy? 7 A. Because my understanding, limited as it was, of any form 8 of litigation process was that all of the data had to 9 come from the audit servers and that is specifically why 10 the SSC were never to touch the audit servers, so that 11 it was completely untouched by those people who had 12 write access to the parts of the system. 13 If I would have known that evidence was being 14 gathered from elsewhere, then I think in my mind that 15 would have put in question the origin of the data being 16 used in a case. 17 Q. Thank you. That document can come down from the screen 18 now. 19 You've mentioned the Lee Castleton case being your 20 sole experience of data from the Horizon System being 21 used in legal proceedings involving a subpostmaster. 22 A. Yes. 23 Q. Can we turn to paragraph 47 of your witness statement, 24 please -- sorry, page 47, and look at paragraph 147. In 25 paragraph 147, under "Conduct of Prosecutions", you say: 60 1 "I was not involved in the case of POL v Lee 2 Castleton, and I did not know of this case before 3 receiving the Request." 4 Can I just understand what you meant by that 5 sentence there, because the "Request", capital "R" -- 6 I am not going to take you right back to it but right at 7 the beginning of the statement you define "Request" as 8 meaning the Rule 9 request that we sent you in January 9 this year. 10 A. That was -- 11 Q. That can't be right, can it? 12 A. When I wrote my witness statement, I was asked 13 a specific question: was I involved in the case of 14 POL v Lee Castleton? 15 Q. Yes. 16 A. And I said no because, at the time, I did not know that 17 that was the case in which Anne Chambers was involved. 18 It's a question of terminology. I didn't know that that 19 was the name of the case. All I knew was that Anne 20 Chambers had had to go to court for a prosecution. Does 21 that ... 22 Q. So, essentially, what you mean by paragraph 147 is two 23 things, "I was not involved in the case which Anne 24 Chambers was involved in, which I now know to be called 25 Post Office v Lee Castleton", full stop? 61 1 A. That's correct. 2 Q. Secondly, "I did not know that the case in where Anne 3 Chambers was involved was called POL v Lee Castleton"? 4 A. Correct. 5 Q. Understood. 6 If we look over the page, please, you set out from 7 paragraph 153 down to 156 your involvement in the case 8 that you now know to be the Castleton case, yes? 9 A. Yes. 10 Q. In paragraph 153, you say: 11 "In this particular case, the person at Fujitsu 12 who was originally responsible/going to give evidence at 13 court declined to go. I cannot recall who this person 14 was or why they declined. My recollection is that Brian 15 Pinder was the Customer Service manager of the security 16 team at the time, and I believe it would have been his 17 responsibility to perform this task within his team." 18 So the way you're describing it there was that 19 there was originally a person within Fujitsu who was 20 going to give evidence at court and they declined. 21 A. That was my impression at the time, yes. 22 Q. Can you recall why they, that person, were originally 23 selected to give evidence? 24 A. No. 25 Q. Can you help us with why they declined to give evidence? 62 1 A. No. I don't know that I was ever told. As far as my 2 recollection, Anne was, to my belief, pressured to go to 3 court. I believed that that was a function of the 4 Security team. 5 Q. Sorry, just stopping there, the function of pressurising 6 her was the function of the Security team or the 7 function that she stood in for was their function? 8 A. The function that she stood in for. I believed that she 9 was being pressured to go to court because the person in 10 the security team was not going to go. 11 Q. Can you help us with -- I'm going to press you on 12 this -- why that person declined to go to court? 13 A. No, I don't know and I'm not sure that I was ever told. 14 Q. Who told you that they had declined to go to court? 15 A. I think that that came out in an argument and I was 16 having the argument with one of three people but I don't 17 recall which one. Specifically, I think Dave Baldwin 18 was the CS director at the time, Naomi Elliot, I believe 19 to have been the Support Services Manager, and Brian 20 Pinder was the head of the Security team. 21 Q. So you had an argument with one of those three people? 22 A. Yes. 23 Q. Where was the argument? 24 A. Probably in a corridor. 25 Q. Did they, the Security team, work in the same building 63 1 as you? 2 A. They did. 3 Q. On the same floor? 4 A. No, they were, I think, 5th floor. SSC were 6th floor. 5 Q. And it was in the course of that argument that you 6 learnt that the person who was originally responsible 7 had declined to go to court? 8 A. That was certainly the impression I got. I don't know 9 if it was specified in those terms. I can't -- 10 I obviously can't remember which one of the three people 11 I was having an argument with, so I certainly can't 12 remember the exact form of words that were spoken. 13 Q. In the third sentence there -- so, second sentence you 14 say you can't recall who this person was or why they 15 declined. That's to go to court? 16 A. Correct. 17 Q. The third sentence, you say: 18 "My recollection is that Brian Pinder ... it would 19 have been his responsibility to perform this task within 20 his team." 21 By that, are you saying that it ordinarily would 22 be Brian Pinder's job to go to court to perform this 23 task? 24 A. No, I am saying that Brian Pinder managed the team 25 within which I believed this task should have been done. 64 1 Q. You say there "I believe it would have been his 2 responsibility to perform this task". 3 A. Yes. 4 Q. You're only talking about going to court in that 5 paragraph. 6 A. Yes. 7 Q. Is that section of the statement incorrect then? That 8 gives the impression, does it not, that your belief was 9 that it was Brian Pinder's responsibility ordinarily to 10 perform the task of going to court? 11 A. No. Perhaps it would be clearer if you read the last 12 part as being "I believe it would have been his 13 responsibility to perform this task from within his 14 team". 15 Q. Or "I believe it would have been the responsibility of 16 a person within his team"? 17 A. Correct. 18 Q. Rather than it would have been his responsibility? 19 A. Yes. 20 Q. You weren't intending to say, by this paragraph, that it 21 was Brian Pinder's job to go to court and he had 22 declined to do so? 23 A. No, I wasn't intending to say that. 24 Q. What was the argument about then? 25 A. The principle of sending an SSC person to court or 65 1 producing a witness statement. 2 Q. Why was that a principle that you were fighting for or 3 against, the idea that somebody should go to court? 4 A. A number of reasons. Firstly, nobody in the SSC was 5 trained to do presentations, certainly not trained in 6 court etiquette or court procedures. 7 Secondly, it's an open-ended commitment for 8 somebody to go to court, which means that I was going to 9 lose one of my most skilled diagnosticians for 10 an unspecified period of time. 11 Thirdly, on a purely personal level, she was 12 clearly being very stressed by it. I wanted to make 13 sure that that did not happen to any of my staff in the 14 future. 15 Q. You say in paragraph 154: 16 "I was instructed by the Director of Customer 17 Services ... whose name I cannot recall, to detail 18 someone from the SSC to go to court to explain the 19 workings of the message store. I strongly objected that 20 nobody in the SSC had any experience of courts, or was 21 legally trained. I was overruled." 22 I think you just named the Director of Customer 23 Services at that time as Dave Baldwin; is that right? 24 A. My recollection -- and my timescales may be off -- at 25 one time Dave Baldwin was Director of Customer Service 66 1 and Naomi Elliot reported to him. At a different time, 2 Naomi was herself the Director of Customer Services. 3 I can't be precise about the timescales because I don't 4 remember. 5 Q. You don't know whether that Director of Customer 6 Services was at the time Mr Baldwin or Ms Elliot? 7 A. That's correct. 8 Q. You tell us in paragraph 155, if we just scroll down, 9 that, essentially, it was up to you to choose somebody 10 from the SSC to give evidence in the case against 11 Mr Castleton. You had a free choice? 12 A. That is my recollection. 13 Q. Was the choice not informed or dictated by the fact that 14 Anne Chambers was the one who had dealt with the 15 relevant PEAK arising from Mr Castleton's calls? 16 A. I almost certainly considered that, yes. 17 Q. Ie you picked the person who knew about the call that 18 was going to be relevant in evidence? 19 A. I'm fairly certain that that would have been one of the 20 criteria that I used to pick her, yes. 21 Q. In this paragraph, you say you picked her because she 22 was the most experienced and technically best in the 23 area of counter code. You had confidence in her honesty 24 and integrity -- 25 A. Yes. 67 1 Q. -- and she wouldn't be rattled? 2 A. Yes. 3 Q. Rather than "I picked her because she was the one that 4 knew about the call"? 5 A. I think that I had forgotten that she was involved in 6 the original call until reading some of the more recent 7 documents that the Inquiry have sent to me. 8 Q. In your discussion with the Director of Customer 9 Services, was there any discussion about whether the 10 witness would be giving evidence as an expert witness or 11 as a witness of fact of what had gone on in the call? 12 A. I don't recall the conversation, sorry. 13 Q. Do you understand the distinction that I've just made? 14 A. I believe so. 15 Q. What do you understand the distinction to be? 16 A. Sorry, can you go through the terms again? 17 Q. Yes. I asked whether there was a discussion over 18 whether the person giving evidence would give evidence 19 as an expert witness or a witness of fact of what had 20 gone on in the call. 21 A. In that case, the correct answer is, no, I don't 22 understand the difference between those. 23 Q. Was there any discussion between you and Anne Chambers, 24 therefore, over the basis on which she was going to give 25 evidence, what she was going to give evidence about, the 68 1 limitations of it? 2 A. My understanding was that she was going to give evidence 3 on the factual basis of the Riposte message store. 4 Q. Were you told, as part of the Director of Customer 5 Services' persuasion or overruling you, that the Post 6 Office was treating the Castleton case as something of 7 a test case and was going to use it, if it won, to try 8 and discourage other postmasters from either bringing 9 cases against the Post Office or defending them? 10 A. I don't recall that being in any way part of the 11 discussion and I don't think that I knew or became aware 12 of those implications until I received documents from 13 this Inquiry. 14 Q. So you weren't aware that, for the Post Office, it was 15 judged that to be the case that a lot was riding on 16 this? 17 A. No. 18 Q. Can we look at some documents, please. Firstly, 19 POL00099397. Thank you. 20 This is an email exchange in 2013, so many, many 21 years later, after you had left Fujitsu, and it's 22 an exchange in which you are not involved, therefore, 23 but there's something in it that I want to ask you 24 about. 25 Can we look at the bottom of page 1 and the top of 69 1 page 2, please. You can see an email from Mr Parker to 2 Mr Winn, yes? 3 A. Yes. 4 Q. Then if we scroll down, please, in the third paragraph, 5 it's the second paragraph on this page, Mr Parker says: 6 "The litigation bit [that's referring to 7 an earlier exchange in a chain that I'm not going to 8 take you to] is all to do with chain of evidence for 9 prosecutions and delivery in court. I'm sensitive about 10 it because in the distant past one of my team was 11 'persuaded' (by our side not yours) [that means by 12 Fujitsu, not the Post Office, in context] to write 13 an evidence statement without fully understanding the 14 implications. As you know, our 'professional witness' 15 for these types of cases is Gareth Jenkins but in this 16 case, because process was not followed, Gareth couldn't 17 do it and preparation for court became very difficult." 18 Firstly, do you understand what the process to 19 which Mr Parker is referring there ought to have been 20 where he says "process was not followed"? 21 A. Not really. If process wasn't followed -- since the 22 Castleton case was the first one that I had come across, 23 I'm not sure I would have known what the process being 24 followed by the Security team was. 25 Q. And, therefore, you wouldn't know in what respect it 70 1 hadn't been followed? 2 A. Correct, except clearly one of my staff was going to end 3 up in court when I did not believe that was appropriate. 4 Q. It says, Mr Parker's email, that, because the process 5 was not followed, Mr Jenkins couldn't give evidence. 6 Was that said to you back in 2006? 7 A. Not that I can recall. 8 Q. Can you think why a process not being followed meant 9 that Mr Jenkins could not give the evidence? 10 A. No, I don't think that I was even aware at that time 11 that Gareth was the nominated person in 2006 to give 12 evidence. 13 Q. What was your knowledge of Mr Parker's involvement in 14 these events? 15 A. Steve Parker would have been involved as the SSC 16 manager -- 17 Q. Back in 2006 I'm talking about. I've asked Mr Parker 18 about this already and he said "You'd better ask Mik 19 about it"? 20 A. I think Steve was involved because in December 2006 21 I was on honeymoon, so he was in charge of the SSC. 22 Q. In your statement, you've told us that it was you that 23 had the conversation with the Customer Services 24 Director -- 25 A. Yes. 71 1 Q. -- and it was you that persuaded Ms Chambers to give 2 evidence? 3 A. Yes. 4 Q. So those things didn't happen whilst you were on 5 honeymoon? 6 A. No. 7 Q. So what did Mr Parker do, then, outside of the 8 conversation that you had with Customer Services and 9 outside of the conversation persuading Anne Chambers to 10 give evidence when he was deputising for you? 11 A. I'm sorry, perhaps I'm being dense. I'm not 12 understanding the question. 13 Q. I'm trying to work out what Mr Parker's involvement was, 14 what his footprint was on this issue back in 2006. Did 15 he have any involvement in it at all, to your knowledge? 16 A. Only if I handed over what had been going own at the 17 time that I was going to be away from the office. 18 Q. Can you recall whether now you had handed over this 19 issue to him? 20 A. No, I can't recall if it was an extant issue when 21 I going on -- I would have briefed him along with all 22 the other things that were going on in the SSC before 23 I went on leave, but I cannot recall this specifically 24 being mentioned. 25 Q. Can we look, please, at FUJ00152300. I'm about to show 72 1 you a couple of documents now that we very recently 2 received from Fujitsu, over the weekend, I think. If we 3 look at the foot of the page, please, an email exchange 4 between you, Brian Pinder and Naomi Elliot, copied to 5 Anne Chambers of 29 January 2007. 6 A. Right. 7 Q. So just to locate that in time, this is after Anne 8 Chambers had given evidence -- 9 A. Yes. 10 Q. -- but I think before judgment. You say: 11 "Brian, 12 "I understand from Anne that you do not intend to 13 have an internal review on the Castleton case." 14 Why would there need to be an internal review on 15 the Castleton case? 16 A. Because Anne was concerned about the process. 17 Q. So it wasn't a review of the case as a whole, to your 18 mind; it was a process by which Mrs Chambers had come to 19 give evidence? 20 A. And her concerns with the evidence that she'd given. 21 Q. "Nevertheless, we are concerned that POA made some 22 errors during the course of this case which could prove 23 critical in any future litigation." 24 The reference to POA there, is that a reference to 25 the Post Office Account within Fujitsu, not a reference 73 1 to the Post Office? 2 A. I would read it as the Post Office Account within 3 Fujitsu. 4 Q. So that should read, essentially, "We are concerned that 5 part of Fujitsu made some errors during the course of 6 this case"? 7 A. That's how I would read that, yes. 8 Q. What errors did you think that part of Fujitsu had made 9 in the course of the Castleton case? 10 A. I don't recall. I only saw this document just before we 11 came in this morning. I believe that there is also 12 another document in which Anne makes her concerns clear. 13 Q. You refer to that in your next paragraph. You say: 14 "... Anne has written up her thoughts and comments 15 [which are attached], and I would welcome your 16 comments." 17 The subject line of this being a "'Mop up' on the 18 Castleton case". 19 A. Yes. 20 Q. Let's look at the paper that was attached to your email 21 then. 22 A. Okay. 23 Q. FUJ00152299. Thank you very much. 24 You will see that this is the paper that was 25 attached to that email. 74 1 A. Yes. 2 Q. It was written by Mrs Chambers, if you just scan, 3 please, both pages of the document at the same time or 4 put both up at the same time, you can see its length, 5 date and authorship. 6 So you can see it's got four headings. It's 7 signed off by Anne Chambers on 29 January 2007. That 8 was the date of your email sent at 11.34 that morning. 9 A. Yes. 10 Q. It's headed "Afterthoughts on the Castleton case". If 11 we can just go through this newly disclosed document, 12 please, starting on paragraph 1 or heading 1 at the top, 13 "Approach to SSC staff". Maybe if that can be blown up 14 for those that are looking online. She says: 15 "In the summer of 2006 I was asked directly by the 16 Security Manager whether I would be prepared to speak to 17 a solicitor about a call I had dealt with in February 18 2004. My initial response was that this was not the 19 normal process ... he reassured me that it was more or 20 less a formality so somewhat reluctantly I agreed." 21 You will see there that Anne Chambers has it down 22 more contemporaneously with events that it was she that 23 was asked directly by the Security Manager, whereas the 24 way you've described it is that it was you that asked 25 her somewhat reluctantly, following a request from 75 1 Security and a row. 2 A. I agree that's way it reads. 3 Q. Does reading this more contemporaneous document help you 4 to remember how matters, in fact, unfolded? 5 A. My memory is probably coloured by the fact that 6 I regarded it as my responsibility because she was one 7 of my staff; so I may well have remembered that 8 I persuaded her when she had, in fact, been contacted by 9 the Security Manager. Her memory is actually 10 considerably better than mine so I would defer to her. 11 Q. She says that she was asked directly by the security 12 manager. Who would that be? 13 A. I suspect that, at this time, it would have been Brian 14 Pinder. 15 Q. So she records being asked directly by the somebody who 16 you think would likely be Brian Pinder and the request 17 was to speak to a solicitor about a call that she dealt 18 with back in 2004. Mr Pinder "reassured me it was more 19 or less a formality", so she reluctantly agreed, and 20 then she said: 21 "Subsequently, before the meeting with the 22 solicitor, he asked me what my availability was in the 23 autumn for the court case. This was the first time 24 there was any mention of the possibility of having to go 25 to court. Repeated assurances that this would all be 76 1 settled before getting to court proved to be unfounded. 2 "I appreciate that there may be circumstances 3 where witnesses are summoned and have no option but to 4 comply, but I was not at all happy about how this was 5 handled." 6 Does this jog a memory in you in one of the 7 elements of unhappiness, the reassurance you're not 8 going to be required but, in the event, having to go to 9 court? 10 A. I cannot recall precisely but I would suspect that this 11 reaction from Anne formed the basis of the argument 12 which I had with one of the three people that I had the 13 argument with. 14 Q. Paragraph 2 or section 2, please, "Review of technical 15 evidence": 16 "When I took the initial call in February 2004, 17 I spent only a few hours on it before deciding that 18 I could not see any sign of a system problem. I only 19 looked at a couple of weeks' information. 20 "While in this case I am now sure that I did not 21 miss anything and my initial analysis was correct, I am 22 concerned that there was no technical review of the 23 Horizon [System] between the original call and the case 24 going to court. It is probable that any system problem 25 affecting the accounts would have shown up to Post 77 1 Office staff who did check the all the figures very 2 carefully, but since the subpostmaster was blaming the 3 system for the losses I think it would have been 4 sensible to have double checked this within Fujitsu 5 before it got as far as court. I was certainly 6 concerned, in the early stages, that there might be 7 something I had missed." 8 Just dealing with that paragraph first, were you 9 aware that between a call coming in about a shortfall or 10 a discrepancy and any action being taken against the 11 subpostmaster, there was no what she described as 12 technical review of Horizon evidence conducted by SSC 13 staff or by anyone? 14 A. That's difficult to say because this was the only case 15 that I knew of. 16 Q. What did you think when you read this? 17 A. I cannot recall reading this until I was presented with 18 it this morning. 19 Q. Would you, reading it now, recognise this as being 20 a serious issue, serious because one is moving from 21 an examination of a fault or error in the context of 22 a Helpdesk call, essentially -- 23 A. Yes. 24 Q. -- and then jumping into a prosecution or civil 25 proceedings without any intervening review? 78 1 A. I would agree. Yes, I would be concerned. 2 Q. Albeit Mrs Chambers saying, "In the event, I don't think 3 in this case it was a problem because my view is that 4 I didn't miss anything"? 5 A. I understand. I agree with her position in both 6 respects. If she said that she's reviewed it since and 7 she didn't miss anything, then I have absolute 8 confidence that that is the case. Her concerns about 9 the process I can only agree with. 10 Q. It's one thing answering a Helpdesk call amongst a stack 11 of tickets. 12 A. Yes. 13 Q. It's quite another thing being used as a witness to 14 speak, as she does in her next paragraph, being treated 15 as an expert witness in answering a wide variety of 16 calls about the system? 17 A. Yes, I agree. 18 Q. You see that she says: 19 "... I found myself being treated as an expert 20 witness and answering a wide variety of questions about 21 the system, although nominally I was a witness of fact 22 and my witness statement covered just the investigation 23 done in 2004. Fortunately I do have extensive knowledge 24 of the system and was able to fulfil the wider role -- 25 but what would have happened if the initial call had 79 1 been handled by a less experienced SSC person? 2 "If there is a similar case in the future, where 3 the system is being blamed, would it not be sensible to 4 have a technical review of all the evidence, at the 5 first indication that a case may be going to court? 6 Someone involved in that review would then be well 7 placed to give evidence in court." 8 Just dealing with those questions that she 9 asked -- in a future case where somebody is blaming the 10 system, wouldn't it be sensible to have a technical 11 review of all of the evidence -- what was done as 12 a result of that question being asked? 13 A. I don't know. 14 Q. We're 2006 here, end of 2006/beginning of 2007 and there 15 are many prosecutions that follow this. Can you help us 16 what happened to the, on the face of it, not 17 unreasonable question that Mrs Chambers is raising? 18 A. I agree that the question is entirely understandable. 19 I know that, at some point, Gareth Jenkins took over the 20 responsibility for doing technical presentations. My 21 recollection is that, even between 2006 and the time 22 that I left, I was not aware of any other prosecutions. 23 Q. Just going back to the email that you got in response to 24 this -- we'll come back to the document in a moment -- 25 that's FUJ00152300, look at the response from Mr Pinder 80 1 at the top, back to the same distribution list, ie you, 2 Mrs Chambers and Ms Elliot: 3 "Mik Anne 4 "Thanks Mik, there was no intention to have a wash 5 up on this particular case as such but I must stress 6 that from the outset this was 'new ground' and 7 a particularly unusual case (1st of its kind in 8 10 years) for all concerned. It involved many ... 9 variables which, at any point in time could have 10 culminated in a totally different outcome. 11 "This enquiry took well over a year to conclude 12 and routine procedures which have served us well for 10 13 years were suddenly being stretched to new limits, but 14 it does highlight how [the Post Office Account] can be 15 called to account and I totally agree we must learn from 16 this. 17 "Anne (many thanks for your comments) you have 18 highlighted some interesting areas of procedure which we 19 need to recognise, and I will discuss these with Naomi 20 and will keep you both informed." 21 On one view, that reads as sort of a pat on the 22 head where nothing much is going to happen or am I being 23 unfair? 24 A. No, I would agree with you. 25 Q. Did anything happen? 81 1 A. I don't recall this email -- 2 Q. What about the substance of the issue then? Do you 3 remember anything happening? He says he's going to keep 4 you both informed. 5 A. I don't have any recollection that I was kept informed 6 of any progress on those issues. 7 Q. Can we go back then to Mrs Chambers' document 8 FUJ00152299 and look at the foot of the page "Disclosure 9 of evidence": 10 "Fujitsu made a major legal blunder by not 11 disclosing all the relevant evidence that was in 12 existence. I found myself in the invidious position of 13 being aware that some information (Tivoli event logs) 14 existed, but not sure whether they had been disclosed or 15 not, since I had not been party to any of the requests 16 for disclosure. It became evident in court that they 17 had not been disclosed. 18 "Quoting from an email received from POL's 19 solicitor after my revelation ..." 20 Then there's a quote from an email. For the Core 21 Participants who are aficionados in this area, if they 22 want to look at that email it's POL00070104. Anyway, it 23 reads: 24 "'In any litigation, the parties involved have 25 a continuing obligation pursuant to the Court rules to 82 1 disclose all documents that may help or hinder their 2 case or the other side's case. In this context, 3 a "document" means anything in which information of any 4 description is recorded, so it includes, just for 5 example, a computer database. Previously, I had asked 6 Fujitsu to let me have all the info it had and had been 7 helpfully given HSH call logs, transaction logs and 8 event logs. I was also recently told that there was 9 a message store which had everything else on it and we 10 invited Mr Castleton to look at this, but he didn't take 11 up the opportunity.'" 12 She continues: 13 "This suggests that disclosure of the message 14 store itself was an afterthought, though it is 15 fundamental to the system. I know that for fraud cases 16 the 'transaction log' and 'event log' are extracted from 17 the ... message store and submitted, but surely the full 18 message store has to be disclosed in all cases? 19 "Many other files are also archived to the audit 20 servers as a matter of course and could hold relevant 21 information, although the Security team are not 22 necessarily aware of their existence or potential 23 relevance. I'd like to suggest that a list of these 24 files is compiled so that similar mistakes are not made 25 in the future. 83 1 "And what about calls on PEAK, which may have 2 evidence attached? And any evidence which might have 3 been kept within SSC? I was not asked whether I had 4 anything that might have been relevant (as it happens, 5 in this case I did not). 6 "Of course there may be subtleties to this that 7 I am unaware of, whereby data may exist but there is no 8 obligation to disclose it. If this is the case, could 9 any future witnesses be briefed appropriately? The 10 response 'no-one has ever asked for that before' does 11 not seem to be a good reason for non-disclosure." 12 Would you agree that Mrs Chambers is raising there 13 a series of reasonable, focused and pertinent 14 questions -- 15 A. Yes. 16 Q. -- against the context of her saying that Fujitsu, the 17 company that you work for, had made a major legal 18 blunder by non-disclosure of evidence? 19 A. I'm not sure that I would have categorised it in that 20 way but the implications of what she is saying I would 21 certainly agree with. 22 Q. Why wouldn't you categorise the non-disclosure as 23 a major legal blunder? 24 A. Because I'm not sure what constitutes a major legal 25 blunder. I agree that if these documents were not 84 1 presented at litigation, then that was an error. 2 Q. The series of questions that she asks (for example, 3 surely the message store has got to be disclosed on all 4 cases), did that happen thereafter? 5 A. This was the only court case that I knew about, so 6 I have no idea. 7 Q. What did you -- she refers there in the second paragraph 8 down: 9 "I know for fraud cases the 'transaction log' and 10 'event log' are extracted from the message stores but 11 surely the full message store has to be disclosed." 12 That's telling you that there's another species of 13 case, isn't it? 14 A. It is but I don't recall seeing this document until this 15 morning and I don't recall being aware that there were 16 any other cases. 17 Q. So you weren't aware that members of your team were 18 speaking to Mr Dunks to help him compile witness 19 statements to prosecute people? 20 A. That's correct. 21 Q. You weren't aware then of your team's indirect 22 involvement in the prosecution of subpostmasters? 23 A. Not that I can recall. 24 Q. You got this document. It was sent to you by Anne 25 Chambers and indeed you sent it on to Customer Services, 85 1 Customer Support, to Mr Pinder -- 2 A. Yes. 3 Q. -- and Ms Elliot? 4 A. Yes. 5 Q. Can you recall what happened as a result of these 6 reasonable, focused and pertinent questions being asked? 7 A. No, I cannot. 8 Q. As a manager, you would want to grip a situation like 9 this, wouldn't you? 10 A. I agree. 11 Q. Can you help us what you did do to grip it then? 12 A. No. I'm sorry, I've no recollection of this. 13 Q. Can we move down to paragraph 4, please: 14 "This case highlighted a common problem, both in 15 2004 and now. The postmaster raised many calls about 16 his continuing losses, both with Horizon and with the 17 NBSC. These kept being bounced [back] and it took weeks 18 before a call was passed to SSC." 19 Is that accurately describing a common problem; 20 namely a subpostmaster continually raising calls about 21 continuing losses which were bounced back or were being 22 bounced at the lower levels of Customer Support? 23 A. It's not something that I recognise. As we discussed 24 earlier, the SSC was supposed to receive the first 25 instance of a new incident. Subsequent incidents would 86 1 become a problem, rather than an incident, and would be 2 handled through the problem management process. So if 3 there were multiple calls on a single problem, then the 4 SSC would receive the first. Subsequent ones would be 5 referred to the problem managers by HSH and SMC. 6 Whether or not calls were being bounced around 7 between Horizon and the NBSC, I wouldn't know. 8 Q. This is telling you that it is a problem and it's 9 a common problem. 10 A. I agree, in which case Anne must have actually 11 researched that to find that. 12 Q. Again, that needs something to be done about it, doesn't 13 it? 14 A. I was not in the -- when I was in the SSC, I was not 15 involved with any relationship between the Horizon 16 Helpdesk and the NBSC, so I can't comment on that. 17 Q. But your staff were. 18 A. In the normal course of events, no, they weren't. 19 Q. How was Mrs Chambers able to write this then? She's got 20 some knowledge. 21 A. She researched -- presumably researched it having 22 recognised that there was an issue. 23 Q. And that it's a common one? 24 A. She may have deduced that from the evidence that she'd 25 got having researched it. 87 1 Q. That's what she's saying: it's a common problem? 2 A. I agree that's what she's saying. 3 Q. So what would you want to do, receiving this? 4 A. I would have wanted to know how big a problem this was. 5 Q. It's a common one. 6 A. That's not very specific. How I would deal with it now, 7 I don't know. How I dealt with it at the time that Anne 8 brought this, I don't remember. 9 Q. Mrs Chambers continues: 10 "Strictly speaking, problems with discrepancies do 11 need to be investigated by NBSC in the first instance, 12 but where there are continuing unresolved problems it 13 should be possible to get the issue investigated 14 properly, and one of the helpdesks should be prepared to 15 take responsibility for the incident. Personally 16 I think the fact that the Horizon Helpdesk is penalised 17 for passing 'Advice and Guidance' type calls on to third 18 line leads to too many calls being closed without proper 19 investigation or resolution. This is very frustrating 20 for postmasters, though possibly not an issue of concern 21 to POL." 22 She's raising a systemic issue there, isn't she? 23 A. She's raising what she believes to be an issue in 24 process, which I believe was handled by the problem 25 management process, not the incident management process. 88 1 Q. So she was somebody who was one of your most experienced 2 and trusted diagnosticians, wasn't she? 3 A. Agreed. 4 Q. You had absolute faith in her competence and integrity? 5 A. Yes. 6 Q. So what did you do with this problem that had been 7 raised? 8 A. I probably discussed it with her and went through the 9 problem management process. 10 Q. What does "went through problem management processes" 11 mean? 12 A. Pointed out to her that the problem management process 13 is the place where repeat issues should have been 14 handled. 15 Q. Is that another form of a pat on the head? 16 A. No, it's me disagreeing with what she said. 17 MR BEER: Sir, I wonder whether we could take the lunch 18 break earlier today at 12.45 and come back at 1.45, 19 please. 20 SIR WYN WILLIAMS: Yes, by all means. 21 MR BEER: Thank you, sir. 22 SIR WYN WILLIAMS: There is just one question, I think, 23 while it is in my mind, Mr Peach, and it relates to 24 questions you were asked about 15 minutes ago now when 25 you were looking at the events which led to Mrs Chambers 89 1 actually giving evidence in the Castleton case, and then 2 shortly after you answered questions about that, Mr Beer 3 took you to an email much later, 2013, when Mr Parker 4 had suggested, or at least it could be thought that 5 Mr Parker was suggesting, that Gareth Jenkins was due to 6 give evidence but there had been a process failure which 7 meant that that couldn't happen. 8 Now, the question I want to ask you is simply 9 this: from your memory when you were speaking either 10 to -- well, when you were speaking to anyone about 11 whether or not Mrs Chambers should give evidence, was 12 the name "Gareth Jenkins" mentioned to you as someone 13 who normally would give such evidence but couldn't on 14 this occasion? 15 A. No, his -- to my recollection, no, his name was not 16 mentioned and the impression that I've got from the 17 email from Brian Pinder was that this was the first 18 occurrence of this sort of event and, therefore, I would 19 not have expected there to have been a process which 20 involved Gareth Jenkins. 21 SIR WYN WILLIAMS: And I ask the question because, as 22 Mr Beer pointed out to you, Mr Parker couldn't really 23 explain how the name "Gareth Jenkins" came to be written 24 by him in this context in 2013 but suggested you might 25 know the answer. But it looks as if we have the 90 1 unsatisfactory position that neither of you can explain 2 that reference to Mr Jenkins. Is that really where we 3 are? 4 A. That's correct. 5 SIR WYN WILLIAMS: All right. Does that mean we start at 6 1.50 instead of 1.45, Mr Beer? 7 MR BEER: Yes, thank you, sir. 8 SIR WYN WILLIAMS: All right, fine. 9 (12.45 pm) 10 (Luncheon Adjournment) 11 (1.50 pm) 12 MR BEER: Good afternoon, sir. Can you see and hear me? 13 SIR WYN WILLIAMS: Yes, I can, thank you. 14 MR BEER: Thank you very much. I said I was going to move 15 to a new topic. I realised that I omitted to ask a few 16 questions on our last one that I ought to ask now. Can 17 we go back, please, to FUJ00152299. This was the Anne 18 Chambers document of 29 January 2007, the two-page 19 document; do you remember? 20 A. Yes. 21 Q. I've taken you through individual parts of it and asked 22 you questions about it. Can I take a step back and ask 23 you to take a step back and look at the document as 24 a whole. 25 A. Yes. 91 1 Q. Would you agree that the document raises a series of 2 fundamental issues about the way that evidence is 3 presented to a court by Fujitsu in proceedings against 4 subpostmasters? 5 A. I don't think I can draw that inference from one case. 6 I agree that what's in this document means that, in this 7 case, I would agree with Anne there were things that 8 were missed, but I couldn't extrapolate that into other 9 cases. 10 Q. Did you ask, "What do we do in any other cases"? 11 A. No, I did not. 12 Q. So this is your total sample size, yes? 13 A. I agree. 14 Q. Why wouldn't you ask, "Is this an outlier, do we do this 15 in all cases? Do we do other cases at all?" 16 A. I don't recall this document and it was only presented 17 to me this morning, so I haven't had the opportunity to 18 research whether or not perhaps I made comments in my 19 monthly report. I just don't know. 20 Q. So we should look, should we, to your monthly -- we got 21 this document on 12 May from Fujitsu -- 22 A. Right. 23 Q. -- which is why you got it recently. 24 A. Yes. 25 Q. If we are to see what action you took as a result of 92 1 this, where should we look? 2 A. I can infer, but not be certain, that having seen this 3 document at the time that could have been when Gareth 4 Jenkins started taking on the responsibility for 5 litigation. Until this morning, I had always thought of 6 Gareth as being a technical expert on one part of the 7 system. I had seen him in the office and I knew that 8 people deferred to him for expert advice. Whether or 9 not his taking over on the litigation side was a result 10 of actions that I took from seeing this from Anne, 11 I just cannot say. I don't remember. 12 Q. Just going back to my question then: where should we 13 look for evidence of action that you took? 14 A. Management meetings, discussions between the CS Director 15 and Brian Pinder. That's all I can suggest. 16 Q. Would you agree that this document, in general terms, 17 standing back, calls for action to be taken? 18 A. Yes, I would agree. 19 Q. Because it's sending a warning, isn't it? It's raising 20 red flags on a number of fronts to Fujitsu about the way 21 that evidence is presented? 22 A. Yes, it is. 23 Q. And I think you would probably agree that, in the light 24 of what we now know, this was a harbinger for many of 25 the events that we were subsequently to see, confusion 93 1 over whether a witness was giving evidence as a witness 2 of fact or an expert witness, non-disclosure of 3 documents? 4 A. I agree to that. As I said before, as far as I was 5 aware, this was the only case that I knew of. 6 Q. Why was Anne Chambers raising it with you if it was the 7 only case? Surely the memo itself contemplates that 8 there are more. 9 A. I agree to that too. It appears that the action that 10 I took was certainly to forward her concerns to the 11 Security Manager. What happened after that I have no 12 memory of. 13 Q. Thank you. That document can come down. 14 Do you remember a man called Alvin Finch working 15 in the SSC? 16 A. I remember interviewing and subsequently offering a job 17 to Alvin. I confess I had not remembered his surname 18 until I saw him on the witness list. 19 Q. Do you remember how long he worked in the SSC for? 20 A. I remember it being a very short time and that his 21 leaving was not happy. 22 Q. Why was his leaving not happy? 23 A. I went on leave. When I came back from leave, I was 24 informed that he'd thrown his pass across the floor and 25 announced to everybody that he was being bullied. 94 1 Q. He's informed the Inquiry in his witness statement that 2 he felt uncomfortable working at Fujitsu because he felt 3 that there was a culture of fear, that he was bullied by 4 a member of the staff and, therefore, walked out, and 5 that in the SSC there was a culture, which he says came 6 from the top, of prioritising short-term fixes over 7 long-term solutions. 8 Taking those allegations in turn, did you know 9 that he felt uncomfortable working within the SSC? 10 A. No, I don't recall any occasion when he raised that with 11 me. 12 Q. That he felt that there was a culture of fear within the 13 SSC? 14 A. I would dispute that. 15 Q. Did you know that he said he was being bullied by 16 a fellow member of staff? 17 A. I knew that because it was reported to me when I got 18 back from leave. 19 Q. And that there was, he says, a culture which came from 20 the top of prioritising short-term fixes over long-term 21 solutions? 22 A. Within the SSC, our role was to generate workarounds for 23 any incident if we could and then to pass the call to 24 development for a code fix. The code fix prioritisation 25 was done at the release management forum of which I was 95 1 a member. I can see that it may be considered that from 2 within the SSC prioritising a workaround in order to 3 keep the system up and running was important but the 4 generation of the code fix was equally important, but 5 probably not visible or as visible to someone of Alvin's 6 level. 7 Q. Can I turn to the KEL system then, please. Mrs Chambers 8 has told the Inquiry that an issue or concern with the 9 KEL system was that service tickets would be passed to 10 the SSC with the wrong KEL quoted on them. Were you 11 aware of that problem? 12 A. Yes, particularly in the early days and I had meetings 13 with the SMC manager when it happened, initially, 14 I recall, monthly, because the SMC's filtration rate was 15 low, later by exception, saying something along the 16 lines of "SMC missed this or got the wrong KEL, why? 17 What was wrong with it? Do we need to amend it? Do we 18 need to amend your procedure?" 19 Q. What was the result of that? 20 A. Frequently an update to the KEL. A lot depended on 21 where the KEL had been written and by whom because 22 technicians at third or indeed fourth line would specify 23 a problem in technical terms, whereas what the 24 postmaster was reporting was clearly something that was 25 happening on a counter, and there were certainly 96 1 occasions where it was not clear from the description of 2 the problem from the postmaster that a particular KEL 3 applied. 4 Q. Did the issue have an adverse impact on Fujitsu's 5 ability and POL's ability to respond timeously and 6 appropriately to complaints made by subpostmasters about 7 Horizon? 8 A. It would have had some impact purely because the call 9 was being passed to the SSC with an incorrect KEL. It 10 was then our responsibility to update the KEL so that 11 that didn't happen again. Did it make any difference to 12 the way in which a specific incident was diagnosed? No, 13 I don't think so. 14 Q. Mrs Chambers also told the Inquiry that the SSC and 15 fourth line support development did not always know how 16 many branches had reported a particular problem because 17 tickets never made their way up to the SSC or to the 18 fourth line, and also because the breadth of a problem 19 in unreported instances of the problem never made their 20 way to the SSC because subpostmasters didn't report 21 them. 22 Were you aware of those two problems? 23 A. When there were multiple instances of a particular 24 incident, then the process for handling that and 25 tracking the duplicates was part of the problem 97 1 management process and, therefore, may not have been 2 visible immediately to the SSC. However, there was 3 frequent contact between the problem managers who were 4 also the Service Delivery Managers responsible for each 5 area. So if an incident was happening multiple times, 6 then the problem manager would be dealing with it as 7 a problem and he would talk to me and say "N" post 8 offices have this problem. 9 Q. How would he find out what "N" equalled? 10 A. From the HSH and SMC. They had a direct line through to 11 the Service Delivery Managers. So duplicates and the 12 trend analysis were part of the problem management 13 process between the HSH and the problem manager. The 14 SSC was tasked specifically with dealing with incidents, 15 that being the first event of a new problem. 16 Q. How would a problem manager know or understand the 17 breadth of a problem by reference to access to HSH 18 databases? 19 A. They would be the ones that knew how many post offices 20 had called in with that problem. 21 Q. What if the post office didn't know what the problem 22 was? 23 A. Sorry, I don't -- 24 Q. They say, for example, "I've just balanced and there's 25 a discrepancy. That's all I know". 98 1 A. Okay. So the initial version of that would come to the 2 SSC for us to do the diagnosis and, if necessary, 3 organise a fix with development. Multiple post offices 4 phoning in saying they had the same problem would be 5 treated as a problem. 6 Q. So were they attributed a code by the HSH? 7 A. I have no knowledge of exactly the way that HSH and SMC 8 communicated that through to the problem managers. 9 I know that the problem managers kept a database. 10 Q. What was that database called? 11 A. I think it was just called the Problem Management 12 Database. 13 Q. What was it populated with? 14 A. I don't know. It was populated by the problem managers. 15 Q. So if you had to describe top to bottom -- or bottom to 16 top, bottom-up, the system that Fujitsu employed to 17 ensure that the nature, extent and breadth of a problem 18 was established, how would you describe it? 19 A. I would describe it as solely within the remit of the 20 Service Delivery Managers who would interface to their 21 counterparts in POL. 22 Q. How would the Service Delivery Managers know about these 23 pockets of instances around the country? 24 A. Because HSH would tell them. 25 Q. What would HSH tell them? 99 1 A. That, I don't know. 2 Q. Can we look, please, at FUJ00079939. You'll see this is 3 a 2005 document entitled "[Post Office Account] Customer 4 Service Incident Management Process Details". 5 A. Yes. 6 Q. The "Abstract" says that the document describes the 7 customer service incident management process and I think 8 we'll see that you're one of the people to whom it was 9 distributed. 10 A. Yes. 11 Q. Can we look at page 9, please. The document says: 12 "The inputs to this process are: 13 "All Incidents reported by Contact with the [Post 14 Office Account] Horizon Service Desk. Contact is 15 defined as voice or Tivoli alert as the methods of 16 communication with the Horizon Service Desk and fall 17 into the following categories: 18 "Business process error. 19 "Hardware or software error. 20 "Request for information, eg progress of 21 a previously reported Incident. 22 "Network Error. 23 "Severity and SLT information. 24 "Evidence of an Error. 25 "System Alerts received automatically from OMDB." 100 1 So is this recording the fact that problems might 2 be identified through a number of routes via a Tivoli 3 alert, via a system alert or via contact with the 4 Helpdesk itself? 5 A. I would agree with everything in that statement except 6 of "problem" instead of "incident". 7 Q. So incident issues? 8 A. Incidents, yes. 9 Q. You don't like the word "issue"? 10 A. I got used to using the ITIL nomenclature for 11 identifying the difference between incidents and 12 problems, which was a struggle for me at times. So you 13 have to make a distinction on an incident between 14 a problem -- you have to make a distinction between 15 an issue and an incident and a problem. 16 Q. So the problem managers weren't really there to manage 17 problems, they were only there to manage incidents; is 18 that right? 19 A. No, it's exactly the reverse. The problem managers were 20 there to manage problems. They were not there to manage 21 incidents. 22 Q. What's an incident? 23 A. A single occurrence of an anomalous event somewhere in 24 the system. 25 Q. Anomalous in what sense? 101 1 A. Just something going wrong. Tivoli -- of all those 2 things in there. So Tivoli error messages being 3 reported by the SMC, incidents being reported by 4 postmasters, errors occurring on the central systems, 5 they would all be an incident. If there were multiple 6 instances of an incident, then HSH would forward that to 7 the problem managers. 8 Q. What was the remit of the problem manager then? 9 A. From personal experience, I don't know. There is 10 a document very similar to this one called the "Problem 11 Management Process" which would detail that. 12 Q. Where did the problem managers sit? Where were they 13 physically located? 14 A. Physically on the 5th floor in Bracknell. 15 Q. So is this identifying incidents that may arise either 16 because of automatic detection by a Fujitsu system or 17 those which are reliant on a subpostmaster calling them 18 in? 19 A. Both of those, yes. 20 Q. If we look further down the page under "Dependencies": 21 "The process is dependent on: 22 "Effective incident handling by the Service Desk. 23 "The Known Error Database being available and kept 24 up to date with all errors as the root cause becomes 25 known to Problem Management." 102 1 Just stopping there, what was the difference 2 between the Known Error Database and the Service Desk 3 Knowledge Database? 4 A. The Service Desk meaning HSH, in this case? 5 Q. Yes. 6 A. HSH used the KEL system as did the SMC. The HSH, 7 because, speaking from memory, 90 per cent of their 8 calls were hardware-related, they kept an additional 9 database which acted as their known error database for 10 hardware problems. 11 Q. It was restricted, was it, to hardware problems? 12 A. As far as I'm aware but the SSC never used it, so 13 I can't be certain. 14 Q. If we just go over the page, you'll see that it's 15 described, in the next bullet point: 16 "Service Desk knowledge database (HSH ONE) ..." 17 Is that what it was known as? 18 A. Yes. 19 Q. It says it's kept up to date with POL business and 20 services knowledge. You're saying, to your knowledge, 21 that was only in relation to hardware issues? 22 A. The only times that I ever came across it, it was 23 relating to hardware problems but the SSC didn't keep 24 it, access it or maintain it, so its exact structure 25 I couldn't tell you. 103 1 Q. Whose responsibility was it to ensure that the content 2 of the KEL database was consistent with the HSH one? 3 A. I don't recall. 4 Q. Was it anyone within the SSC's responsibility? 5 A. No. Because the SSC did not have access to HSH ONE. 6 Q. So would you think that it was the HSH's responsibility 7 then? 8 A. I really couldn't comment. I have no knowledge of the 9 contents of HSH ONE. 10 Q. Whose role was it to ensure that KELs were communicated 11 effectively to HSH members? 12 A. SSC. 13 Q. How was that done? 14 A. We made sure the database was there, we made sure that 15 HSH and SMC, which in this context we would treat as one 16 unit, made sure that they understood it, had meetings 17 with their manager to ensure that they were able to use 18 it. SSC made sure that the server that contained it was 19 up and running all the time. 20 Q. So maintaining the content and continuity of operation 21 of KEL? 22 A. Yes. 23 Q. Was there a system in place to see how consistently HSH 24 were using the KEL database? 25 A. Initially there were monthly reviews between myself and 104 1 the SMC manager, who was my interface. After the 2 filtration rates of the SMC rose sufficiently, then that 3 was done on an ad hoc basis. 4 Q. What do you mean by that? 5 A. I mean, if SSC staff said, "SMC have raised this KEL and 6 we don't agree with it", then I would go back to the SMC 7 manager and say, "Why is this, what's happening? Do we 8 need to update the KEL?" 9 Q. Can we go to page 15, please and look that foot of the 10 page at paragraph 2.5. The document records that: 11 "If the incident is not routine, the Service Desk 12 agent checks for Known Errors listed in HSH ONE and the 13 SSC KEL against records relating to the incident 14 symptoms. If a match is found, the agent informs the 15 caller of the workaround or resolution available and 16 links the call to the master incident record." 17 Would you agree that, quite aside from the 18 importance of the KEL, the content of the HSH ONE 19 knowledge database is key to how and whether a problem 20 was escalated or resolved properly? 21 A. Yes, and I would agree with that statement. 22 Q. If we go on over the page, please, to paragraph 2.7, the 23 document records that: 24 "If no match is made against the Problem Database 25 ..." 105 1 The "Problem Database", is that referring to 2 something other than the KEL and the HSH ONE? 3 A. That's how I would read it. 4 Q. Is that the thing that you were describing earlier? 5 A. Yes, I think so. 6 Q. "If no match is made against the Problem Database, the 7 Service Desk continues with first line resolution of the 8 Incident assisted by the Product Support Engineers 9 (PSEs)". 10 So that's if there's no match on HSH ONE, no match 11 on KEL and no match on problem database, the incidents 12 refer back down to first line support; is that right? 13 A. I'm not sure who Product Support Engineers are in this 14 term. 15 Q. Are they people that are concerned with hardware? 16 A. Reading the next paragraph I would read it as both, 17 since the SDUs are Service Delivery Units that would 18 include the hardware and the SSC. 19 Q. You have to explain that answer in a bit more detail. 20 Just sticking with 2.7 for the moment -- 21 A. Okay. 22 Q. -- if there's no match of the incident in HSH ONE or KEL 23 and no match on the problem database, the incident gets 24 sent back to first line resolution, yes? 25 A. The way I was reading this was that we were currently in 106 1 the first line when they were checking HSH ONE and the 2 SSC KEL. So the Product Support Engineers I would read, 3 as being part of the HSH. 4 Q. Okay. You think they are people within the HSH? 5 A. I would read that that way, yes. 6 Q. I have suggested that they are hardware specialists and 7 then you referred to paragraph 2.8. What in 8 paragraph 2.8 causes you to think that Product Support 9 Engineers may also be involved with software? 10 A. Because it refers to SDU which stands for Service 11 Delivery Unit and SSC is listed in many documents as one 12 of the Service Delivery Units. 13 Q. So if the other people within first line support can't 14 resolve the incident, it goes back to SSC; is that 15 right? 16 A. I don't think so. Can you reword that or ... 17 Q. Yes. You said that SDU, Service Delivery Unit, is 18 defined in a number of documents as meaning the SSC? 19 A. No, the SSC is one of the Service Delivery Units. 20 Hardware engineers would also be a Service Delivery 21 Unit, as would the Network team. 22 Q. Okay. So it's an umbrella term, SDU, referring to 23 a group of different units? 24 A. Yes. 25 Q. So this paragraph 2.8 means: if HSH doesn't contain the 107 1 answer, if KEL doesn't contain the answer, if the 2 Product Support Engineers can't find the answer, then 3 you go back to one of the units within SDU using the 4 support matrix in HSH ONE? 5 A. You don't go back to, you go -- in my terms, you go up 6 to. 7 Q. Up to. 8 A. If first line can't resolve the incident and there are 9 no known errors on any of the systems, then they will 10 pass it through to, effectively, the third line support 11 units. 12 Q. Why do you say, "effectively, the third line support 13 units"? What about the second line? 14 A. Well, because for -- as I can recall the process, HSH 15 would send calls which were clearly hardware to the 16 engineering team. They would send calls which were 17 clearly or possibly software issues to the SMC and, 18 whether or not the interface to the Network team went 19 through HSH direct or through the SMC, I can't recall. 20 Q. Can we move forwards please to page 18 and 21 paragraph 4.4. The document records: 22 "Where this incident has a number of calls 23 referenced to it, or where there is a probability that 24 proactive action is required to prevent further 25 occurrences of this Incident the IMT [I think that's 108 1 Incident Management Team] initiate a Problem record to 2 be authorised by the Service Management Team and passed 3 to Problem Management." 4 What do you understand that process to involve? 5 A. I understand that to be the process by which HSH, having 6 received a number of calls, would attach those calls to 7 the incident and then report to the problem managers for 8 handling as a problem. 9 Q. How many calls would be sufficient to justify doing so? 10 A. I believe that the problem management process says on 11 receipt of the second call. Whether or not that was 12 modified to say "Only do it with 'N' calls", I don't 13 know. 14 Q. Alternatively it says: 15 "... where there is a probability that proactive 16 action is required ..." 17 Who would decide whether it was probable that 18 proactive action would be required? 19 A. Reading that, I would assume HSH because they're the 20 ones that are having the conversation with the problem 21 managers. 22 Q. How would they decide the probability that proactive 23 action is required? 24 A. I don't recall if I ever knew the details of the 25 interaction between HSH and the problem managers. 109 1 Q. Thank you. Can we move on to the PEAK system, please. 2 At paragraph 42 of your witness statement, which is on 3 page 17, you say: 4 "If the SSC recognised that a particular problem 5 could have implications for multiple branches, this was 6 added to the PEAK and the KEL. It is important to note 7 that problems which occurred in overnight processing 8 sometimes had the potential to affect all Post Office 9 branches, but not every potentially affected branch 10 would be listed on the PEAK." 11 How would SSC recognise that a particular problem 12 could have implications for multiple branches? 13 A. In two ways. If HSH has received multiple calls, then 14 they had the ability to attach the call references to 15 the master call which was presumably by this time in the 16 SSC. If an SSC person, when analysing a problem, 17 realised that it might have wider implications, then 18 they would actually trawl the system to find out if 19 other post offices are being affected by it. 20 Q. How would they trawl the system? 21 A. I think that would vary widely depending on the nature 22 of the problem. If I can think of an example ... if we 23 take as an example, and it's hypothetical, an issue in 24 a certain type of reference data, then having received 25 the initial call, the SSC would go through the 110 1 correspondence server message stores in the central 2 systems and identify which other post offices could be 3 affected by this particular issue. 4 Q. And what would they do then? 5 A. Note the details of those post offices on the PEAK, 6 probably on the KEL as well and, if it was felt to be 7 a problem rather than a single incident, then we would 8 talk to the problem managers in that area. 9 Q. Whose responsibility was it to get in touch with the 10 Post Office to tell them that a problem had -- 11 A. Problem managers -- Service Delivery Managers who became 12 de facto problem managers for their areas had interfaces 13 through to Post Office management. 14 Q. Now, I think earlier in the process, before calls got up 15 to the SSC, you were aware, obviously, that, from what 16 we've discussed so far, that the HSH had access to their 17 own knowledge database and KELs. 18 A. Yes. 19 Q. Were you aware that they were provided with scripts? 20 A. I have become aware that they were provided with 21 scripts. I must have been aware at the time because it 22 was clear from some of the PEAKs coming to us the 23 responses from postmasters to questions were being 24 repeated. So I was aware they were running on scripts. 25 The exact origin of those scripts, I cannot remember. 111 1 Q. Did you ever see the scripts? 2 A. Not that I can recall as scripts. I can remember seeing 3 on PEAKs a series of questions and answers which were 4 clearly the same between some PEAKs. So a series of 5 questions, a series of different answers. So, yes, 6 I was aware there were scripts there but the full nature 7 of the script I don't recall ever seeing. 8 Q. So you saw extracts from what appeared to be scripts cut 9 into a range of PEAKs -- 10 A. Yes. 11 Q. -- and saw them in that way? 12 A. Yes. 13 Q. Was it part of the SSC's function to either write or 14 verify the accuracy of the content of scripts? 15 A. I don't recall it ever being a responsibility of the SSC 16 to produce scripts or verify the accuracy. I do recall 17 on occasions Service Delivery Managers saying, "What 18 questions do we need to ask in order to find out the 19 root cause of this particular issue?" 20 Q. So contributing to the content in that way? 21 A. Yes. 22 Q. But were you aware of any quality assurance process of 23 the scripts? 24 A. No. 25 Q. We've heard evidence from some subpostmasters that, when 112 1 they contacted HSH, they were aware that the person they 2 were speaking to was reading from a script and, amongst 3 other things, they were told they should take no action, 4 that a shortfall would probably resolve itself. Did you 5 see any scripts of that nature? 6 A. Not that I can recall. As I said, I don't recall seeing 7 any complete scripts at all. 8 Q. And that others have suggested that they were told that, 9 under the subpostmaster contract, they were liable to 10 repay to Post Office any and all shortfalls and, 11 therefore, they should do so? 12 A. I was not aware of the terms of the contract between 13 Post Office and subpostmasters probably until March this 14 year when it became obvious from the documents that the 15 Inquiry had sent to me. If I had been made aware of 16 that, then I don't recall when it was done and it 17 obviously didn't register. 18 Q. Can we turn to paragraph 43 of your witness statement 19 which is just underneath. Thank you. 20 You speak about the Service Management Portal -- 21 A. Yes. 22 Q. -- between this paragraph and paragraph 49. You say it 23 was: 24 "... an initiative proposed by the then customer 25 service director (Dave Baldwin), to be written by the 113 1 SSC as something completely separate from the live 2 estate and the formal development of Horizon ... created 3 in about 2006, and it was a prototype to demonstrate the 4 ways in which Fujitsu could display data of interest to 5 POL." 6 Then over the page, please, you summarise in 7 paragraph 44 -- I'm not going to read it all now -- the 8 incidents which the SMP would deal with or the matters 9 that the SMP would deal with, including a number of 10 system functions, including Alliance & Leicester Network 11 Banking, overall failure account on Network Banking 12 transactions, ePay connections for E-Top Ups, 13 et cetera? 14 A. Yes. 15 Q. What was your involvement in the creation of the Service 16 Management Portal? 17 A. I wrote large portions of it and included it in several 18 tools that had been developed within the SSC. 19 Q. Can we look, please, at FUJ00142216. If we look on the 20 second page, we can see your name as an "Originator". 21 Does that mean author? 22 A. Yes. 23 Q. Then, if we look at page 4 of the document, please, 24 there should be an index, if we just zoom out, please. 25 You'll see that we get an idea of the size of the 114 1 document and its coverage, yes? 2 A. Yes. 3 Q. So, after the introductory section there are five main 4 sections: Section 4, "The Management Monitors"; 5 Section 5, "Major Incident Management"; Section 6, 6 "Reports and SLTs"; then, over the page, please, 7 Section 7, "Operational Business Change"; and section 8, 8 "Operational Change". 9 Overall, standing back from the document and 10 without going through all 70-odd pages of it, was the 11 idea of the portal to provide a high level overview of 12 the status of the component parts of the Horizon System? 13 A. Both high level overview in real-time monitoring and 14 more detailed description of performance against SLTs. 15 Q. It was also a record of business and operational changes 16 that it was proposed would be made? 17 A. Yes. 18 Q. Its purpose was not to record or to summarise particular 19 bugs or the effect of particular bugs on subpostmasters? 20 A. No. 21 Q. Neither the document nor the Portal? 22 A. No. 23 Q. In practice, is this right: the Service Management 24 Portal was not used to record or to summarise particular 25 bugs or the effect of bugs on subpostmasters? 115 1 A. No, I don't recall that ever being my intent when 2 I wrote it. 3 Q. And, in practice, it wasn't used for that purpose? 4 A. No. 5 Q. Thank you. 6 If we just go back to paragraph 44 of your witness 7 statement, please, which is on page 18, you say in that 8 paragraph that the Post Office (POL) could also check on 9 the current state of an individual Post Office? 10 A. Yes. 11 Q. The data which was used to provide the information was 12 detailed on the Service Management Portal. 13 A. Yes. 14 Q. Do you know who within POL was provided with access to 15 the system in this way? 16 A. At the time, I knew in terms of names because I had to 17 set up individual users with a username and password. 18 Their roles within Post Office, no, I never knew. 19 Q. How many were there? 20 A. No, I'm sorry, I can't remember that. 21 Q. Are we talking 5, 50, 500? 22 A. We're talking tens but not hundreds. 23 Q. What sort of information was held on this system 24 concerning individual branches? 25 A. As I recall, there was a map of the UK with areas that 116 1 would go red if a post office was not operating, details 2 of the nearest working post office and, for any post 3 office, if you put in the FAD code, you would get back 4 details of address and, I think, number of counters. 5 Q. And that was the extent of the data? 6 A. In that particular section, as I recall, yes. The user 7 guide would detail exactly what was available in there. 8 Q. Yes, and we've got it. Thank you. 9 A. Okay. 10 Q. Lastly from me, can we discuss another way of sharing 11 information with the Post Office and look at 12 paragraph 107 of your witness statement, please, which 13 is on page 35. I should start with 105 to give some 14 context. Under "Sharing information" you say: 15 "There were no procedures or work instructions of 16 which [you were] aware that restricted the flow of 17 information about any workaround or potential bug 18 anywhere in Horizon." 19 A. Correct. 20 Q. Then, 107, you say: 21 "When incidents were closed, this would be 22 communicated to the postmaster who raised the incident 23 by the HSH/SMC. In cases where the SSC was 24 communicating with a postmaster about an incident, SSC 25 staff would sometimes agree closure of the incident with 117 1 the postmaster. If a workaround was being applied, POL 2 would sometimes liaise with the postmaster as to when 3 the workaround was to take place -- for example, if 4 messages needed to be inserted into the counter message 5 store. As I recall, these types of workaround required 6 liaison between Fujitsu, the postmaster and POL because 7 (a) the postmaster would have to have the Post Office 8 branch open (otherwise transactions would appear after 9 end-of-day processing and cause failures), (b) Post 10 Office branch staff would have to log out of their 11 counters (otherwise the Riposte sequence number would 12 mismatch and cause error), and (c) POL would also be 13 asked to confirm that the inserted transaction had been 14 successful. However, workarounds that were applied to 15 data centre systems were not always agreed, or discussed 16 with POL." 17 Can you explain why workarounds that applied to 18 data centre systems were not always agreed or discussed 19 with POL? 20 A. If I may, I would like to correct some of the statements 21 that come before that. 22 Q. Yes, go ahead. 23 A. I think I was asked specifically would it be possible to 24 put messages into a message store out-of-hours. I can't 25 remember who asked me that, to be honest. 118 1 So what is written there was in response to that 2 particular question and why it would not work. It's not 3 to be taken as detail of the process. Was that not 4 clear? 5 Q. No, it's not clear. What's the correction you wish to 6 make? 7 A. Yes, I recall that it required liaising between Fujitsu, 8 the postmaster and POL but the reason behind that, in 9 the working day, was because the counter needed to be 10 left switched on, logged in and left alone. That was 11 against POL procedures, which is why we sought POL's 12 authorisation for any such addition to a message store 13 and why it required us to talk to the postmaster to 14 arrange a specific time. 15 Q. Can we go back to the last sentence: 16 "[The] workarounds that were applied to data 17 centre systems were not always agreed, or discussed with 18 POL." 19 Why was it that workarounds that applied to data 20 centre systems were not always agreed or discussed with 21 POL? 22 A. As I recall, there were occasions where the error had 23 arisen in the calculations inside Fujitsu code in the 24 data centre. Again, a hypothetical example: three post 25 offices sell three stamps and the system then turns 119 1 around and says, "Yes, ten stamps sold, receipts for 9". 2 Clearly, that's a software bug. Equally clearly, there 3 is a correction required to the data but it does not 4 affect any of the individual post office's branch 5 accounts and, therefore, there is no point in 6 contracting them. 7 Q. Was it the case that no contact was made principally on 8 the basis that, for logistical or technical reasons, 9 which you list in this paragraph, it was necessary to 10 make contact in order to allow the workaround to be 11 carried into effect? 12 A. I'm sorry, I'm not sure I understood that question. 13 Q. You have described in this paragraph occasions on which 14 liaison was required between Fujitsu, the subpostmaster 15 and POL. 16 A. Yes. 17 Q. You told us that some workarounds were effected, you 18 have just given an example of one, without such contact 19 with either POL or the subpostmaster. 20 A. Yes. 21 Q. Was the reason that no contact was made with them 22 because it wasn't necessary to make contact with them to 23 carry out the fix? You didn't need to ask them to leave 24 their machine on, for example? 25 A. No. The whole process about leaving the machine on 120 1 refers solely to the very rare occasions where messages 2 needed to be inserted into the branch message store. 3 Where it was not necessary was where corrections were 4 being made to the central systems and where there was no 5 implications for the individual branch's accounts. 6 Q. Wouldn't POL wish to know that there was a bug in their 7 system? 8 A. Yes, and they would know because the daily reports on 9 receipts and payments would not match. 10 Q. I thought you told us that it would have no impact. 11 A. It would not have an impact on individual branches. It 12 would have an impact on the overall concatenation of the 13 data. Where it would have an impact on an individual 14 branch, then we would know which branches and the 15 branches would have been told by SSC. 16 Q. So is it your evidence that, for each and every bug that 17 was discovered by Fujitsu, either the postmaster was 18 told directly or POL were told indirectly through the 19 provision of the data that you've just mentioned? 20 A. No, that's not true for every bug because there were 21 bugs in the central systems which would not have 22 implications of that sort. 23 Q. What do you mean implications of that sort? 24 A. Any sort of financial implications on either branch or 25 POL business. 121 1 Q. Were they the only ones that were kept from the Post 2 Office then, ones where Fujitsu assess that there was no 3 financial impact on anyone? 4 A. No, I think that's reading something into what I said 5 that isn't there. There was, in my mind, no case in 6 which POL could not be told of any bug in the system. 7 Q. And I'm asking you, of all of the bugs of which you were 8 aware, were they told? 9 A. No, because I am aware that there were incidents in the 10 central systems which would not be communicated unless 11 they became problems. 12 Q. Is the existence of a bug not a problem about which POL 13 should be told? 14 A. No. The existence of a bug can cause an incident. Only 15 if it is deemed important or there are multiple sorts 16 will it be a problem and, as far as I am aware, problems 17 were always communicated to POL. It's very difficult to 18 be specific about this without looking at an individual 19 problem that may have occurred in a central system 20 overnight. 21 MR BEER: Thank you very much, Mr Peach. Those are all the 22 questions I ask. 23 MR STEIN: Sir, I have a few questions for Mr Peach. May 24 I go ahead now? 25 Questioned by MR STEIN 122 1 MR STEIN: Mr Peach, my name is Sam Stein. I represent 157 2 subpostmasters and mistresses. I'm instructed by the 3 solicitors Howe+Co. 4 Can you just help us a little bit more to 5 understand what people within the SSC were doing by way 6 of the more general nature of work. Is it right that 7 not everyone was engaged in fixing bugs and problems but 8 other work was going on; is that correct? 9 A. Yes, it is. 10 Q. Now, the other work that was going on, and just help us 11 a little bit more on this, is that there was pressure to 12 complete testing; is that correct? 13 A. No, SSC were not directly involved in testing except for 14 workarounds that were produced in the SSC. 15 Q. Right. So regarding workarounds that had to be looked 16 at by different members of the team; is that correct? 17 A. No a workaround for any specific incident would need to 18 be tested before going out to the live estate. 19 Q. Right. So that work was ongoing? What about new 20 features for the Horizon System? Was that part of the 21 work that was ongoing? 22 A. Only in as much as we would review the technical 23 documents that came from development and architecture 24 prior to them going live and review support guides for 25 each product that was going into the estate that was 123 1 new. 2 Q. Later on, did that SSC also play a role in relation to 3 Horizon Online? So as we move from the early days of 4 what we call Legacy Horizon into Horizon Online, did it 5 play a role there as well? 6 A. Yes. We specified a large number of requirements to 7 Development in order to help us support the system and 8 one of my staff was put on more or less permanent 9 secondment to the Development teams to try and ensure 10 those requirements were met before the system went live. 11 Q. Okay. So we've looked at, first of all, with Mr Beer, 12 King's Counsel, in the questions that he has asked you 13 today, about the different roles of the different tiers 14 of problem solving, the taking the calls, the more that 15 developed, the more expert individuals that engaged in 16 solving problems, the workaround, the testing within 17 that that had to go on and then also the work that was 18 ongoing as regards development of Horizon Online. 19 A. Yes. 20 Q. Have I missed anything? 21 A. Yes. 22 Q. Okay. Because it seems as though there's a lot going on 23 here. 24 A. Yes. SSC developed both the KEL system and PEAK, so we 25 were effectively fourth line support for those 124 1 particular applications. We did a large number of ad 2 hoc requests for data which came usually from POL and 3 usually through the Service Delivery Managers. So POL 4 would need certain information, they'd talk to the 5 Service Delivery Managers, the Service Delivery Managers 6 would come and ask us whether it was feasible, how long, 7 how much. 8 There were other things going on but I'm 9 struggling to remember. 10 Q. Okay. All right, so we've got a very good idea in 11 outline terms about the number of different things that 12 were happening. 13 Now, presumably, this was all part and parcel of 14 the delivery by Fujitsu of its contractual obligations 15 to the Post Office, yes? 16 A. Yes, and in the case of the ad hoc data requests beyond 17 the contract. 18 Q. Just to clear up one matter, was the SSC or any of its 19 staff members also engaged on other projects, so working 20 on maybe development of other systems for other 21 companies? 22 A. Not while I was manager. 23 Q. So we know that the SSC, therefore -- you used the term 24 workarounds. One of the big drivers here is to make 25 sure that the system remains up and running, is 125 1 servicing and need of the branches and the Post Office; 2 is that right? 3 A. Yes. 4 Q. So a number of things all going on. There's 5 a contractual obligation to try to do what it can for 6 the Post Office, whole thing needs to be kept going, 7 yes? 8 A. Yes. 9 Q. It sounds as though it's a fairly, how can I put it, 10 pressured or busy place? 11 A. Busy, yes. Pressure is subjective. 12 Q. All right. Now, help us little bit more in relation to 13 the way that things worked. We know from your evidence 14 that what you are saying is that you weren't aware until 15 around, what was it, 2006/2007 that there was the use of 16 the Horizon System to support court cases. 17 A. And, to my memory, only in 2006 and I didn't -- I wasn't 18 aware that it was being used for any other cases. 19 Q. Right. Help us a bit more with your line management, 20 okay. So you're in charge of the SSC. Who's in charge 21 of you? Who's your boss? 22 A. On occasion it was the Customer Service Director. More 23 usually -- 24 Q. Name -- can you name the individuals as we go, please? 25 A. Customer Service Director, Stephen Muchow, Martin 126 1 Riddell, Dave Baldwin, Naomi Elliot, Wendy Warham, 2 possibly two or three more I've forgotten. 3 Q. The different names you are giving, are they all people 4 who held a line managerial job relation to you over 5 a different period of time or were they all together 6 managing you? 7 A. No, over different periods of time. So those were the 8 Customer Services Directors. I did not report direct to 9 all of those. A lot of the time I reported to the 10 Support Services Manager who, in turn, reported to the 11 Customer Service Director. 12 Q. Okay. So, in terms of your knowledge about the use of 13 the data from Horizon System, it going to support or 14 going to be used as part of court cases. Now, you 15 probably know by now the sort of court cases we're 16 talking about are the prosecutions of individual 17 subpostmasters and mistresses? 18 A. Yes. 19 Q. You know by now that that also included some people 20 being prosecuted, a failure of disclosure and people 21 going to prison. 22 A. Yes. 23 Q. You must have read all about that? 24 A. Yes. 25 Q. You also must be aware by now that we're not just 127 1 talking about the use of the criminal system but also 2 data from the Horizon system was used in relation to 3 civil prosecutions, civil cases before the county 4 courts, yes? 5 A. Yes. 6 Q. A serious business. 7 A. Yes. 8 Q. Now, help us understand a little bit more. The people 9 that you have mentioned, the people in your line 10 management, who out of those people do you think should 11 have told you that, "Look, Mik, we've got this system, 12 it's going to be used by the Post Office, they're going 13 to prosecute people with it, just so you know, you know, 14 make sure you design the thing, make sure you get your 15 people working to that level". 16 A. Yes. 17 Q. Who should have told you that? 18 A. That's extremely difficult to answer because it's one of 19 those cases where you don't know what it is that you 20 don't know. Given that I was not aware that data was 21 being used in that way, how could I say who should have 22 told me? 23 Q. Let's try it another way round. 24 A. I would have expected it to come through my line 25 management. 128 1 Q. Right. Was that Mr Muchow? 2 A. I was not aware of any cases when Steve Muchow was my 3 manager. 4 Q. All right. Let's try this another way. Do you think 5 it's unimportant, do you think it doesn't matter at all, 6 that the SSC was not put into the state of knowledge 7 that this stuff is going to be used for core purposes, 8 think it doesn't matter? 9 A. No, I certainly do not think that -- 10 Q. No, well, quite. Let's go back then to my question: who 11 should have told you? If it's something that was 12 important and you weren't told, who should have told 13 you, Mr Peach? 14 A. Again, I would have expected that to have come through 15 my line management. 16 Q. Right. Well, over that period of time can you name the 17 individuals that should have told you this important bit 18 of the business? 19 A. Over which period of time? 20 Q. Well, why don't you start from the first manager that 21 you can think of that should have told you that and name 22 them, please. 23 A. The first case that I knew of was 2006. 24 Q. No, Mr Peach, let's try this another way round. We've 25 got a situation whereby you agree it's important that 129 1 you should have been told that the information from the 2 Horizon System was going to be used in court 3 proceedings, right. You're saying you weren't told 4 that, okay? 5 A. Right. 6 Q. Right. Who out of your line managers should have told 7 you that? 8 A. If they knew, the entire list of my line managers. 9 MR STEIN: Excuse me one moment. (Pause) 10 Thank you, sir. 11 SIR WYN WILLIAMS: Any other questions? It looks as if 12 that's it, Mr Peach. 13 MR BEER: No, it's not there's two lots of questions to 14 come. 15 SIR WYN WILLIAMS: I see. 16 MR BEER: I think Ms Page is going to go first and then -- 17 SIR WYN WILLIAMS: I'm conscious that we have been going for 18 longer than an hour and I don't want things to be too 19 difficult for Mr Peach. Do we need a short afternoon 20 break? 21 MR BEER: I will just check, if they can hold up their 22 hands, for how many minutes they are going to be. Five 23 for one and ... 24 MS PAGE: It may be more like 15. 25 MR BEER: I think that sounds like an afternoon break then, 130 1 sir. 2 SIR WYN WILLIAMS: Let's have our break now then, sir. 3 MR BEER: Can we come back at 3.15, please? 4 SIR WYN WILLIAMS: Certainly. 5 MR BEER: Thank you. 6 (3.00 pm) 7 (A short break) 8 (3.15 pm) 9 MS PATRICK: Sir, can you see and hear me. 10 SIR WYN WILLIAMS: Yes, I can thank you. 11 Questioned by MS PATRICK 12 MS PATRICK: Thank you. 13 Mr Peach, my name is Angela Patrick and I ask 14 questions on behalf of a number of subpostmasters who 15 were wrongly convicted and then acquitted and I am 16 instructed by Hudgell Solicitors led by Mr Tim Moloney. 17 We only have a couple of questions about two documents. 18 One you will be quite familiar with, and I'm not 19 going to spend too long on, is Mrs Chambers' January 20 2007 afterthought document, which Mr Beer has taken you 21 to at some length. It's FUJ00152299 and you will be 22 happy to hear I only want to look at the last two 23 paragraphs of the document, which are on page 2. 24 I don't think we need to read it again. These are 25 the two paragraphs that deal with what she, 131 1 Mrs Chambers, calls the common problem of SPMs, 2 subpostmasters, being bounced from one Helpdesk to 3 another and she says about some calls being closed 4 without proper investigation or resolution when they are 5 bounced back, and those calls are problems with 6 discrepancies needing to be investigated but personally 7 they are being penalised for passing some back. 8 Is that fair? She's basically saying there might 9 be a problem about calls, about discrepancies, being 10 bounced back inappropriately; is that right? 11 A. I agree that, from what she says, calls were being 12 bounced between Horizon and the NBSC but she does make 13 the point in there that that resulted in it taking 14 a long time before the call got to the SSC. 15 Q. Okay. 16 A. I can't answer for the processes in NBSC and HSH, 17 sorry -- only for the SSC. 18 Q. Okay. Do you recall now whether at the time you took 19 any specific steps to follow up on that concern she was 20 raising? 21 A. It sounds as though I'm being evasive and I'm really 22 trying not to be. The first time I saw this document 23 was at about 9.45 this morning, other than when it 24 happened and I didn't recall. I would think that Anne 25 would have made sure I took some action but what that 132 1 action was, I couldn't say because I don't recall the 2 document. 3 Q. Just to be absolutely clear, and I know that it's some 4 time ago, having had a bit of time to reflect, can you 5 recall whether it was raised with the Post Office? 6 A. I can't recall but, since NBSC were part of Post Office, 7 I would have expected them to have known from NBSC. If 8 calls being bounced between Horizon and NBSC had been 9 raised within Fujitsu, then it would have been raised 10 with the Service Delivery Manager, ie problem manager, 11 responsible for that area and I would have expected it 12 to have been notified to POL through that route. 13 Q. Thank you. Taking that on board, can we look at another 14 document and it's FUJ00080096. 15 If we can look at the top of the page to start 16 with, can you see that, Mr Peach? 17 A. Yes. 18 Q. I think this is a joint working document. We've seen 19 these before. It's got both logos on the top; is that 20 right? 21 A. Sorry -- yes. 22 Q. You'd recognise that as a joint working document, would 23 you? 24 A. Yes. 25 Q. We see the title there. It's a document on the 133 1 operation of the Horizon Service Desk and it's 2 identified expressly as a joint working document. If we 3 look at the very bottom of the page we can maybe put 4 a date on it. Thank you very much. 5 Can you see at the bottom right-hand side there 6 it's dated 4 September 2008. So that would have been 7 about a year, or more than a year, after Mrs Chambers' 8 feedback in her afterthought note, which was 9 January 2007? 10 A. Yes. 11 Q. If we see there above in the approval authorities, you 12 weren't an approval authority for this document, would 13 you, Mr Peach? 14 A. No. 15 Q. If we look at page 27 though, near the bottom of the 16 page, we can see you're named there four boxes up from 17 the bottom as a key contact for major software problems? 18 A. Yes. 19 Q. Can you recall whether at the time you may have been 20 consulted on this document, which was, as we see, 21 a joint working document for the operation of the 22 Service Desk? 23 A. I don't recall the document. I would have expected to 24 have been consulted on procedures relating to software 25 problems, which were going to go through the SSC. 134 1 Q. Okay. Now, if this includes instructions for the 2 Service Desk, including when something would be directed 3 towards the SSC or elsewhere, as the manager of the SSC 4 at this time in 2008, would you expect to have been 5 shown it? 6 A. This document, yes, I would have expected to have been 7 shown it. 8 Q. Can we look at page 11, please. About part of the way 9 down this page, you can see at the bottom there's 10 a header 3.4 "Inappropriate Calls". This is the part 11 that I'd like us to look at together. It reads: 12 "Inappropriate calls are defined as below; all 13 inappropriate calls are logged on the Incident 14 Management System. If the incident has been 15 investigated by an Agent and the call is deemed to be 16 an inappropriate call, the desk Agent will give the 17 postmaster an Incident number and either transfer the 18 postmaster to the NBSC or cold transfer the postmaster." 19 We see a list there of what calls are considered 20 to be inappropriate. Can you see the fourth bullet 21 point there, Mr Peach: 22 "Caller requests advice on cash 23 account/discrepancy issues." 24 Is this document advising the HSH, the Helpdesk, 25 to treat calls from subpostmasters about cash account 135 1 and discrepancy issues as inappropriate calls? 2 A. The way I would read that would be to say, before any 3 call of that sort was dealt with by HSH, it should have 4 gone to NBSC first. 5 Q. But the direction there is "inappropriate calls are 6 defined as below" -- 7 A. Yes. 8 Q. -- to include calls about cash account and discrepancy. 9 A. Yes. 10 Q. And the direction is to transfer to NBSC or cold 11 transfer. What was a cold transfer? 12 A. I have no idea. 13 Q. But, essentially, they are being told to bounce them 14 back to the NBSC? 15 A. I would not disagree with your interpretation of that. 16 Q. Was that exactly the problem that Mrs Chambers was 17 raising concerns about in January 2007? 18 A. That's the way I would read her statement, yes. 19 Q. If Mrs Chambers' feedback had been taken up with Fujitsu 20 management and with POL management, as you said you 21 would have expected it to have been, it wasn't reflected 22 in this new joint guidance to the HSH, who were 23 providing front line support to subpostmasters, was it? 24 A. It does not appear to have been but, as I said before, 25 I do not know the processes that were operating in 136 1 either NBSC or HSH in this area, so I don't know what's 2 behind that. 3 MS PATRICK: Thank you. I don't have any more questions. 4 SIR WYN WILLIAMS: Thank you. 5 Ms Page next then. 6 Questioned by MS PAGE 7 MS PAGE: Thank you, sir. 8 I also act for a group of subpostmasters, 9 including Mr Lee Castleton. 10 A. Right. 11 Q. Can I start, please, with the document which you've seen 12 some of but I'd like to take you to a different bit of, 13 which deals with the quality of the EPOSS code and the 14 question of whether it was going to be rewritten. If we 15 could look at WITN04600104. You were taken initially to 16 page 9 and you told us that, when it came to a debate 17 about whether the code should be rewritten, you were not 18 surprised that that wasn't communicated to you because 19 that was a matter for the Development team, yes? 20 A. Yes. 21 Q. If we look at page 10, we can see what happened with 22 respect to the debate about whether it ought to be 23 rewritten and, in that column headed "Agreed Action 24 Commentary", we can see that, ultimately, on 10 May -- 25 and this is in the year 2000 -- that large block at 137 1 bottom: 2 "Following response received from MJBC: as 3 discussed, this should be closed. Effectively, as 4 a management team we've accepted the ongoing cost of 5 maintenance rather than the cost of a rewrite. Rewrites 6 of the product will only be considered if we need to 7 reopen the code to introduce significant changes in 8 functionality. We will continue to monitor the code 9 quality (based on product defects) as we progress 10 through the final passes of testing and the introduction 11 of the modified CI4 codeset into live usage in the 12 network. PJ, can we make sure this is specifically 13 covered in our reviews of the B&TC test cycles. 14 Closed." 15 What I'd like to ask you about is the suggestion 16 there that there needed to be a monitoring of the code 17 quality based on product defects as the modified codeset 18 is put into live usage. As the front-line team dealing 19 with complaints about the quality of the code, you were 20 the team to do the monitoring, were you not? 21 A. No. As I've explained, if there were multiple instances 22 of an incident or if a particular issue was considered 23 important, then problem management process would deal 24 with it. SSC would deal with finding the technical 25 solution to an individual incident. That having been 138 1 said, if the problems encountered were clearly issues 2 with code, then we would know the number of calls that 3 we had passed to Development requesting code fixes. 4 Q. So it would make sense, wouldn't it, for your team to 5 have been told about this so that you could do that job 6 of counting up how many calls you had had about EPOSS 7 code and doing that sort of collecting together 8 exercise? 9 A. That's -- with any release of any software, at some time 10 the decision is made "We're not going to keep rewriting 11 this, we will fix forward", and that is the point at 12 which the code is delivered to the system and the SSC 13 becomes responsible for supporting it. 14 Would I have liked to have known about this 15 decision? Yes, I think I probably would. Was it my 16 decision to make? No, and the people who made the 17 decision were the people empowered to make that 18 decision. 19 Q. It's not about blaming you, Mr Peach. I'm asking really 20 because the question is: would it have been sensible for 21 your team, as the front-line team dealing with the first 22 initial software code problems, to have been able to 23 know about this, so that you could have fed your team's 24 findings in to that ongoing decision about whether there 25 ought to be a rewrite? They were keeping it under 139 1 review, that question? 2 A. Yes, and, as I said, I probably would have liked to have 3 known. Would it have made any material difference to 4 the way we handled incidents, no. 5 Q. No, but it's about how they could have handled the 6 question of whether they should have rewritten the code. 7 A. At the point that this document is written, I don't 8 believe that this code was in live. It's talking about 9 a specific release and it's a discussion between 10 development and testing and a management decision at 11 that point. I would have liked to have known that there 12 were those issues in testing but I don't think I would 13 have added significantly to the discussion. 14 Q. All right. Well, I'll leave it in that case because I'm 15 actually trying to focus on not on whether you would 16 have added to the discussion. But whether it would have 17 been sensible for your team's feedback to go into that 18 discussion. 19 Now, I mean, if you don't want to take that any 20 further, I'll leave it there. 21 A. My team would not have had any feedback to this code 22 because this release had not been -- had not gone to 23 live. 24 Q. No but once it had gone into live? 25 A. Once it had gone into live, knowing that there was 140 1 specific focus and there was supposed to be more 2 monitoring in place, yes, I would have appreciated 3 knowing that. From this document, I can't demonstrate 4 that we were not asked to specifically monitor. I don't 5 remember that we were but I can't demonstrate that we 6 were not. 7 Q. You have no recollection of being asked to monitor. 8 That's as far as we can take that line, I suppose. 9 Thank you. Let's just turn then to another issue, 10 which you've already touched upon and which I'd like to 11 just explore a little further with you. The document 12 FUJ00088082 is one that you've seen a little of. If we 13 could turn to page 15, please, if we scroll down 14 a little from the bit that you've looked at and just 15 a little further -- sorry, I may have got the wrong 16 page. Could we just scroll on a bit further. 17 Sorry, I had the wrong page. So if we go down to 18 7.4. Oh, dear, I seem to have made a bit of a mess. 19 I don't know if this is a document that you've 20 already had sight of but there's a passage where it 21 talks about the SSC having a form of access on a needs 22 must basis. Is that something that you recall seeing? 23 A. It is but I don't think it's this document. I think it 24 was ... 25 Q. You may well be right. In which case, it may be 141 1 FUJ00087994. But it probably doesn't matter actually. 2 I won't ask for that to come up. If you've seen the 3 document, we've all seen it as well. 4 A. Right. 5 Q. That is a document which sets out a sort of a form of 6 access that had developed because it needed to be; is 7 that a fair way of describing it? 8 A. Yes. 9 Q. When that was put to Alan d'Alvarez, he testified that 10 it had gone entirely against what the access control 11 policy says should happen. 12 A. Yes. 13 Q. So this was in 2002. You were obviously aware of it. 14 Was your line manager aware of it, this needs must 15 access? 16 A. Yes. 17 Q. Who would that have been in 2002? 18 A. I don't think that I was working direct to Stephen 19 Muchow at that point. It would probably have been Peter 20 Burden. 21 Q. Peter? 22 A. Burden. 23 Q. Burden. Presumably there were some less than ideal 24 discussions around that issue. I mean, it can't have 25 been something people were happy with? 142 1 A. No, none of us were happy with it. The original access 2 control policy, from my reading of it now and my vague 3 memories of it at the time, assumed that the SSC would 4 be providing detailed code support to the data centre 5 boxes and I believe the assumption was that, in the 6 majority of cases of any problem on the counter, then 7 the counter would be replaced: a straight hardware swap. 8 So the SSC responsibilities with regard to counter 9 were very limited in the beginning, according to the 10 access control policy. My recollection is that there 11 were one or two problems on counters which affected 12 a number of counters and for which hardware swaps were 13 clearly not going to be an option and, therefore, SSC 14 were allowed access to the counters in order to fix 15 those problems. 16 Immediately that that became obvious and that we 17 were contravening the terms of the access control 18 policy, the security architect, who was Glenn Stephens 19 (now, sadly dead), started looking for a solution that 20 would give the SSC the access they needed to fix the 21 system but in a controlled and auditable way, and that's 22 where SSC came from. 23 Q. What I'd quite like to understand -- and thank you for 24 that -- is who -- where did this land at the most senior 25 level? Who was the most senior person who knew about 143 1 this problem? 2 A. I don't know. 3 Q. Was it seen as sufficiently serious for this to have got 4 to the board? 5 A. I'm sorry, I don't know that either. 6 Q. Your evidence about what then happened in terms of the 7 secure shell system, the SSH, was that you understood 8 that that meant that, from that point onwards, there was 9 a full keystroke audit trail for everything that SSC 10 members did when accessing subpostmaster accounts? 11 A. When accessing any part of the live system. 12 Q. But certainly subpostmasters accounts? 13 A. Certainly access to counters. Indeed, SSC staff used 14 their PCs to get into the SAS servers and, from that 15 moment onwards, whenever they went into the live estate 16 it is my understanding that all keystrokes were logged. 17 Q. That was your understanding. You told us, however, 18 that, although that was your understanding, you hadn't 19 actually seen evidence of that, as such, but that was 20 your understanding. You never saw any printouts that 21 showed full keystroke logging? 22 A. I did not but then, to be honest, I don't think I ever 23 saw a message store. That was just too technical for 24 me. I know from discussions with one or two SSC staff 25 that the logging was there but only because they said it 144 1 was absolute torture to look at it, because if you typed 2 "T-E-H" instead of "the" and then went back, you would 3 actually see "T-E-H, backspace, backspace, delete, 4 delete, T-H-E". So to read it was very difficult. 5 Q. That suggests that definitely somebody saw it at some 6 stage? 7 A. Yes. 8 Q. If I told you that there was a document considerably 9 later which says that it wasn't happening, it certainly 10 wasn't happening by that stage, you're not aware, are 11 you, of any decision that was taken to stop doing it? 12 A. No, I would question the date on that subsequent 13 document. 14 Q. It was after you left. It was into the next decade. 15 A. Okay. Then -- 16 Q. But you weren't aware of a decision during your time to 17 stop doing keystroke logging? 18 A. No. 19 Q. We've also been told that there was a four-eyes process 20 and it was important to record the fact that one person 21 was going into the message store had another person was 22 watching them do so, and that there was then a narrative 23 account of what the insertion into the message store had 24 been and what it was for put into the PEAK. 25 A. Yes. 145 1 Q. So some effort would go into that and, indeed, it would 2 involve a doubling up of manpower? 3 A. Yes, and it wasn't just for corrections being made in 4 the counters. Corrections made on the data centres with 5 any financial impact would also be subject to four eyes. 6 Q. If you believed there was a full keystroke audit trail, 7 what was the point? Why did you bother with the four 8 eyes process? 9 A. That's a good question. Because it never occurred to me 10 to remove the process. 11 Q. Finally, if I may turn to the subject of Anne Chambers 12 and her evidence in the Castleton trial, it's your 13 evidence, if I understand it correctly, that Ms Chambers 14 was put into a difficult position when she was asked to 15 give evidence in the Castleton trial; is that right? 16 A. Yes. 17 Q. According to her document after the event, she really 18 identified various failings of others within Fujitsu who 19 had failed to take full investigative steps ahead of the 20 trial, failed to do a proper investigation ahead of the 21 trial, yes? 22 A. Yes. Again, I would stress that the first time I can 23 recall seeing that document was early this morning. 24 Q. But the result of that really, according to those two 25 things put together, her being put in a difficult 146 1 position and her understanding that others should have 2 done a better investigation, is that she was really 3 rather unfairly put into the firing line. 4 A. I completely agree. 5 Q. Who do you hold responsible for that? 6 A. No, I couldn't say. I really ... 7 Q. You had an argument about this. You were angry with 8 somebody about this. 9 A. Oh, yes. 10 Q. Who was that? 11 A. As I said this morning, I know that I had a stand-up 12 argument and I know it was one of three people but what 13 I can't recall is which one of those it was. 14 Q. Were you angry with all three of them? 15 A. Very probably. 16 Q. So one of those people was the Head of Security, 17 Mr Utting -- so sorry, not Mr Uttering, that was POL's 18 Security -- Brian Pinder, Fujitsu's Security, and one of 19 them was a person who may or may not at that time have 20 been the Customer Services Director -- 21 A. Yes. 22 Q. -- her name being Naomi Elliot? 23 A. Yes. 24 Q. Did she come from a Security background before she 25 became Customer Services Director? I ask that because 147 1 the document that we just looked at with Ms Patrick 2 named her alongside Mr Pinder? 3 A. I don't know Naomi's background. What I can tell you is 4 that she was very astute technically, so I would have 5 thought she would have come from an engineering 6 programming background, rather than Security. 7 Q. Was Security under the same director as your team, in 8 the sense that they were also under Customer Services? 9 A. Yes. 10 Q. So we can take it, can we, that either the Head of 11 Security or the director who was in charge of Security 12 was ultimately responsible for putting Anne Chambers in 13 the firing line? 14 A. Ultimately responsible, yes. 15 MS PAGE: Thank you. I have no further questions. 16 Questioned by SIR WYN WILLIAMS 17 SIR WYN WILLIAMS: I'm sorry, just to revisit what Ms Page 18 has asked you but I wanted to be clear that I got your 19 earlier evidence correct. 20 At paragraph 154 of your witness statement -- it 21 doesn't have to come up, I can just read the relevant 22 words -- you say: 23 "I was instructed by the Director of Customer 24 Services at the time, whose name I cannot recall, to 25 detail someone from SSC." 148 1 A. Yes. 2 Q. You actually recall it was the Director of Customer 3 Services but you couldn't identify that person by name. 4 When Mr Beer was asking you questions, you appeared to 5 narrow it down from three to two, in that the relevant 6 Director of Customer Services was either Mr Pinder or 7 Ms Elliot, all right? Now, you have said something 8 slightly different to Ms Page, so I just want to be 9 clear what it is that I'm left with on this topic, which 10 is obviously of some importance. 11 So am I correct to keep to my earlier note that it 12 was the Director of Customer Services who instructed you 13 and that, by a process of elimination, was either 14 Mr Pinder or Ms Elliot, or do you want me to amend that? 15 A. My recollection, sir, is that it was the Customer 16 Services Director who instructed me. Brian Pinder was 17 Head of Security reporting to the Customer Service 18 Director. He was not Customer Service Director. At 19 some point, Naomi Elliot, I believe, was Customer 20 Service Director and at a different time. Dave Baldwin 21 was Customer Service Director, and I cannot recall which 22 of those two was in that post at that time. 23 SIR WYN WILLIAMS: I'm sorry I misread my own note. That 24 is, in fact, what you said to Mr Beer. I take it that 25 this conversation, this instruction, must have been 149 1 happening some time in the autumn of 2006? 2 A. I have deduced that from the same documents without 3 direct memory. 4 SIR WYN WILLIAMS: So that if we were to ask Fujitsu who was 5 the Director of Customer Services between, shall we say, 6 August 2006 and December 2006, we'd probably deduce who 7 that person was? 8 A. Indeed. 9 SIR WYN WILLIAMS: Fine. All right thank you very much. 10 I think that concludes the questioning, am I right 11 in that, Mr Beer. 12 MR BEER: Yes, you are, sir. 13 SIR WYN WILLIAMS: All right. So thank you very much, 14 Mr Peach, for coming to give evidence to the Inquiry and 15 for answering very many questions over the course of the 16 day and also thank you too for a detailed witness 17 statement in advance. 18 So that I think, Mr Beer, just leaves tomorrow 19 and, ordinarily, the assessors and I would appear 20 tomorrow but, unfortunately, because of an appointment 21 which I can't change, which is early in the morning in 22 South Wales, we'll be appearing remotely. So that's the 23 reason why I have asked for a 10.30 start tomorrow. 24 My understanding is that we've got two short 25 witnesses and then you may wish to say something about 150 1 the written evidence relating to this phase, Mr Beer, 2 and then we're going to hear closing oral submissions. 3 Now, just for the avoidance of any doubt, there 4 will be -- this is not the complete closure of Phase 3 5 tomorrow because we have Mrs Chambers to return and 6 Mr Gareth Jenkins to appear, and that's going to happen 7 I believe, in July so that those who are making final 8 submissions, whether orally or in writing, can rest 9 assured that if either of those persons say anything 10 when they appear which they wish to address me about 11 they can do it when they are making submissions at the 12 end of Phase 4. 13 Does that sound reasonable? 14 MR BEER: Sir, yes, that's all correct and it does sound 15 reasonable. Thank you. 16 SIR WYN WILLIAMS: Fine. Well, then I will see you remotely 17 tomorrow. 18 MR BEER: 10.30, sir, thank you. 19 (3.52 pm) 20 (Adjourned until 10.30 am the following day) 21 22 23 24 25 151 1 I N D E X 2 MICHAEL EDWARD PRYOR PEACH (affirmed) ............1 Questioned by MR BEER ............................1 3 Questioned by MR STEIN .........................122 4 Questioned by MS PATRICK .......................131 5 Questioned by MS PAGE ..........................137 6 Questioned by SIR WYN WILLIAMS ................148 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 152