WITNO0460100
WITNO00460100
Witness Name: GARETH IDRIS JENKINS
Statement No.: WITN00460100
Dated: 6 FEBRUARY 2023
POST OFFICE HORIZON IT INQUIRY
FIRST WITNESS STATEMENT OF GARETH IDRIS JENKINS
I, Gareth Idris Jenkins, will say as follows:
1. I make this witness statement in response to a request under Rule 9 of the
Inquiry Rules 2006 regarding Phases 2 and 3 of the Post Office Horizon IT
Inquiry (“the Inquiry”). Time was extended for my response to this request.
2. I want to make clear (as has previously been stated on my behalf) that I support
the work of the Inquiry and I wish to assist it. To that end, I asked the Inquiry
to consider seeking an undertaking so that I could provide the Inquiry with
evidence whilst at the same time being subject to an ongoing criminal
investigation. That criminal investigation overlaps, to the best of my
understanding, with the same issues that this Inquiry is investigating, in
particular in Phase 4. I understand that the Inquiry is not minded to seek an
undertaking at the present time, but rather has adopted a ‘wait and see’
approach.
3. I was made a Core Participant to this Inquiry on 16 November 2002. Since
then, I understand that my lawyers have been provided with access to the
evidence which the Inquiry has disclosed to Core Participants. However, I also
understand that, as matters stand, there is not much evidence that relates to
Phase 4 which has currently been disclosed. I understand that this is because
Page 1 of 15
WITNO0460100
WITNO00460100
disclosure is on a rolling basis and that witnesses are not due to be called in
Phase 4 until later in 2023.
The position that I find myself in is that I am subject to a criminal investigation
which is continuing. I have attended voluntarily for interviews but I understand
that I will continue to be asked to do so as the police continue to analyse
evidence. I do not yet know the totality of the police investigation into my
conduct, the totality of the allegations that I face or the totality of the evidence
the police may have gathered. At the same time, I am asked to give evidence
to the Inquiry but without having seen all of the evidence which the Inquiry
holds (or is seeking) particularly in relation to Phase 4.
The Inquiry asked me a number of questions, many of which contained detailed
sub-questions. I wish to maintain my reliance upon the privilege against self-
incrimination in relation to certain of these questions for the time being, but I
want to make clear my reasons for doing so. As matters stand, my lawyers are
working their way through the evidence disclosed to Core Participants by the
Inquiry and will continue to do so. I understand that, at the present time, they
are unable to advise me fully on the privilege against self-incrimination,
particularly in relation to Phase 4 because their review is ongoing and because,
as explained above, the existing evidence disclosed in relation to Phase 4 is
limited.
I therefore wish to ensure that it is understood (as I also understand my lawyers
will communicate in correspondence to the Inquiry) that I will continue to keep
the question of whether I waive my rights under consideration. My lawyers
have undertaken to the Inquiry to notify it in good time when they are in a
position to advise me fully. I want to emphasise how seriously I take these
issues. I am simply asking that I be permitted the opportunity to understand
better what allegations are made against me or what evidence relates to me
(and how either relates to the criminal investigation I am currently subject to).
I will continue to prepare for the Inquiry so as to ensure that when my lawyers
are in a position to advise me, that it is not a cause for delay.
Page 2 of 15
10.
11.
WITNO0460100
WITNO00460100
The Inquiry has also provided me with 93 documents (“the Documents’) to
inform the basis of my answers and/or invited my comment on them. Many of
the Documents consist of lengthy and detailed technical data. Whilst this is a
significant volume of material to consider and comment on, it is only a small
snapshot of the material which I saw at the time, and which may be relevant to
the Inquiry’s questions.
The events that I am asked to recall occurred between 10 and 20 years ago.
Many will have assumed a greater importance as time has passed and events
have unfolded, than they did at the time. My memory of them is far from perfect
and I am relying to a large degree upon the Documents to assist my recall. I
am not certain what I knew or understood at particular points in time. This is
made more difficult without a complete documentary record and because I only
learned of certain events after they happened.
In light of the above, and the fact that the events in question span a period of
over twenty years, I have approached this statement, as I have been asked,
by answering the Inquiry’s questions in light of the Documents specifically put
to me and by relying on my memory of the relevant events. I think it is important
to be clear that I have written this statement adhering to this approach, as
requested by the Inquiry. As noted above, my solicitors and I are working
through the Inquiry’s disclosure to all Core Participants and if there is evidence
which touches upon my answers set out below, I will draw that to the Inquiry’s
attention.
I have sought, where possible, to indicate when my knowledge or recollection
is not contemporaneous but is based on subsequent events or what I have
learned after those events. Where it is possible to do so, I have indicated what
other material would assist me in recalling what the position was at a given
time. Again, if there is further material that supplements or changes my
understanding of the events, or my recollection of them, I may wish or need to
supplement or amend what I have said in this witness statement.
Page 3 of 15
WITNO0460100
WITNO00460100
Relevant background
12. I graduated with a degree in Mathematics from the University of Cambridge in
1973 and started work for International Computers Ltd (“ICL”) in September of
that year. I continued working for ICL, which became Fujitsu Services Ltd
(‘Fujitsu’) in 2002, until I retired in 2015. From my retirement until August
2022, I was on a retainer with Fujitsu to provide ad hoc consultancy services.
Throughout all of this time, I worked on the design and development of
computer software.
13. ICL introduced a Distinguished Engineer scheme in the 1980s. I became a
Distinguished Engineer at some point in the mid-1990s, at which point I believe
there were approximately 100 employees who held that title. I believe it was a
purely honorific title designed to recognise the key technicians within the
company. Of the Distinguished Engineers employed by ICL/Fujitsu, I believe
that at least six worked on Horizon at any one time.
14. The Inquiry has asked me to provide an overview of my involvement with the
design, testing, implementation and operation of the Horizon IT system.
Horizon was a bespoke product designed by ICL/Fujitsu specifically for Post
Office Ltd ("POL”). It was a huge, multi-layered processing system that acted
as a conduit for data between POL branch counters (which initially relied on
software called ‘Riposte’) and the various back office functions and third party
agencies (which relied on software based on ‘Oracle’).
15. There were two main versions of Horizon. The original version, now called
Legacy Horizon, was first piloted in 1996 and installed throughout POL
branches between 1999 and 2002. The replacement version, Horizon Online
(also referred to as HNG-X), was piloted in early 2010 and installed throughout
POL branches during the middle of 2010. Legacy Horizon and Horizon Online
were completely different systems as far as POL’s branches were concerned.
16. I hope it is helpful if I clarify that, as far as I can recall, during my employment
by ICL/Fujitsu, there were six main departments within the POL account:
Page 4 of 15
17.
18.
WITNO0460100
WITNO00460100
customer services, requirements, development, testing, program office and
sales & commercial (this later became five teams when requirements merged
with development). From around 1996 to around 2000, I was part of the agent
team, which sat within the development department, along with other teams
including architecture & design, host, counter, security and systems
management. In around 2000, I moved to the architecture & design team within
the development department, until its later merger with the requirements
department, where I remained until I retired in 2015. Throughout all of this time,
my involvement in relation to both Legacy Horizon and Horizon Online was
technical and not managerial or supervisory in nature.
I never held or exercised line management responsibilities for other Fujitsu
staff, whether within or outside the teams in which I worked. Within the POL
account team in ICL/Fujitsu — which comprised hundreds of staff working on
the many facets of the Horizon systems — I was never in a leadership role either
managerially or technically. Throughout this time, there would always have
been one or two people senior to me on the technical side (the Chief Architect
and sometimes a Team Leader below them and above me). I recall that Dave
Hollingsworth was the original Chief Architect, succeeded by Alan Ward, then
Dick Long (although I think he had a different title), then Tony Drahota. After
Tony Drahota left in around 2005, he was replaced by other Chief Architects
whose main focus was on Horizon Online. I recall someone named Giacomo
being in charge, but cannot remember his surname. From the early days of
Legacy Horizon until shortly after the rollout of Horizon Online, I was one of a
team of 10-20 architects working on the project, each with a specific focus on
a particular component or process. This number reduced to three or four by
the time I left. On the management side, I recall there were two or three
gradually increasing levels of seniority between the Account Manager and me.
I started work on Legacy Horizon in 1996. My initial role was to try to integrate
Riposte and the Oracle-based software. I formed a team that developed an
“agent layer” that would allow these two types of software to communicate with
each other. I continued as both the chief designer of this agent layer and as
one of the technical interfaces to Escher (the designers of Riposte) until around
Page 5 of 15
19.
20.
21.
22.
WITNO0460100
WITNO00460100
2002. During this time, extensive testing of Legacy Horizon took place. I
supported ICL/Fujitsu’s testing teams and assisted in identifying and resolving
technical problems that they found.
Between about 2000 and 2002, I was involved — with teams of individuals at
Fujitsu, POL and IBM — in specifying the interface between Riposte and the
online banking engine provided by IBM, as well as enhancing the agent layer
so that it could handle banking transactions. Shortly after the introduction of
network banking (which went live around 2003), I performed a similar role to
enable Legacy Horizon to interface to Streamline, the merchant acquirer that
POL chose to support payments by debit card.
Once these major systems went live, I moved away from the agent team and
took on the role of defining the technical changes required by Project Impact,
which was the name POL gave to the programme to replace their back-end
accounting system. After Project Impact went live in around 2005, I worked on
smaller changes to Legacy Horizon such as the implementation of bureau de
change and the acceptance of credit cards.
In about 2008, I commenced work with POL analysts in defining the technical
requirements for Horizon Online. Horizon Online retained many of the back-
end systems and interfaces from Legacy Horizon, but was otherwise an
entirely new counter system, e.g. it removed all use of Riposte. I supported
POL’s requirements team in specifying “use cases” for the counter functionality
(i.e. distinct functions that the software was required to perform, such as selling
a stamp or making a withdrawal from a bank account), and then worked with
the Fujitsu counter development team in implementing those use cases and
testing Horizon Online.
A further element to my role at this time was assisting in producing the
mechanism by which a counter would transition from Legacy Horizon at the
end of one business day to running Horizon Online the next. Once migration
from Legacy Horizon to Horizon Online was well under way, I worked with POL
on developing further enhancements to Horizon Online to support POL’s
Page 6 of 15
23.
24.
25.
WITNO0460100
WITNO00460100
business requirements. I continued working on these matters until my
retirement in 2015.
I was one of a group of anywhere between around 20 and 100 software
designers and developers who would provide “fourth line” support for Legacy
Horizon and Horizon Online, alongside their day-to-day roles developing new
functionality. The first line of support comprised two helpdesks. SPMs could
phone a helpdesk at POL for any business related issues, whereas for
technical issues, they could phone a helpdesk at ICL/Fujitsu. The second line
of support was an ICL/Fujitsu data centre (the Systems Management Centre,
“SMC”) that received automatically generated reports concerning significant
technical events affecting Horizon, e.g. server failure. The third line of support
was the ICL/Fujitsu System Support Centre (“SSC”), which resolved
operational problems and investigated potential software faults. If the SSC
concluded that the problem was one that they could not resolve or it required
a code fix, only then would they escalate it to fourth line support, which is where
I worked.
The first three lines of support were staffed by dedicated employees whose
sole role was to provide the relevant level of support. Fourth line support was
different; for the software designers and developers who provided support at
this level, this was only a small adjunct to their main day-to-day roles in
producing new functionality for Legacy Horizon or Horizon Online. I estimate
that my responsibilities in fourth line support comprised no more than 10% of
my overall workload, although this varied over time.
As ageneral principle, at the time, I would only have been aware of those bugs,
errors and defects specifically allocated to me by third line support, or where I
was asked to route the issue to the correct person in fourth line support, or
where I was asked for specific input on the problem because of my particular
expertise. Over time, I was gradually allocated more bugs, errors and defects
in my role in fourth line support. However, as noted above, there were about
20 to 100 software designers and developers working in fourth line support at
any one time; all of them were allocated bugs, errors and defects depending
Page 7 of 15
26.
27.
28.
29.
WITNO0460100
WITNO00460100
on their own experience with, and technical knowledge of, the component parts
of the Horizon system architecture. The software designers and developers in
fourth line support did not check or quality assure each other's work. I would
not have been aware of many of the problems referred to, and addressed by,
fourth line support.
In my role in fourth line support, my main functions (which varied in their
intensity over time) were as follows. First, when third line support allocated a
bug, error or defect to me, it appeared in ICL/Fujitsu’s systems as a so-called
PEAK (or before that as a so-called “PinICL”), which was a computerised report
that allowed ICL/Fujitsu to review and track progress on fixing the problem. I
was responsible for addressing and closing the PEAKs allocated to me (unless
I re-allocated them to someone else). Secondly, on occasions, I assisted
ICL/Fujitsu’s Release Management Forum in explaining the nature and impact
of certain bugs, errors and defects, as well as the corresponding fixes. Thirdly,
when asked to, I joined conference calls with POL where I explained to them
the nature and impact of certain bugs, errors and defects, as well as the
corresponding fixes. I address these matters in more detail below.
This fourth line support infrastructure existed throughout my time working on
Legacy Horizon and Horizon Online, both in their testing and operational
phases. PiniCLs and PEAKs were generated and allocated to specific fourth
line software engineers to investigate and address. PinICLs and PEAKs
generated during the testing phases were retained on the system once Legacy
Horizon and Horizon Online went live. I understand that there was an archiving
process for PinICLs and PEAKs but I was not involved in this process.
For the purposes of this statement, I thought it would assist if I addressed the
evidence given to the Inquiry in writing and orally by David McDonnell that
relates to me. I do not have a clear recollection of Mr McDonnell but I was able
to watch his evidence and he looked familiar.
He is in error, in his witness statement, when he attributes to me the role of
Chief Architect. At the time he describes (around 1998), the Chief Architect
Page 8 of 15
WITNO0460100
WITNO00460100
was Alan Ward. As stated above, I have never been a chief architect in relation
to Horizon.
30. It is also important that I make clear that, during the period he refers to, I did
not have a direct role in relation to EPOSS. In his oral evidence, Mr McDonnell
stated that in 1998 my responsibility was specifically for the EPOSS counter
system. This is not correct. I had no responsibility for the EPOSS counter
system (I was just responsible for the Agents). I do not recall that I had anything
to do with EPOSS at that time. I had no responsibility for any concerns arising
about EPOSS nor for responding to such concerns. I do not recall knowing
anything about the formal concerns Mr McDonnell states were raised in
relation to it.
31. In his oral evidence, Mr McDonnell expressed surprise that I was based at
Bracknell. I assume that this is because he is under a misapprehension as to
my role. I was based at Bracknell because that is where the Agent team was
based. However, I probably visited Feltham a couple of times each week to
meet with other senior designers.
RMF
32. RMF stands for Release Management Forum. I never held any formal position
on the RMF and was not a regular attendee. However, I probably received
emails about which issues they were discussing and outputs from the
meetings. I do not know the period during which it operated. My understanding
is that the RMF was an ICL/Fujitsu management forum that examined the
severity of bugs, errors and defects within Legacy Horizon and Horizon Online,
defined the work involved in producing fixes, and scheduled the release of fixes
for those bugs, errors and defects (together with any new functionality) at
regular intervals. I attended the RMF occasionally — either when I wanted to
argue for a particular bug fix to be given a high priority in the scheduling of
fixes being released, or alternatively when I was asked to attend by a member
of the RMF in order to provide technical explanations of bugs, errors and
defects, or to make recommendations as to the prioritisation of certain fixes.
Page 9 of 15
WITNO0460100
WITNO00460100
Relationship between the Post Office and Fujitsu
33. As explained above, I believe that my involvement with POL technical staff
commenced in around 2000. I always felt that our relationship was positive and
productive. We worked collaboratively as a team without much regard to the
distinction between “POL” and “Fujitsu.” Information was shared between us
as and when needed to resolve any particular issue we were working on
together. If there were any political difficulties at more senior levels in the
relationship between Fujitsu and POL, I had no direct knowledge of them and
they did not seem to affect us.
34. I believe that my involvement with POL business staff commenced with Project
Impact in around 2002. Project Impact began with a number of “requirements
workshops”, which POL organised in order to define the specifications of the
changes that they wished to implement. I explain my involvement in Project
Impact in further detail below.
35. As explained above, I also worked with POL technical and business staff during
the design, testing and implementation of Horizon Online.
36. Alongside these development roles, I had a support role which I believe I took
on once Project Impact went live in around 2005; this involved my talking to
POL occasionally about the nature and impact of certain bugs, errors and
defects, as well as the corresponding fixes.
37. The Inquiry is also aware that I had contact with individuals in POL in relation
to prosecutions brought by POL. I have assumed that I am not being asked in
this statement to explain this role given that it will be explored in phase 4 of the
Inquiry.
38. My experience was that ICL/Fujitsu and POL (at least at the level I was working
at) seemed to have a good and professional working relationship.
Page 10 of 15
WITNO0460100
WITNO00460100
39. The Inquiry has asked me to describe the processes by which any “issues”
were raised and shared between the teams, either by POL or ICL/Fujitsu. I am
not entirely clear what is meant by “issues” in this context given the potential
scope of that term and in the context of a very longstanding commercial and
technical relationship. If the question is intended to be specific to how
ICL/Fujitsu shared with POL information about bugs, errors and defects within
Legacy Horizon or Horizon Online, my understanding is that the ICL/Fujitsu
customer services team — in particular the so-called “problem managers’ within
this team — handled these matters. They would be alerted to problems by the
SSC or (for operational issues) by the Ops team in Belfast. I was never a part
of the customer service team. I had no understanding of their interfacing
processes. I was one of a number of sources occasionally called upon to assist
when the customer services team wanted technical explanations of the bugs,
errors and defects, or where it wanted a software engineer to talk to POL about
these bugs, errors and defects, and the corresponding fixes. However, I would
not do any of these things of my own initiative. I would only talk to POL about
these issues if ICL/Fujitsu’s customer services team specifically asked me to
do so, or if POL contacted me directly. My understanding is that ICL/Fujitsu’s
customer services team and POL also asked other software engineers to
speak to POL about issues on which they had particular expertise.
40. My understanding is that POL had to sign off each new release of fixes and
functionality for Legacy Horizon or Horizon Online. These releases occurred at
regular intervals. If that understanding is correct, POL would have received
details about the bugs, errors and defects that ICL/Fujitsu had identified and
fixed.
The IMPACT Programme
41. The Inquiry has asked me to describe the background to the IMPACT
programme, including its main business objectives. I do not recall the exact
date when the changes proposed by Project Impact were first raised. However,
my understanding is that it was an initiative generated by POL in around 2002.
As far as I remember, the main business objective was to replace POL’s back-
Page 11 of 15
42.
43.
44.
WITNO0460100
WITNO00460100
end accounting system. I understood that this would have the effect of reducing
POL’s staff costs.
I do not know precisely how POL created the concepts and designs for Project
Impact. I probably contributed some technical ideas that I thought would assist
in implementing POL’s objectives, but I was not involved in taking the final
decisions about these concepts and designs. POL had a team that analysed
the business and held a number of “requirements workshops” to define the
specifications of the changes that they wished to implement. In order to
understand what POL wanted and what the technical feasibility of
implementing those changes was, I attended many of these workshops. Those
workshops led to POL inviting bids from various suppliers for different parts of
the project. I understood that Fujitsu bid for all parts of the project, but the only
work it actually received concerned changes to Legacy Horizon that would
enable it to interface with the new back-end system.
The Inquiry has asked me to explain the process that led to the decision to
remove the ability of branch staff to post discrepancies to a suspense account
and then roll into a new trading period. I believe that POL took this decision. I
do not know who within POL was involved in the decision-making process or
what factors they took into consideration. I do not know what tests POL carried
out to determine the business effect of implementing this change. The
relationship was not such that I could submit or dictate to POL what changes
in functionality should be made; that was a business decision for POL to make.
Rather, it was my job to understand what POL, as the customer, wanted and
to work out how to achieve those objectives on a technical level.
The Inquiry has asked me whether there were any technical reasons relating
to the changes to POL’s back-end systems that prevented branch staff from
posting discrepancies to a suspense account and then rolling into a new
trading period. I do not recall any. POL took what I assume was a business
decision that this practice should be stopped.
Page 12 of 15
WITNO0460100
WITNO00460100
45. As explained above, I believe that ICL/Fujitsu was only involved in one aspect
of Project Impact, which was implementing changes to Legacy Horizon that
would enable it to interface with the new back-end system. This was
implemented in two phases as part of the S60 release in 2004 and the S80
release in 2005. My involvement was to design the changes which POL
wanted.
46. I remember that these phases of the work required a major re-engineering of
the accounting processes in each POL branch, e.g. removing the cash account
and introducing the branch trading statement.
PEAKs and PinlCLs
47. The Inquiry has asked me whether the PEAK and PinlCL systems as designed
were effective. In my opinion, they were effective, because they allowed
ICL/Fujitsu to understand, prioritise, monitor and remedy bugs, errors and
defects in a timely fashion. However, it is difficult to state definitively that a
system is “effective” when that system necessarily relies on humans inputting
and processing the information in it. Used properly, I believed that it was a
good tool, but only as good as the users handling it. I cannot account for the
effectiveness of every PinICL and PEAK. I should also point out that I cannot
compare it to any other system because I have not worked with any other
system over the past 25 years. I have no point of comparison to offer the
Inquiry.
48. The Inquiry has asked me what pressures there were to resolve or close
issues, or attribute them to user error. I recall that there were periods where
ICL/Fujitsu management would consider the statistics on open PinICLs and
PEAKs, and would encourage us to spend more time investigating and closing
those that could be closed. However, as far as I am aware, we only closed
PEAKs if satisfied that we had investigated the problem. I do not recall any
examples where I was involved in closing PEAKs without investigation and
simply attributing the underlying problem to user error.
Page 13 of 15
WITNO0460100
WITNO00460100
Reflections
49. The Inquiry has asked me whether I think I could have done anything differently,
and whether Fujitsu or POL could have done anything differently. In the context
of the questions I have addressed in this statement, I cannot think of anything
that I could have done differently, but I am happy to respond to specific
suggestions.
50. Again, in the context of the questions I have addressed in this statement, I
cannot think what Fujitsu could have done differently, but I am happy to respond
to specific suggestions.
51. In relation to what POL could have done differently, it is difficult for me to
comment because I have little direct knowledge of exactly what they did and
why. However, my impression is that POL did not provide enough support to
SPMs when they were struggling to use Legacy Horizon and Horizon Online.
Instead of investigating the issues that the SPMs reported (and trying to assist
them), my impression is that POL blamed them instead.
52. The Inquiry has asked me who was to blame for the problems that have arisen
from the Horizon IT project. Again, it is difficult for me to comment because I do
not have the full picture in front of me and I am not sure that there is an easy
answer to this question. The demands on the Legacy Horizon and Horizon
Online systems (in terms of what they had to deliver) were substantial. Their
scale was significant. They were not perfect computer systems. No computer
system is perfect. However, I believed at the time — and I still believe now — that
ICL/Fujitsu did a good job in investigating and fixing bugs, errors and defects.
53. Finally, the Inquiry has asked me whether there is anything I would wish to add
to assist the Chair. In the context of this witness statement and the questions
which I have answered, I do not think that there is more that I can say which
would be of assistance. I would like to say however that I was shocked to read
Page 14 of 15
WITNO0460100
WITNO00460100
the judgments that quashed the wrongful convictions of the SPMs. As I have
stated at the outset of this statement, at this time, and pending the ability of my
lawyers to advise me, I will revisit the question of whether I will rely upon the
privilege against self-incrimination related to those matters connected to POL’s
prosecution of SPMs.
Statement of Truth
I believe the content of this statement to be true.
s I GRO
Dated: 06/02/2023
Page 15 of 15