FUJ00122218 - Witness statement of Gareth Idris Jenkins

Evidence on official site

FUJ00122218
FUJ00122218

Witness Statement

(CJ Act 1967, s9; MC Act 1980, ss 5A(3)(a)
and 5B, MC Rules 1981, r 70)

Statement of Gareth Idris Jenkins

Age if under 18 Over 18 (If over 18 insert 'over 18")

This statement (consisting of 3pages each signed by me) is true to the best of my knowledge and
belief and I make it knowing that, if it is tendered in evidence, I shall be liable to prosecution if I
have wilfully stated in it anything which I know to be false or do not believe true.

Dated 24th dayof March 2006
the

Signature

I have been employed by Fujitsu Services, working on the Post Office Account, formally ICL
Pathway Ltd., since 1996 as a Customer Solutions Architect, involved in many aspects of design
and implementation of the computer system known as Horizon, which is a computerised

accounting system used by Post Office Ltd.

Audit Record Queries’s (ARQ’s) 401, 459 and 460 requested information in relation to Gaerwen
Post Office (Branch office code! 60604). I was asked to produce information relating to ‘Nil’
transactions during the periods specified. I have provided three spreadsheets which I now
produce as Exhibits GIJ/01, GIJ/02 and G1J/03. I’m not sure about this. I’ve had nothing to do
with producing the spreadsheets. All I’ve done is made some statements based on what is in the
spreadsheets. I assume that Neneh or Penny produced the spreadsheets, but I have no personal
knowledge as to what was included within them and what was excluded. For all I know, you

could have typed them up from scratch

The reports are formatted with the following headings:

For ARQ 401

Signature Signature witnessed by

CSO11A (Side A) Version 3.0 11/02
FUJ00122218
FUJ00122218

Witness Statement
(CJ Act 1967, s9; MC Act 1980, ss 5A(3)(a) and 5B, MC Rules 1981, r 70)

Continuation of statement of Gareth Idris Jenkins

Nodeld — The Counter Identifier at which the transaction took place

Userld — The User Identifier of the user who was logged on when the transaction was carried out
Date — Date of transaction

TranType Decscription — The type of Transaction carried out together with a description of its
meaning

Amount — The amount of money that actually changed hands

HTxnNum — The Unique identifier for a banking transaction

RespCd Description — Response Code generated for a particular transaction together with a
description of its meaning

For ARQ 459 — Some of the columns as listed in 401 plus the following below

Id — Same as Counter Id above

Time — The time of transaction.

Application — Initiating Application (usually Banking)

RtngGwy (Routing Gateway) - which is the identifier of the route to the financial institute
where the transaction is going

TranType — Same as TranType Description above

PAN (Primary Account Number) — Account number on the card

SaleValue - Same as Amount Above

RespCd — same as RespCd Description above

For ARQ 460 - Some of the columns as listed in 401 and 459 plus the following below

Counter Position - Same as Counter Id above

Should the spreadsheets be regenerated so that they are all in the same format and so that the
correct text comes out for RespCd 23? I can’t explain why the formats are different (I assume it
is changes to the tools used to generate them). There are three reasons why a zero value
transaction may be generated as part of the banking system:
The transaction has no financial effect (ie a Balance Enquiry or a PIN Change)
The transaction has been declined by the Bank

(This is a really poor choice of words which seems to accept that failures in the system are
normal and therefore may well support the postmasters claim that the system is to blame for the
losses !!!!) Please can you suggest something better then? What we have here are genuine
failures of the end to end system which are not part of normal operation, but are anticipated and
the system is designed to cope with them. Some such failures could be engineered as part of a
malicious attack (but that doesn’t apply to those failures that appear in the evidence presented).
In all cases the system is designed to identify such failures and handle them in such a way that
the Customer, the Postmaster, Post Office Ltd and the Fls are all clear as to the status of the
transaction and any necessary financial reconciliation takes place. I guess one option is to delete

Signature Signature witnessed by

CSOt1A Version 3.0 11/02
FUJ00122218
FUJ00122218

Witness Statement
(CJ Act 1967, s9; MC Act 1980, ss 5A(3)(a) and 5B, MC Rules 1981, r 70)

Continuation of statement of Gareth Idris Jenkins

the paragraph since it is purely an introduction to the following more detailed description.

Each transaction has associated with it a Response Code field, which identifies what has
happened. These values are included (together with their descriptions) in the spreadsheets
produced. In summary:

RespCd = 1, means that the transaction was Authorised by the Bank. If it is for zero value,
then this means it had no Financial effect. In all the example, such transactions were PIN
Changes or Balance Enquiries where this is indeed the case.

RespCd has a value between 2 and 10 means that the Financial Instituation declined the
transaction. The actual value indicates why (eg 3 means that the PIN was invalid and 4 means
Insufficient funds)

RespCd with a value greater than 10 implies some sort of system failure. The actual value
provides further information as to the nature of the failure within the overall system

Within exhibits GIJ/1, 2 and 3, the following Response Code values explain the reasons for the
‘Nil’ transactions :

1 - Authorised OK

3 - Declined - Invalid PIN

4 - Declined - Insufficient Funds

6 - Declined - Usage (amount)

8 - Declined - Other

23 - Failed by FI: Unexpected Response Code (Note this appears as Timeout in the spreadsheet)
26 - Failed by FI: System Unavailable - Return Card

Note that the text varies in the ARQs, but the meaning is basically the same. (what does this
mean, does it needs to be in the statement ?) I’m just repeating the point about the ARQs being
in an inconsistent format. If anyone looks, they will spot this, so I thought I’d better say so.
However I’m happy to delete the sentence if it will cause more problems.

Signature Signature witnessed by

CSOt1A Version 3.0 11/02