Too Many Secrets...

Further to yesterday's post about Gareth Jenkins testimony to the Post Office Scandal Enquiry, I stand on my thesis that he was put up by Fujitsu to take the flak that was headed their way in the light of the fact that their Horizon software systems for the Post Office created problems over an extended period of time, as is now a matter of record and widely known. It would appear as I hinted at yesterday, that the message-passing back down the line from PO Ltd. to Fujitsu regarding software issues was partial and restricted at best. That there was corporate/government subterfuge in the attempted cover-up of issues with the system is now also a given. What needs to be drawn out is the extent and scope of the conspiracy to pervert the course of justice, which now appears to lie at board level within both companies, and within government itself.

There have been hints of some deeply unpleasant coercion of this week's witness to comply with the instructions of his 'superiors' in supporting their actions to cover up and deflect attention from the wholly-flawed Horizon systems. Although the two systems are completely distinct from one another at an underlying software level, the issue still remains that both were foisted onto the Post Office estate at an early stage of development, rather than being 'sandboxed' until the majority of obvious flaws could be caught and treated. In the original 'Legacy' Horizon's case, the foundation upon which the 'solution' was  based was simply the wrong tool for the job in the first place: what was required was an enterprise level database management system.

What they used was effectively a flat-file messaging platform which was never designed for anything like the purpose that Fujitsu proposed would be the solution for the Post Office's needs. As the difference in licensing costs between the platform they chose - Riposte - and a mainstream enterprise database system such as Oracle were huge at the time, it would appear that at some level of management within the company, the decision was made to go for cheap rather than standard and reliable. The difference between the two can be summed up as the difference between a properly coordinated and rule-based filing system conforming to double-entry accounting rules, and a large cardboard box full of random receipts. Legacy Horizon was the latter.

The problem was that all the data that was loosely collected in this way from Post Office branches, had to be scraped into a proper database at the Post Office end and translated into spreadsheet reports where necessary. This had to be done by using what is colloquially known in the software game as 'glue' code: pieces of software written to connect two dissimilar systems in order to transmit data from one to the other whilst preserving the sense of those data. The problem appears to have been at least twofold: the quality of the initial data entered onto the system at branch level - the Legacy Horizon system had no data format checking as such: a sub-postmaster/mistress could make a typo which would simply be stored in the system as was, due to the fact of its data agnosticism/ignorance - which would screw up the overall 'picture' of the branch accounting; and inept programming of the 'glue' code that was 'designed' to parse out the data for inclusion in the back-end database further compounded the problem.

This meant having a monitoring system/team dedicated to trapping errors that should been caught and flagged at the code level automatically [as in any properly-constructed software system]. These were all down to poor base platform choice and some coding of extremely dubious quality. However, the flagging up of issues was left to the Post Office side of the equation, with sub-postmasters having to contact the PO's technical 'help-desk' in order to flag up any problems. This appears to have been pretty much a dead end for them, the blame for their accounting issues for the most part laid at their feet alone, with the first port of call for PO Ltd. being accusations of theft, false accounting and at the very least, mis-operation of the system, most often leading to sub-postmasters at best stumping up money they didn't owe to cover losses generated by the system, or worse, being prosecuted and fined or imprisoned for offences they never committed.

As to the Horizon Online system that replaced the original, which 'appears' to now be heading towards being in some way robust in nature, the fact remains that Fujitsu/PO dropped what was effectively beta software straight into a live, commercial environment, that threw up exactly similar problems for their hapless employees and sub-contracted postmasters to deal with without any real help or support and under the same distraint as before, with further false prosecutions forthcoming as a direct result of upper-tier management decisions to ignore the problems they instigated through their enitirely venal management/commercial decisions as to how to operate their companies. Add the nebulous, historical government connections into the mix, and you've got a recipe for disaster, at least for the 'little' people at the sharp end of it all.

How much information about these issues was actually transmitted to Fujitsu and how much of it the PO itself forwarded to the sub-postmasters is a central question that the Enquiry is seeking to tease out. But underlying all of this, of course, is the politics of business, as I wrote last night. The roots of this affair lie at a very high level between the boards of Fujitsu's UK operation & Post Office Ltd., and the UK government at the head of its Civil Service, in an attempt to paint a very rosy picture of a business for sale: and therein lies the crux and substance of all of this; and today, Jenkins used the word 'trap' and the phrase '... [the position] that they put me in...' in answer to questions from KC's Beer and Steen, who both knew, I think, as does the Enquiry I would suggest knows, that Jenkins was under some considerable duress from a far higher power than his line-manager to have done as he was instructed. We're far from done with this one, and my surmise is that Jenkins will be, in the final analysis, considered in a similar light as the hundreds of other victims of this travesty of business and governance. Still no mention of The Official Secrets Act, though...