There is a certain absurdity about this current trial on Horizon errors. By Post Office Ltd (POL) own admission the Horizon system is no longer fit for purpose and it should be replaced. I am sure they will do so in due course, but when that happens, by everybody’s own admission surely, it will introduce a raft of new errors as happens when any new software system is released. If POL pretend otherwise and proclaim the new system as being robust and reliable just as they did in 2000 with the current system (and again in 2010 with HOL) then we are back at square one. Presumably though POL will take more care when dealing with the new errors than they did in the past. Well based on what we are hearing in this trial the safest assumption is that they will not as they are fighting their past mistakes tooth and nail and one can only surmise that they truly believe the methods they have in place at the moment are the best way forward. Of course they will also be relying on their present employees to support the new system – the very people in court right now as witnesses testifying to the mistakes they have made.
One thing that I thought had been extremely reliable over the years of Horizon was the counter computer equipment that staff used. The keyboard, the touch screens, printers and computer itself. Yes there were breakdowns and replacements needed but by and large they were quite reliable and they lasted well past their sell by date. But they were getting extremely old. The workstation computer was at best a 486 processor I believe – ancient and slow in today’s world. So recently POL decided to replace them all with the latest technology and decided they would plump for the cheapest on the market. Yet another example of incompetence and one that has truly backfired on them. Horizon, the not fit for purpose system, cannot handle the new equipment and the errors that this new equipment generate.
I am not going to detail every type of error that the new equipment has introduced but concentrate on one that I was reminded of yesterday in the trial. Up until then I had no idea about problems with the old touch screen monitor causing phantom transactions but I can understand how these errors may have arisen. Actually an important point – 10 years owning a post office and I was never told about the possibility of this error occurring. Anyway, hearing of this error reminded me of Seema Misra’s trial and how her expert IT witness Prof MacLachlan had extensively criticised the keyboard layout AND the touch screen layout but he never gave any thought to the possibility of problems with the screen hardware itself. This error would have been in the system from the start until that hardware was replaced, It would be impossible to trace in log files. Mmmm.
The discussion in the trial yesterday also talked about the 00 key and the mistakes that that has enticed out of SPMRS over the years, just as Prof McLachlan pointed out in 2010 and that key and the problems it makes for subpostmasters is still there.
I am no one man call centre but I do have access to social media platforms where thousands of SPMRs meet and discuss PO issues so when the rollout of the new counter terminals started across the network I was able to note reports of strange errors starting to happen. One of these was and as far as I know remains, the most prevalent. The built in swipe card reader causes lots of problems but the generic effect and most likely the cause are the same. A user swipes a mag stripe card and, occasionally, instead of the system accepting the data on the card as it should, whether that be an account number or a name or in some cases a balance, the system instead interprets it as a sequence of key strokes entered into the keyboard by the user and this causes a wide variety of different outcomes such as purchases of stamps, cash deposits etc in fact any event that can be entered via the keyboard during normal operation. This effect would also happen extraordinarily fast so all the user can see is the contents of the screen once the computer has processed the string of characters from the mag stripe card. The reason I know this is the cause is because several times users have reported the final screen as showing a string of characters in a field on the screen such as an address for a postage item that could only have come from the card because of the format. Not only that but some of that data is confidential and can now be clearly seen by a third party.
It seems to me that this error is down to the cheapness of the equipment that was bought. I am no expert in programming for mag stripe card readers but I do know that they have a start and end byte/character that should be captured, processed and checked to ensure that the data the system has received in between the start and end characters is that which is recorded on the mag stripe. I think that this is generally done by firmware on the card reader itself and error codes are returned to the calling program should the data not be verified after the card is swiped. Whatever the reason, whether Lenovo do not include this industry standard feature on its card readers, or Horizon itself does not interpret the data received properly, it is a mess and clearly this computer equipment was not tested thoroughly before it was purchased for no doubt a very significant sum.
Just another item to add to the long list of incompetency displayed by the current management of POL who continue to think they are the ones who can fix the problems they created. New blood is desperately needed in this organisation so I hope Tim Parker chooses Ms Vennells replacement wisely.