Original URL: https://www.theregister.com/2009/04/03/computer_says_no/

How gov scapegoats systems for man-made errors

Dead pupil letter shows it is human to err

By Jane Fae

Posted in Legal, 3rd April 2009 08:02 GMT

If you want to understand what is wrong with public policy when it comes to IT in the UK, look no further than the recent tragic case of the letter sent by a school to the parents of dead schoolgirl Megan Gillan, demanding that she improve her attendance.

It was one of those bleak and bitter accidents that inevitably occur from time to time, leaving sysadmins everywhere breathing a sigh of relief that on this occasion it was nothing to do with them.

Total error-free operation is not a realistic goal. Bugs do get built into systems, and no matter how perfect the system, they still need people to operate them - fallible people.

What could be avoided is the misleading aftermath: the spokespeople for official bodies running round and blaming the system, the reinforcement of the public myth that computers are somehow alien, given to working in mysterious ways quite beyond the ken of the average punter.

When we first read this story – a single paragraph in one of the broadsheets – the event was described as a "system error". Our instant reaction was that this was unlikely; most system errors usually turn out to have a very human origin.

However, the same story was elaborated upon in the BBC report of the event. Cheshire County Council confirmed that Megan’s school had been using Capita’s School Information Management Systems software (SIMS) to maintain her details. A spokeswoman for the Council was reported as saying: "Megan's name had been taken off the school roll when she died, and removed from the main school database.

"However, unknown to the school, her details had remained in a different part of the computer system and were called up when the school did a mail merge letter to the parents of all Year 11 students about their prom.

"The letter called up details of each student's attendance for the whole year to date and because Megan had been on roll in September, she was included."

So the system, for some peculiar reason, holds multiple representations of the same data? If true, that would be an accident waiting to happen.

Except, according to Capita, that is not so. After wading through their initial slightly woolly response, which committed them to a software change that would make it impossible to send attendance letters out to pupils who have left a particular school, we put it to them that the systems architecture implied by the above statement would make their system unfit for purpose.

They took the bait and politely, but firmly, explained that our conjecture was wrong. The system only contains a single data table for pupils' names and addresses. The issue was pretty much as outlined above.

We are still not totally convinced of the need for a software fix. We have since spoken to the Head of ICT in a school that runs SIMS, who was equally scathing of the idea that this was a "system error", pointing out that the software allows schools to set a deceased flag, after which, as far as he is aware, all communications in respect of a given pupil are automatically blocked.

If true, the danger of putting in a "fix" of the kind that Capita have proposed is that it adds to the complexity of the system, and increases ever so slightly the possibility of a real bug being introduced. It is cosmetic rather than necessary.

So why is a County Council putting out a statement that is at best misleading, and at worst calculated to divert attention from an individual failure or even a more systemic failure by the Education Authority to ensure that users of the SIMS system are properly trained?

Their initial reaction was that such speculation was unfair. They have now passed on a statement from the Head Teacher of the school in question, who is adamant that the correct procedures - including flagging the pupil as deceased - were followed. This does not quite explain why their first statement made reference to a second database.

Does this matter? We think it does, because as long as the general public are encouraged not to understand the difference between system and human error, it is impossible to have a sensible debate about the use of IT in the UK. On the one hand are the eternal optimists in government, focused wholly on the benefits that properly functioning IT can bring; on the other are the naysayers, who do not quite understand the issues but have been educated to believe that many of life’s ills can be put down to "computer error".

That in turn makes it hard to spot when a system is badly structured. Earlier this year, we were told of failings in a package called "Truancy Call" that sends out automated calls to the parents of absent school pupils. One problem with this package was that it did not appear to have properly sorted out its rules for calling where parents were separated or divorced, which led to it generating inappropriate and almost certainly DPA non-compliant calls to non-resident parents.

That is a "system error", normalising bad practice, and inculcating some very bad data habits indeed into those who believe that if the computer says so, it must be right. Stephen Clarke, Managing Director of Truancy Call Ltd, commented: "Truancy Call extracts data live and in real-time from schools' management information systems which schools are responsible for updating." In other words, if there is an issue here, it is to do with the way the schools hold data - not Truancy Call's interpretation of it.

The issue in both cases is that the explanation is neither simple nor straightforward. The Macclesfield error was reported as system error, looks more like human error, and is finally claimed as something subtle and systemic. The Truancy Call problem appears to be about the unintended consequences of software interacting with someone else's database. In neither case is understanding aided by sweeping general claims from officialdom. ®