Friday 24 February 2017

The best of intentions - but a bad design


By trade I am a systems developer. I design and build large data applications. A good system can be transformative in both efficiency and culture. A while back I developed a system for handling aircraft repairs to ensure they were completed on time and in order to eliminate paperwork. When you're dealing with safety and quality systems the paperwork soon mounts up.

As the system matured, providing us with a very useful warehouse of data, we began to make further efficiencies. Not long afterwards we saw the advantage in rolling the system out to suppliers who carried out similar work. We would often use another company to manage our peak load work. Due to various constraints we were forced to develop a slightly different system for suppliers but they were essentially the same.

Thankfully we were able to use existing network infrastructure to ensure the systems talked to each other. This worked well over a number of years. Eventually though, corporate governance got involved and identified the data link between the two systems as a security threat. Had we considerably upgraded our network technology we would have been permitted to continue but alas the costs were prohibitive. We were forced to sever the link.

Thankfully we were given some notice and so we were able to come up with a contingency system whereby the systems would communicate via email exchange, sending data in zipped attachments. It was far from perfect but since all the other options were closed down we were forced to make the best of a bad job.

The application itself was fairly comprehensive, being capable of producing airline certification documentation, but these features depended heavily on having a live link into the parent company. The system had been developed with certain assumptions in mind - one of which being that there would be a continual live link. As soon as this was severed, only the basic functionality worked and years of development were rendered useless.

In some areas of the business this caused a reversion to previous systems or simply adding more work for engineers who were forced to use Word templates and call in by phone for an official document number. It was a huge disappointment and a major step backwards.

The corporate view though was that the system was only reliable as long as I was in place to manage it. Throughout the question was asked "what happens if you get hit by a bus?". Being slightly accident prone I assured them that I have in fact already been hit by a bus, twice in fact, and a third time was statistically unlikely. This was insufficient reassurance.

If honest, I can see their point. Had they given the system more of an official standing it could have been supported but there was no budget for it. Worse still, it would have given me total leverage and I could, if so inclined, have bled them dry - as indeed do most software vendors. In terms of functionality my system was a net benefit but in the risk margins it was considered a liability.

In that regard, if that was their view they should never have allowed it to be developed in the first place. I'm sure they wouldn't had we told company headquarters we were building it. However, for the time it was running, while the going was good, we had a useful and reliable system that kept aircraft in the sky.

And this lesson is pertinent to Brexit. If you are going to build a system on which so much depends, seek the consent of those it affects, ensure you have a contingency plan and don't put all of your eggs in one basket.

As much as the EU was never built with proper consent, with much decided in secret, it is a system vulnerable to the whims of voters, whose reasoning can be often as spurious as that of any bureaucracy. They very often don't see the merit in what you are seeking to achieve while at the same time enjoying the benefits of it.

My mistake in developing an ever more intricate and capable system was the assumption the same consent would be a constant and that once the merits were there for all to see, consent would follow. Neither I nor the EU ever sought consent, overreached in what we are trying to do, and ultimately, without a buy-in from the people it affects, it could not last - however good it was.

I have since learned that the survivability of a system is its adaptability. One should keep it narrow in scope and limit the number of dependencies. Had the EU considered this we would not now be facing this mess. It was never developed with democracy in mind.

No comments:

Post a Comment