Specifically, I want to introduce regression testing in physical design work, and have a worldwide list of test cases across all companies making a given thing.

(Warning: lengthy rant ahead.)

( the office coffee jar has a design flaw )

( and the office itself has a design flaw )

( watches and clocks have design flaws too )

( and one of the office back doors used to be downright dangerous )

There's something deeply disturbing in the above list. In all these cases, there's a design flaw – sometimes quite serious – in an everyday item that as a species we have been building for decades (digital alarm clocks) if not millennia (buildings). How can we still be making this kind of fundamental mistake, on what appears to be a regular basis?

I'm not blaming any of these designers, per se, for failing to anticipate a particular design flaw. Each of the flaws I've described above is something I think one could reasonably be forgiven for not having foreseen, the first time. It's only because I can't believe it was the first time in every case that I get angry about it.

My guess is that there are two factors in operation. One is that there's no central Design Authority which works out the best possible shape for a coffee jar or an alarm clock; instead, each individual company does its own design independently. This is great from some points of view – it means innovative new designs can prove their worth without having to convince a committee of closed-minded theoreticians – but one area in which it falls down is fault-fixing, because companies don't appear to routinely go around learning from one another's mistakes.

The other factor is that companies don't internally seem to be good at keeping track of the reasons for their design decisions. Casio had been making watches with flat display faces for years and years and years. I suppose it's possible that they had previously never even considered anything else, that at some point they discovered how to make curved faces cheaply, tried it, found the watches all scratched, and returned to Plan A, and that I happened to be a victim of that ill-fated venture; but it just doesn't seem likely to me. Far more likely, I think, that someone actually put in some thought at one stage and decided that flat faces were a good idea, and later on some overzealous designer thought it would be fun to move to curved faces, and somehow the good common sense of the first guy didn't communicate itself to the second.

So it would be a good start to have companies who design things just do regression testing, even if only as a thought experiment. Every time someone fixes a design flaw, they should add the flaw and the fix to a list somewhere. ‘1988-04-05: Curved display faces can scratch in the middle. Reverted to using flat ones universally.’ Then arrange that it's a serious gaffe for any designer to make a gratuitous change in a product without first checking that list and making sure they haven't reintroduced a problem that someone had carefully fixed fifteen years ago.

Better than that, though, would be to have some kind of a cross-company repository of good common sense: things to take into account when designing buildings, doors, clocks, coffee jars, whatever. I realise that this is the seriously unrealistic bit, since it requires competing companies to cooperate, but at the very least there should be some mechanism whereby a brand new building firm, starting up in business for the first time and landing a large contract for a great big office block, can somehow benefit from the experience of all the office building that has happened in the previous fifty years.

I wonder if there's any way that could be arranged to happen.