It's tried and tested before leaving the factory. Every bug that was traced has been programmed out and, according to Quality and Assurance, it's ready to roll. It's installed and for the next six months (minimum) there are huge letters sprawled across the front....

"System Under Test"

How many times have you seen this rather infuriating notice on a piece of equipment!? UK railway stations seem to take delight in displaying what is to you highly critical travel information followed by "... but please don't believe this, we are still testing". Yes, testing they are, especially one's patience!

But what is the 'system' they are testing. They would like you to believe it's the hardware and software that physically makes up the product - hey!, even those who installed the bits and pieces think this - but it runs far deeper.

Yes, there is the argument that if anything will fail, it is likely to be in the first few weeks of service - the old "bathtub curve" argument. Now, long baths are really nice but isn't 6 months stretching things a bit!? Ah-ha! It's got nothing to do with baths. It's all to do with "health and safety". The companies involved with writing the software, building the hardware, and installing the system all don't want to be sued should, as in our example, they send you on the wrong train, plane, etc. and the result of this happens to be rather dramatic. Again, is 6 months a realistic time to test. Heck! there's been a timetable change in between.

So, we've proved it's got nothing to do with failures, or health and safety. Any other argument along these lines will also come to just as an abrupt end. So what is the "system" that is under test? Why does it take such exceedingly long times before a system is truly proven?

The system being tested is engineering principles as a whole. Everyone is so intent on using the correct (or is that the latest?) methodology that the real engineering practices are being lost along the way. "Do our software methods comply with the latest UML levels?", "Do our records hold up to our ISO9001 rating?", "Does our project bear any relation to the accepted model?" all over-shadow the ultimately important question.....

"Will the flaming thing work?"

Can any amount of engineering principles, methodologies, or models answer this most basic question? No. Good old fashioned basic engineering practices, however, easily answer this question. If methodologies are that good why are they being reviewed as often as they are? Why have there been so many adjustments? Best of all, why have so many companies now abolished them? Simple. They cost time and money without offering one shred of guarantee the product will even make it to the market, never mind survive once out there.

Imagine, horror of horrors, the engineering fraternity suddenly realize that the latest methodology had some serious flaws... and the product was about to roll. Imagine the costs of having to adopt new policies toward the end of a project - and there is many a project leader who can relay stories of having to "back-track" on a project just short of the end.

If it is not irritating enough to have the wrong information fed to you, how about the thorn in our sides called Swanwick. If there was ever an advert to not employ such methods then this one rates as the best of the bunch. Remember, it was over 5 years late and many millions over budget. The cute clipping of a couple of aircraft controllers biding their time with a game of ping-pong became irresistable. Sadly this is about as far as the joke goes as I have personally experienced two of the many-hour delays of which there have been four in as many months!

What is not being spoken of here is test methodologies. That is a whole different ball-game and one that is seriously being neglected in favour of the design ones (why do you think the newly installed hardware keeps failing!). If the modern design-manufacturing concern realises that it is imperative that the product roll out in a 100% working state in the shortest amount of time, then they will win the battle.

So, the next time you see this annoying little statement splashed across the latest hi-tech information system, or are stranded in an airport for many hours longing for home, just think... have we really progressed? Oh, you too are thinking that if you had written software like that you'd have been fired!

Marc Dekenah.

"Chattering Classes", Sunday Telegraph, Pg 34, 19 May 2002.

© Marc's Technical Pages 11.07.01 / 19.05.02