why it's Important That Software Fails

[adam]

http://www.berglas.org/Articles/Importa ... Fails.html

[teamonkey]

No problem there. I can write software that fails in my sleep. I'm an asset to productivity and the economy.

[piku]

Making software's hard, but not for the obvious reasons. Having the skills to actually write the stuff, or design it correctly isn't the difficult bit. The really hard bit is understanding what your client is going on about.

I used to write bespoke business software, and our clients often knew what they wanted the software to do (usually "turn this mess of paper into something computerised") but had no idea how to tell us that in simple ways. We often found they had really inefficient processes for handling data that had evolved over time... and didn't want to change it, they just wanted to computerise it.

However that's not the really hard bit, because understanding your client is part of your job.

The hard bit is understanding the business itself and all its subtleties and technical details. You can't be this clinical software developer that impartially observes things and somehow creates a bit of software. You've got to talk to the employees and learn how to do their jobs. Without doing that you can't make sense of what they mean. "Well the Ali-angle 310mm sometimes comes from the foundry as ali-plate and we shape it, then rebook it in as the new product, but keep the old product on file too because if it's cold rolled obviously you can't reshape that or it'll snap so we blah blah".

[isomies]

Interesting article and certainly borne out by my experience (15 years in LOB applications development).

Increased sophistication in system functionality leads very often to more bureaucracy rather than improved efficiency.