1 comment

[ 3.3 ms ] story [ 15.4 ms ] thread
Summary: Software has no glitches, they are just there (and users are accustomed to them) because the process and the organisation around the software is not sane.

I share the pessimism - I've seen such things even in the smaller scale.

CS grads come to the interview (for a thesis) and say, well I'll design the system and write a spec, but I won't touch code (meaning I don't know what, maybe 'I'm no idiot' or the like).

I saw a PhD design a knowledge representation system and asking me (no PhD) where is the button in eclipse to run the code - ten lines of an API example from the docs (give me teh codez, eh?).

And I just learned that the official software development model for Germany's federal institutions is a modified V-model: I counted around 30-40 steps (excluding the artifacts) and just one read: "Coding the system". I laughed from the heart. Now imagine replacing this step with "Go for a walk" and the whole project, well, what would it be? Failed? Incomplete? I don't know.

In short: Organisations, as Nietzsche put it, are much more childish and immature as individuals. (Corporate) software development as a relatively new engineering discipline will have long years of struggle to come. Keep your heads up (and sane)!