> It's a curious thing, but a lot of software development organizations
> start out wanting to build high rises, but approach the problem as if
> they were knocking out a dog house.
partly i think it comes from the perception that code is cheap (no materials
only time and a good coder can write it pretty quickly) and easy to test
(near instant compile and run). Why spend time designing code when you could
spend that same time actually writing it.

And some certainly are guilty of overplanning, writing plans that are
basically an alternate representation of every peice of code is rather
wastefull. In this area i think there is a similar issue to with commenting,
the schools teach people they should plan thier programs and comment thier
code but the programs simply aren't big enough to require the use of
planning or comments.

Equally though as soon as more than one person is involved and sometimes
before that good planning of the boundry interfaces, discussion of exactly
what is needed from what part of the sytem and descisions about the timing
(who has the CPU when and associated management interfaces is a major issue
when desiging single threaded systems) are required.

ultimately it's getting the balance between time spent on design and
time/bugs saved by design right that is so difficult.

another issue seems to be lack of oversight. With any building (except
smallish temporary structures) you are going to have to get your plans past
planning and building control (for non-uk readers planning is concerned with
looks and other affects on the local area, building control is about making
sure the building is safe, energy efficiant, disabled accessible etc). On
the other hand for software (with few exceptions) the only person making
demands of software is the client. Until clients demand and pay for quality
software engineering clients won't get quality software.


-- 
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist