Lists Home |
Date Index |
On Fri, 2004-12-31 at 17:20 +0000, Bill de hÃ“ra wrote:
> I agreed with you up to here. No doubt there are terrible programmers
> out there, but poor engineering is usually a symptom of something else.
> I'd speculate that most poor software is not down to poor engineering
> but because software projects are rarely predicated on the amount of
> work to be done.
Which may be down to a flat statement from some
project management type that 'you've got n man days to finish this'
| 'we must bid less than X man days to get the project'
| 'The Proj manager thinks your % of the work is Y man days'
i.e. the program estimation is based on available finance rather
than software estimate from the people who are going to do the work...
Not that they are in a position to give those estimates when its
a project that has a zero length code base at the time of the estimate.
That's when the mythical man month comes into its own.
Guess it'll be 163 man days. Multiply by 3, add 10 percent, then
open your mouth to the project engineer, who will immediately
cut it back by ZZ percent, smiling all the while and assuring you
he's looking after your interests.
Add QA, iso9000 level documentation into the mix, peer group
walkthrough's and system build documentation, and the multiplication
factor shoots through the roof.
Then go program it in some language without a stack, mathematically
proven as being traceable through any program path.... and DoD/MoD might
just buy it, at some small percentage of the cost.
Perhaps I'm just jaundiced.
Either way, happy new year xml-dev'ers
XSLT + Docbook FAQ