[Tutor] facing a new project

[alan.gauld@bt.com]

> A rough design and working code sounds like a good approach.  
> According to Richard Gabriel, it's more important to get 
> something out that sorta works and then gradually fix that

This guy obviously never worked as a maintenance programmer.

Seriously, I spend a lot of my time fixing systems where 
not enough thought went into design - sometimes none at all!

Its still true that 80% of the cost of a program comes 
after the first release - bug fixes and enhancements. 
Lack of a design makes both much more difficult.

BUT I would say its worth doing some prototyping to
get the ideas of how it should work clear in your mind, 
then throw it all away and re-do it properly.

The cost of recoding is usually recovered several 
times over downstream.

The only exception to this rule is programs with 
known short lifespans. If you know you will only be 
using it once or for a few months then just hack
it together whatever way works.

But for this specific case you need to seriously 
consider how to data drive the scheduling algorithms, 
the people profiles etc. A bit of up front design 
will help.

Alan g.