<x-flowed>Most people I know tend to write code first before they even know for
sure what the whole program is going to be. I do it that way myself.
The program grows in fits and starts. Design decisions are put off in
favor of coding the really fun parts of the program. The problem gets
worse if there are multiple people working on the same project. How
is my little prize piece of coding going to mesh with your code?
If there is going to be a "Grand New Firewall Paradigm", then there
will need to be some major prep work to be done prior to the coding.
What it is going to look like. What to implement now and what to
implement in the future. How to make it extensible. What will the
objects look like. Primarily this means there should be a spec.
Recently I found Joel Spolsky's web site and have been reading it.
There are a couple of specific articles from his site that I
recommend for reading that are listed below. For background, Joel was
the spec writer for Excel 5.0, which he said went to around 500
pages. He has experience and he writes well.
Functional Specifications Made Easy
4 chapters
http://joel.editthispage.com/stories/storyReader$212
User Interface Design for Programmers
9 chapters
http://joel.editthispage.com/stories/storyReader$51
His archives for further reading:
http://joel.editthispage.com/stories/storyReader$94
_______________________________________________
Leaf-devel mailing list
[EMAIL PROTECTED]
http://lists.sourceforge.net/mailman/listinfo/leaf-devel
</x-flowed>
