On Thursday 25 October 2001 18:09, Rafiq Ismail wrote:
> On Thu, 25 Oct 2001, Nathan Torkington wrote:
> > I can have p5ee.perl.org created, and set up people to edit the
> > content, but I want to see some progress before p5ee gets 'blessed'
> > like that.  Similarly, I'll make a CVS repository when we have code
> > to version :-)
>
> Like wise, code shouldn't be written until it has been clearly spec'd.
> Like wise, spec's shouldn't be written until we know our requirements.
> Like wise, requirements's won't be known until we understand the language
> and nature of the model we're modling.

I think we can try to be more iterative than that. I wouldn't mind writing 
some code (with the obvious caveat that it has strong chances of not being 
final at all) in order to flatten out the rough edges of the specs and to 
isolate clear requirements. Spec'ing without some reference implementation 
and a clear idea of what is a good/bad idea in actual code imho often leads 
to problems (and I'm not even into XP ;-).

-- 
_______________________________________________________________________
Robin Berjon <[EMAIL PROTECTED]> -- CTO
k n o w s c a p e : // venture knowledge agency www.knowscape.com
-----------------------------------------------------------------------
Heisenberg may have been here.

Reply via email to