Hi Christian,
On Thu, 2006-11-02 at 00:47 +0100, Christian Lohmaier wrote:
> But surely the specification needs to be "final" or "stable" or
> whatever you want to call it before the code gets into the master.
Sure - it needs to be in a good state of agreement with the code,
although as we know, currently the specs bit-rot drastically as soon as
we get past this point.
> So adjusting the spec as your likings after throwing random junks of
> code into the tree is not what I have in mind when talking about
> wiki-based specifications.
Nor me :-) clearly "throwing random junks of code" into HEAD is stupid.
> And if you need to change your whole feature multiple times, then you
> ought to thing before. (and again this doesn't relate on how to actually
> code it, but on what the feature is supposed to do)
Anyone that thinks they can sit down and design a perfect system and
then implement it, without some (perhaps substantial) degree of
iterative fixing is [ I think ] deluding themselves. When I worked with
hardware design, it was -unheard-of- for revision A. hardware to work
without modifications. [ the hardware design flow including a ton of
design, simulation, review, etc. ].
There are some great examples of unsatisfactory usability in OO.o that
have specifications ( I guess packed with screenshots to match :-).
There are other great examples of all-out-over-design around the code,
where an iterative approach would have saved both time, money,
complexity etc.
I don't expect to convince you that iterative development is a good
idea for you, it seems you prefer a cathedral approach :-) but at least,
it appears to work out rather well for other people.
All the best,
Michael.
--
[EMAIL PROTECTED] <><, Pseudo Engineer, itinerant idiot
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
