On 2/13/06, Ben Collins-Sussman <sussman@red-bean.com> wrote:
> This sounds like a typical conflict beween "top down" and "bottom up"
> approaches to writing software. Do you spend weeks in theoretical
> discussion and design documents, then code at the very end? Or do you
> hit the ground coding, dissect the results, code again, dissect, etc.
> until you arrive at a design?
>
> While you're certainly welcome to start coding mockups of merge or
> history-tracking behaviors, I think what Branko's trying to say is
> that historically, this project has always been dominated more by the
> "top down" sorts of people. We much prefer to spend two months
> discussing theoretical frameworks before we write any code at all. I
> think you're likely to get more participation and general engagement
> if you take that route.
FWIW, I would have zero problem with someone prototyping this sort of
thing, but I think doing so outside of the actual subversion codebase
is kind of silly. It seems likely that you'd end up with a prototype
that can't possibly be implemented in our actual system because it
depends on the capabilities of an SQL database, something that neither
of our back ends actually have.
On the other hand, I do agree that it would probably be beneficial to
have some concrete discussion of requirements going on before that
happens.
-garrett
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Mon Feb 13 19:13:59 2006