The order of development
Tue, 21 Jul 1998 17:14:50 -0700 (PDT)
On Sun, 19 Jul 1998, David Manifold wrote:
> I think I've changed my mind...
> Writing a spec is like trying to write the document before the program.
And I've seen commercial programs that have done that. It's sad.
You usually end up fixing yourself into a bad track which seemed good in
the beginning, or you end up straying from the documentation making it
worthless in the first place...
> What if the program idea changes during development? So, I think I'll
> dive right in and start making something. Actually, waiting for a done
> spec before starting implementation is a concession to expediency. If a
> spec is already written, the product can be implemented immediately
> without any false starts or rewriting.
Writing is rewriting. Book authors have said that for hundreds of
years, I'm sure. We're going to end up rewriting Tunes several times
no matter what course we take. Might as well gain practical knowledge
and experience now...
> Of course, it probably takes
> longer to write the spec without having any idea what you're specifying,
> so I'm going to take some false starts as a Good Thing(tm).
Tunes has always been a proof-of-concept OS. Might as well get around to
proving some of those concepts, right?
> It's kind of like this project never even started. Maybe if I have some
> lousy code that just gives an idea of what we're looking for... at the
> very least there will be a topic for discussion on this list.
I've agreed with that for a long time...
> Flames and comments appreciated.
> Happy Tunesing!
> David Manifold <email@example.com>