[svlug] Wish there was a program that..

[Dan Martinez]

Scott Hess wrote:

> The problem is that people do things backwards. They _start_ an
> open-source project, _then_ they start writing code. The better
> approach is to start writing code, get to the point where you've
> actually accomplished something, _then_ make it into an open-source
> project. This is effectively how all successful open-source projects
> were originally created.

To amplify this point a bit: the specific problem Scott refers to is
that a lot of people will write the beginning scraps of a program --
if they actually write any code at all to start with -- declare that
they've started an open-source project, and believe that they're
halfway to victory. History suggests that in practice such projects
run an appallingly high risk of dying on the vine.

Writing the skeleton of an ambitious program and expecting others to
fill it in for you generally doesn't work very well. If people
download your tarball, compile it, and realize that it doesn't
actually do anything interesting yet, they're likely to simply shrug,
delete it, and move on to something else.

If, on the other hand, you build something modest that actually works,
users are more likely to play with it, realize that with just a
*little* more effort it could be extended in some useful fashion, try
it, and get hooked on making improvements to the software.

Jamie Zawinski talks about this sort of thing a bit in his
"resignation and postmortem", in which he talks about some of the
problems that nearly wound up strangling Mozilla in its crib. (I'm
thinking in particular of "Excuse #2".)

    http://www.jwz.org/gruntle/nomo.html

A fairly perceptive guy I used to work with summed it up succinctly
and memorably when he said, of open-source projects, that "the
interesting stuff happens at the edges, not at the core." (Hi, Scott!)

Dan