Subject: Re: Exploring the limits of free software
From: Brian Behlendorf <brian@hyperreal.org>
Date: Mon, 24 May 1999 08:51:15 -0700 (PDT)

On 24 May 1999 craig@jcb-sc.com wrote: 
> Mainly, we need to do a good job reminding people (as we seem to keep
> reminding ourselves) that OSS is not itself the magic bullet, except
> insofar as the market uses it in the gun to achieve whatever objectives
> it has.  If the market values bloatware over stability, OSS will help
> it get there, probably faster than MS ever did.

In my view it still boils down to the management of the OSS mainline
project, e.g. your management of g77 or the Apache Group's management of
the HTTP server.  IMHO we in Apache have been fairly good about being able
to resist bloat, thanks to a really good API (which allows us to tell
most people "put your desired functionality into a module and
distribute it yourself") and a good modules directory, so that people
don't feel like no one will use their code if it's not in the core.

So I think I agree with your thesis that there may be more pressure on the
OSS project maintainers to integrate code developed in the sXc, and that
there is the risk of "bad code" being developed (code which doesn't
generically solve the problem the way it should), but I'm hoping that the
peer reviewer could play a role in helping mitigate this.

	Brian