Re: Coffee and Open Source

[Jeff Licquia <jeff@licquia.org>]

On Wed, 2003-01-29 at 18:13, mike808@users.sourceforge.net wrote:
> > Again, there's the problem of comparing physical devices to software. 
> > It costs to make copies of physical objects, which explains how people
> > can make decisions to not copy them for posterity.
> 
> I'm not talking about the cost to *copy*.
> I'm talking about the cost to *create* in the first place.

OK.  But in your analogy, management killed off the free coffee because
of unintended usage of a scarce resource.  And it was related to product
activation, which is fundamentally an act to restrict copying, not
creation.

> > But this is fundamentally a movement about people doing what they see as
> > important.  If something isn't maintained, then evidently few out there
> > see it as important.
> 
> Not quite true. If Linus and everyone else stopped working today on the kernel,
> it wouldn't stop people from making copies and using it. What people see
> as important today is to get the free (as in beer) software. The vast
> majority don't give a rip about how it got there in the first place.
> Any more than you care about how much a company spent on R&D or M&A to
> create non-free software.

Yeah, but we don't need a majority of users to work on something, just a
persistent minority.  And as long as something continues to be used,
you'll always find that someone out there isn't satisfied with just
using it.  See the original Halloween Documents for an example.

> > And if you see something important being left by the wayside, the best 
> > thing to do is to help out.  A thousand points of light, and all that.
> 
> I agree. That's the idea the OS movement needs to do more selling of.
> The process of actually contributing code (and documentation!) to various 
> projects needs to be somehow more efficient than it is. It's definitely
> a *long* way from where it was before the likes of SourceForge and Savanna,
> Bugzilla, RT, CVS repositories, and all. 

Yup.  I submit the current flurry of activity in the version control
field as an example of continued improvement on that score.

> And while OS has gotten easier to contribute, it's also getting difficult
> to jettison cruft and obsolete code. Stroll through CPAN to see some.
> It's easy to add, but hard to prune. Which is fine when we're starting up.
> But as our garden matures, our pruning skillz need some help.
> 
> We need to work on making code management and maintenance also more
> efficient and easy for the (drumroll - new buzzword alert) 
> "casual contributor" to participate. Obviously, this is all "writ large"
> and general, so YMMV.

Well spoken.  My mileage certainly doesn't vary. :-)

I brought up version control before.  IMHO, one improvement stands out
among the rest: getting a VC system that does changesets and operates in
a distributed fashion.  This will encourage casual contributors more
than any other change I can think of at the moment.  As it turns out,
the two features are intertwined, so any system that supports one
inevitably supports both.

Unfortunately, the only examples of this that exist today are
proprietary: BitKeeper, Perforce, etc.  Right now, the best OSS/Free
version control system is Subversion; even in its prerelease state, it's
much better than anything else, including CVS.  Changeset support and
distributed operation are slated as post-1.0 features, though; their
main goal for the present is to simply fix the bugs and release 1.0.

So if anyone out there is looking for something practical to do to help
improve the OSS/Free development process, you now should have at least
one idea. :-)
-- 
Jeff Licquia <jeff@licquia.org>

-
To unsubscribe, send email to majordomo@luci.org with
"unsubscribe luci-discuss" in the body.