What is to be released to the world?

[Joe Klemmer]

Andrew Conkling wrote:

> Hello Joseph (no Jimi that time... doh!),
> Your post leaves me with a few impressions.  Firstly, that this is a
> call to developers to expect less from users who wish to test/find
> bugs, or otherwise help in a non-coding way.  I can appreciate this;
> with the complications of building unstable software that doesn't
> often sit well on top of a stable system, it's hard to test Xfce if
> you also want to use Xfce.  (Disclaimer: Xfce happens to be some of
> the most stable development code I've run.)
> 
> On the other hand, it would seem to discourage those willing to take
> the time to keep up with development code to "go the extra mile" to
> help in this way.  I mean, if developers had a "look but don't touch
> (unless you intend on submitting patches)" attitude about their code,
> I feel like this would sacrifice a bit of openness, if only in spirit.
> 
> Did I read you correctly, or have I got it wrong?

	It's not a easy line to walk because you (i.e. developers) have to walk 
the line between the two perspectives.  The strength of open source is 
that it does get much better exposure and more of the proverbial "eyes" 
that make bugs shallow.  But at some point you have to draw a line and 
say, "Ok, this is where we set the marker and everything will move on 
from this point."  CVS/SVN code should be "look don't touch" but the 
time code is in this state should be very short.  Beta releases on a 
project like Xfce could conceivably be put out every week but I would 
think every couple of weeks would be better.

	It's a balancing act.  Otherwise the code will be perpetually in 
CVS/SVN mode and never actually be a release.