[Zope] 2.3.0 really not Beta?

[Brian Lloyd]

> > Because I am not a programmer. I am an administrator and
> > webdesigner. Without this class of users, there would be
> > no use for Zope. Except for your hobby.
> 
> I don't want to participate in a war here. But if you do not
> intend to test betas and report problems (which an administrator
> can still do) then I suggest that you stay one to two versions
> behind and NEVER install a ".0" version of anything.  You won't
> have the latest functionallity but you won't have the problems
> associated with it either.  That's the RealWorld(tm)! <g>

I too am not interested in a war - there is too much volume 
on this list already. So I will attempt to clarify some 
things and leave it at that.

There are many types of users in the Zope community. Some are
early-adopter types who are happy to try out early alphas and 
use the public CVS. Others are very dependable in using beta 
releases in their environment and are very active in reporting 
and diagnosing problems. Some (quite reasonably) don't have 
the time or inclination to do these things and that is fine 
too.

We have a documented release process (see dev.zope.org), where 
we introduce new features (in feature releases) in alphas. In 
the beta cycle, no new features are added - only bugs are fixed. 

The reality is that we get quite a few people who use the betas 
in many different environments with many different configurations, 
installed products etc. - for these folks we are all grateful. 

But even a lot of people saying "works for me" in a lot of different
environments does not guarantee "works for everyone". There is a 
practical limitation to the coverage that a beta cycle can provide. 

There is a curve of diminishing returns, bug reports tail off 
as the part of community using the betas get to "works for me". As 
a practical matter, at some point a release has to be made, because 
of the simple fact that *you will not find the rest of the bugs 
until it is used by the wider community*, and the expanded 
coverage that implies. You could stay in beta for the next year 
and you will still have this problem, just a year later.

If this offends you, then by all means take the advice earlier 
in this thread: stay with an earlier release until you are 
comfortable. Don't take _our_ word for it, take the community's
word. Ask around to people with similar configurations. This is 
not an "open source software" issue - its not even just a software 
issue.

If the new 2002 model of your favorite sports car comes out 
"totally redesigned", you have the same choice to make. Surely 
the engineers who designed did their level best to find problems 
before putting it out there, but does that guarantee that there 
won't be any? No - you need to decide whether you need that new 
sunroof today or if you're going to hang onto your '99 and wait 
for any problems with the 2002 to shake out.

To be fair to the many folks who worked on 2.3, I think that it 
is easily the most stable '.0' release that Zope has had. That 
is thanks to a lot of hard work and attention to detail on the 
part of the contributors, the Fishbowl and a generally more 
mature development process. I don't think anyone who was around 
for earlier '.0' releases would argue this point. 

As with any .0 release, we know full well that there will need 
to be a .1 relatively soon afterward to take care of any 
important issues. This will happen with a 2.3.1 in the next 
few weeks.

</war>


Brian Lloyd        brian@digicool.com
Software Engineer  540.371.6909              
Digital Creations  http://www.digicool.com