Re: [Exist-open] New RC release?

[Andrzej J. T. <an...@ch...>]

Jason wrote:

> Shouldn't we strive to do everything we can to make sure that each
> release of eXist is as stable and bug-free as possible?  And if all it
> takes is a little extra SCM process and a couple of weeks (in a
> development cycle that has taken many months), is that too big a price
> to pay?

Following this approach would result in no releases ever making it out the door.

To some degree, Caveat Emptor (or should that be Caveat Openus Sourceus?) should apply.

It goes without saying that for those users that are still on 1.2 (!!!), they should not switch a production system over
to 1.4 without significant testing first.  Regardless of how much testing the dev team has done, or how supposedly
stable/bug-free the release is. And many will not begin this testing process till 1.4 is official.

If stuff doesn't work then the users have a number of choices:

1) Stay on the older release (1.2 or whatever) since that is known to work. There's nothing forcing users to upgrade!
2) Dive in and help test/isolate/document and even maybe fix the bugs they find in 1.4.
3) Hope someone else finds the bugs and wait till 1.4.1 (though that can be an infinite recursion)
4) Find workarounds or live with the issues with 1.4 and deploy into production.

All of which could be potentially "optimal" strategies, depending on each users individual circumstances.

And I would also suggest that those users who have a lot of "investment" in eXist and use it heavily should seriously
consider doing SVN trunk builds and testing frequently against the latest code base. That way, when a release comes out,
there shouldn't be many/any surprises. That is what I do, and it's very effective, plus it has the benefit of actively
testing and improving eXist on an ongoing basis and keeping our application in close sync with eXist's evolution.

Better still, next time maybe you could start your testing sooner, Jason (eg. using SVN trunk) and help to improve the
final release rather than whining that more SCM processes and delays are required at the nth hour?  And let me tell you,
your comments do sound like whining to me.

So to answer your question, YES, it's too big a price to pay. At some point you just have to cut things off and deliver
a release. Wolfgang (and others) have put in massive amounts of work to get this release ready, and many people are
awaiting it anxiously.

Release early and release often should be the mantra. Not "hide under your mom's skirts till it's guaranteed to have no
bugs whatsoever". LOL

+1 from me to create a 1.4 distribution release immediately, and start work on 1.4.1 and 1.5.

My 2 cents worth as one of the committers (albeit not a major one).

-- 
Andrzej Taramina
Chaeron Corporation: Enterprise System Solutions
http://www.chaeron.com