[dev] Fwd: HEAD

[Eric Rostetter]

Quoting Michael M Slusarz <slusarz at bigworm.colorado.edu>:

> 1.) It's been a while since the release of the last major versions (year and
> a half almost?)

Nothing wrong with that.  But as you say, time to at least start thinking
about the next release, so I won't be 3-4 years before we get to it.

> 2.) For anyone who has read Eric Raymond's "The Cathedral and the Bazaar",
> one of the most important statments he makes about open source software is
> "release early and release often".

I think this has been taken too literally recently.  RedHat and others have a
strict 6 month rule now, which basically means "a release every six months
whether or not it is ready" which is a bad thing.  Just look at RedHat 7.0
to see how bad that kind of "rush to release" can be.

> 3.) There's all sorts of new stuff in HEAD that is just yearning to see the
> light of (mass distribution) day :)

Yes, and we could have released it a while back, just before we did the
new UI interface.  That was a stable time, with lots of new features, and
no real problems/debates other than making a major all-encompassing
UI change.

But we didn't make a release, and instead we made a major change to the UI.
Which means we obsoleted all the docs.  Which means the new UI isn't
completed and needs work.  Which means that some of the other apps have
yet to be updated to work well with the new UI.  Which means we have a lot
of work before we can release the next version with the new UI.

My point is basically that we should have discussed if we wanted to make
a new release -- or branch -- before we did a major UI change.  By not doing
so, we may have delayed any release much further into the future, or caused
a release to be incomplete and be forced to make rather major changes to the
UI and docs after the release.

We could have done a Horde 3.0 release, then made the UI changes and in a 
short time released the new Horde 4.0 with the new UI. That would have been
a good "release early and release often" plan.

> Maybe we can at least identify what needs to be cleaned up in each component
> so we know what to shoot for.

Number one thing is the new UI.  Number two on *my* list is the charset
code (e.g. its use concerning access keys for just one example).

Less important, but also on my list, is support things.  New user's guide
would be great.  New php/pear releases with RPMS would be great.  While not
part of the software, having these things ready with the release would make
it a lot easier for people to upgrade.  Since the new Horde would require
new php versions, pear versions/modules, and new docs, it would be great
if we were prepared to help people here also.

> Maybe start releasing the code as 'alpha' or 'beta' to induce more people to
> download and try (I realize we do snapshots, but I think people are still
> hesitant to try something out that isn't labeled at least 'beta').

I can't even get people to try my release candidates...  I don't I could
get them to run pre-release candidates.  The release candidates are already
beta, so we'd have to go alpha.  And we can't endorse alpha code for 
production machines, so I doubt we will really increase our user base
this way.
 
> Call me names, tell me my reasoning or stupid, whatever.  Just wanted to
> give my current feelings.

I pretty much agree with it all, except I'd warn against taking the
"release often" bit too strictly. Release early and release often but
only release something when it is ready, and never before it is ready.

Right now, I don't see it as ready...  In a short while, maybe it will
be.  What we need perhaps is for someone to jump up and down and scream
and shout when we hit those points where it is ready, so we can get a
release out at that time instead of making major changes which preclude
a new release coming out any time soon.

Okay, so now call me names, tell me my reasoning or stupid, whatever. (To
steal Michael's line).

> michael
> ______________________________________________
> Michael Slusarz [slusarz at bigworm.colorado.edu]
> The University of Colorado at Boulder

-- 
Eric Rostetter
The Department of Physics
The University of Texas at Austin

Why get even? Get odd!