I appreciate Craig's focus on users as well. Having said that, I agree
with David's points. The only thing I'd like to add is that people I
have talked to outside the Pluto community are really excited to see
1.1.0 final.
Elliot
David H. DeWolf wrote:
Craig,
Thanks for the thoughts, I think you're correct in considering what the
users would want, and hopefully some of them will speak up.
From my perspective, I think we should consider the following:
1) Pluto is about the container, all else is just gravy to make it
easier to test/use. As such, the quality should be based off of it.
2) We've gone through the cycles of implementing persistence before and
I really don't think it's all that important. The reality is, that
portals embedding pluto will provide their own services for that. From a
developer perspective, if you need a consistent page layout, then modify
the underlying config file once and you're good to go.
3) The biggest question to me is are we ready to guarantee binary
compatibility in the 1.1.x branch. If so, and we're confident in the
quality, then let's label it GA.
4) Just because we release the 1.1 container as GA, doesn't mean it's
the time for everyone to upgrade. If they depend on a portal feature in
1.0.x (though I really doubt many people do), then there no reason why
they can't continue to depend on it. The fact of the matter is though,
that all work on 1.0.x has practically stopped.
5) The persistence mechanism we've discussed requires Java5. I'm
totally against requiring Java5 in 1.1. Do you have another alternative
that we should use in 1.1? If not, let's leave it for 1.2.x and just
start working on 1.2 immediately.
6) If your concern is that these features won't be released for a while
if they miss the 1.1 boat, then the real issue is that we're not
releasing often enough. I'm hoping that changes as we grow and more
portals depend upon us and more developers use pluto, but all of that
can only happen if we cut a release. That said, I have more time for
Open Source (Pluto, Struts, Tiles. . .) available now that I'm out on my
own and working gigs that actually use those projects, so I fully expect
to be able to help driver releases more often than every six months if
required :), YEA!
Those are my thoughts, but I can also understand yours. Anyone else?
David
[EMAIL PROTECTED] wrote:
Thank you, David, for working through all the Jira issues you have for
the past few days. It seems that Pluto 1.1 GA is in our sights.
But, from a users standpoint, I'm not sure if releasing Pluto 1.1
without page layout or preferences persistence is such a good thing.
It seems to be a step back from Pluto 1.0.1 which has both of those
features. Again, I'm thinking about our users, who seem to largely use
Pluto for portlet development. Imagine having to reset your page
layout and preferences every time you have to restart the server while
you are developing a portlet. That would be a real drag. They might
not think the upgrade is worth it.
It might be too late to do a simple preference persistence
implementation (file based?) for Pluto 1.1 GA, but we might want to
shoot for some sort of file persistence of the page layout stuff you
created which simplifies the publishing process. We also need to make
sure the maven and ant deployments work too (see PLUTO-257).
Finally, we should look at the remaining outstanding Jira issues.
PLUTO-234 particularly troubles me since it is the only testsuite test
that fails. I will try to look at this and others as time allows.
That's all I can think of right now. Please add your thoughts on this
issue.
/Craig
