The constructor change really seemed to fix the issues we had. I
remember really liking the idea. And I also liked using it. But the
truth is that it lacks certain flexibility for complex situations and
this is not something that could have been easily predicted. I don't
think that blaming anyone for supporting/advocating the change is the
way to go.
And I don't even thing that it is amateurish. You really have to try
some things and get burned, there's no other way. We are not
implementing a JSR. We need to experiment. And this one doesn't really
worked better that what we had before. Which doesn't mean that it's
completely wrong. It's just that the pros don't outweight the cons. But
we needed to try it first to find that out.
I think right now it's pretty sure that 2.0 is not the way to go. So
what we need to decide is whether to backport the models to 1.3 or not.
My personal opinion is - let's do that. I don't really think it's gonna
be that hard to convert old models to new. Simple models will just get
removed the component argument, property and compound model will just
become inner subclasses. Is this really such a big deal?
I think we should have a regular vote on this. One way or another, we
need to come to a conclusion. I wish we could just sit somewhere, have a
beer and discuss this live :)
-Matej
Eelco Hillenius wrote:
first of all this has nothing to do with a "power play". it has to do
with a
simple fact that betas are meant to be stable releases that users can
develop against without worrying much about us pulling stuff from under
them.
Threatening to veto is power play.
Also, I don't agree with your definition of beta being non-breaking.
You are in fact contradicting most of your own behavior in the
project.
from the list i gather there is a large number of users that are
waiting for
a beta before they start upgrading their projects for the above reason.
Not many oppose the model change so far though. It should be quite an
easy change for most occasions really. I think you and Al are making
it larger than it really is.
second of all, yes i was the advocate. i was trying to find a way to
solve
the problems everyone were complaining about. i found away, i proposed
it,
we had a discussion and a vote. we all voted +1. now that i had a
chance to
extensively use the api i dont think it was worth it and so i started the
discussion again. it will take another vote where everyone votes to
resolve
this. i have two large projects on trunk, so i think i am affected
more then
others by this rollback with regards to hours of work. so take that back.
Don't underestimate the amount of work that goes into rewriting and
checking half a book of code example, ALL the wicket-stuff projects I
converted, MOST of the other wicket projects I converted. We can all
feel the pain here, and it is great that at least we tried to come up
with the best solution and worked hard on trying to realize it. But we
also have to admit that this whole situation is pretty amateuristic,
and is not something we should repeat in the future.
The fact that the rest of the team +1-ed on it, makes us all
responsible yes. But +1 a proposal is something else than actively
advocating it and I currently feel kind of had to be slapped back for
giving my support in the past.
Eelco
