[EMAIL PROTECTED] wrote:
>
> On Thu, 15 Mar 2001, Geir Magnusson Jr. wrote:
>
> > [EMAIL PROTECTED] wrote:
> > ....
> > >
> > > That's why we need the extra " or sponsored by a jakarta subproject".
> >
> > I don't think so. I think that additional phrase will lead us back to
> > the current situation. As one of the problems that I hope Commons would
> > solve is making sure that 'released' components are documented,
> > installable and supported. If it's just a tagged set of classes in the
> > sandbox, we are back to the same problem we have already - released code
> > that a project depends upon that is not the projects main 'deliverable'
> > is undocumented, not independantly buildable, testable, and installable,
> > and largely hidden from public view because it's not separated and
> > promoted.
>
> Than what's the problem ??? You think the problem the commons would
> resolve is "released" components - you have that.
Quite frankly, I am not 100% sure the commons would solve that, but I
want to try.
The problem that I was addressing was that 'sponsored by a Jakarta
subproject' is vague to me, and seems pointless - if a project, say
Velocity, wantd to use the 'unreleased' code in a velocity release, I
assume we would just include the code we needed. I can't see other ways
that would be practical for the end-user.
I gotta do this line by line :
> Why don't you let other people to try a different thing ?
Why do you believe I want to stop you? I don't. I respect and support
the idea, your idea, of Agora/playpen/sandbox, for what its worth. And
even if I did, I can't.
And anyway, I thought my point above was limited in scope to the simple
phrase "or sponsored...".
> I am interested
> in sharing the work on some common components -
As I am - the problem that led me here is that we have code duplicated
throughout jakarta, and having some way of getting the separate groups
to work on common basic things would be great.
> I don't care too much
> about releasing the components, it is not my main goal.
Ah. There we differ, and I think it doesn't matter - if you don't want
to release, don't release. You can't be forced to release, except by
your own collaborators, and thats between you and them.
I do think that releasing components would be a good thing - certainly
every other Jakarta project seems to believe that releasing software for
others to use, predictably, is a good thing. So I will try to release
anything I work on that I think is good and valuable to others. You
don't have to.
>I think this will
> happen as a side effect, because the code will be subject to a bigger
> community and in direct contact with the projects using it.
I am not trying to stop anything you want to do, and I wish you would
explain to me how I would be stopping you.
I think that if we look back, while I have been very concerned about
getting the things I want into this, it hasn't been to the exclusion of
other approaches.
On this specific issue, release, I think that the status quo *isn't*
*working*, and depending upon the release process to come out as a 'side
effect' might be unsatisfactory. That's what I saw when I went looking
for a DBCP - Turbine has one, buried. Struts has one - less buried, but
still, if not hidden, certainly 'unemphasized'. They aren't going to
release those parts as separate entities.
I think that if you have a component, and you want to 'release' it, if
it isn't released via inclusion in another jakarta project, it should be
done via commons - that is one of the purposes of the project.
>
> But the fundamental issue is this intolerance I find again and again and
> again in jakarta - why does everyone things that there is only one
> solution and one way to do something ?
Why do you say that? (I assume you are addressing this to me
directly...)
> It is very likely that agora will fail - as far as I can tell intolerance
> is the rule here, and Agora's can't succeed in such environment. But we
> can try - maybe there are people who can repect other's ideas and not try
> to impose their will and solution.
>
> What's the problem ? Is the library the guardian of Apache quality, the
> only absolut experts in components ? Are jakarta projects some stupid
> entities who can't be trusted to release something to the public ?
<sigh> I hope you really don't think I believe that. To me, and I
haven't been around here that long, the Jakarta projects are entities
that serve their specific charter, and produce software for the public
that meets that charter. In the course of that mission, the Jakarta
projects appear to develop building blocks that are necessary for their
mission, but incidental to their 'product'. They therefore do not
package and release the building blocks in a way that is immediately
accessable and useful outside of their specific deliverable.
The DBCP is a great example - everyone needs one, seems like everyone
wrote one, but no one made it so you could just take it and use it (See
Poolman as the great counterexample).
That's what I want to see come out of this - joint development and
productization of common shared components and parts. Note, that is NOT
TO THE EXCLUSION OF OTHER THINGS, LIKE AGORA.
> Your development model is wonderful - but it's not the only one.
I don't think it's wonderful, nor do I think it's the only one. But
some of the ones I see now aren't meeting a need, hence the endless time
spent here trying to get this going....
geir
Geir Magnusson Jr. [EMAIL PROTECTED]
Developing for the web? See http://jakarta.apache.org/velocity/
