[moving a thread from ant-dev to general]
Conor MacNeill wrote:
>
> As far as I can tell from the minutes, the underlying issues of
> both these threads were dealt with at the PMC meeting. Given
> the resulting action on guidelines, voting, etc, we could hope
> that Jakarta as a project, would "work".
That is my opinion too (as recorded in the minutes, BTW).
> So how does the PMC set the overall direction for the Jakarta
> project?
As we did in the above two issues. Primarily by being the carriers of the
ASF culture, and secondarily by the selection (and possible culling) of
subprojects. See below.
> > in essense what do we (Jakarta) want to be when we grow up?
>
> Indeed, this is the interesting issue. In a Jakarta project
> containing Ant, tomcat, log4j, ... what is the common goal
> that binds them into some cohesive community. Perhaps, Pete
> is right and that common goal is something very broad such as
> "the best Java software (not just the best free Java software
> :-)" How would the PMC set the direction on such a broad
> scope?
It could not. As an extreme example, should Sun ever decide to open source
the JDK, this would not be an appropriate place to house it IMHO. Such a
large piece of work would deserve a home of its own as Apache and the JDK
would compete for focus.
What we have tried to do is select projects that complement the need for
servlets. Taglibs are obvious as are most other technologies which build
upon servlets. In other cases, we have grown by necessity when a suitable
alternative is not available (or not under the appropriate license). As an
example, regular expressions are often needed to validate forms. Had these
tools been available under an appropriate license, they would have simply
been used as is.
We have tried to control the rate of growth. We have tried to maintain
focus. We haven't always succeeded.
Ant is a special case. Originally it was simply a tool needed to build the
servlet engine in a cross platform manner. Nobody predicted it would take
off the way it did. I could argue that it is the tool of choice to build
the Jakarta subprojects, and no suitable alternative exists. So
technically it meets the criteria above. What appeals to me more about
Ant, however (and potentially someday in the future, Alexandria and my
tinderbox) is that they help define the community. Everybody who becomes
aware of Ant becomes aware of Jakarta, and vice versa. And everyone in
Jakarta uses Ant.
One other aspect worth mentioning. Subprojects are not simply included
into Jakarta based on merit, but also by the choice of the owners. The
log4j project development team sought inclusion of their project into
Jakarta. The cocoon and junit folks made other choices. Note: I agree
with the former, but am saddened by the latter, for example, see:
https://sourceforge.net/support/index.php?func=detailsupport&support_id=113180&group_id=1
https://sourceforge.net/support/index.php?func=detailsupport&support_id=112649&group_id=1
In closing, neither Apache nor Jakarta are perfect. We have a much higher
barrier to entry than SourceForge. That's actually one of our strengths.
- Sam Ruby
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
