On Wednesday, November 12, 2003, at 06:36 AM, Greg Stein wrote:
On Tue, Nov 11, 2003 at 05:24:59AM -0800, Rodney Waldhoff wrote:On Mon, 10 Nov 2003, Greg Stein wrote: ...Look. Geir, Rodney: if you guys don't believe in A-C, then fine. You don't
have to. But some others do, and it appears that some J-C components will
move. Why is that so threatening?
On the contrary, I think apache-commons has several things going for it,
and I'd very much like to see the project, or more generally, its mission
of creating reusable library code under the ASL, succeed. Frankly, I'm a
bit insulted you'd characterize this interest as feeling "threatened" by
the possibility, and I'm disappointed that you'd respond to any difference
of opinion about how to best accomplish that mission by with comments
along the lines of "and if you disagree, then go away".
I don't want you guys to "go away". Quite the contrary. I'd prefer if you
offered constructive criticism, assistance, and insight. But to be honest,
a lot of what I have seen is "but why? J-C does this already. let's just
promote that instead." or "why should they give up their community?" or
"svn creates barriers" or ...
There has been very little, "this is how J-C operates, and to make them feel at home in A-C, we should do X."
To be honest, Rodney: yes, I may have lumped you in unfairly with Geir,
whose positions are very much more "J-C does this right. why should anyone
bother with A-C?" But my response was built from an overall impression,
which is that I'm not seeing "we can fix things <this> way".
I guess my problem is that I don't yet buy into A-C as a forgone conclusion as the solution to J-Cs problems, when there's incredible potential to fix what is deemed wrong with J-C. I do see value in the language specific community of J-C, given the richness of interaction due to binary interoperability. Also, I obviously think that the 'commons' notion is a good idea. Math *is* a good example of a candidate to move to A-C is there's interest in doing math in some other language for speed, and then the j-c math would use JNI to get to the fast libraries in A-C. But there's a few ways to skin that cat too.
It appears to me that A-C is being bootstrapped from top down rather than bottom up. J-C was a bottom-up bootstrap put together by a bunch of us in jakarta who noticed a problem (lots of duplicated utility code in the jakarta sub-projects) and went about solving it. It was a community solution - many, many jakarta members debated [seemingly endlessly] about our charter, proposed it to the PMC of Jakarta, and got it going. It's been a runaway success. (Which is the problem - you have a previous post where you point out the obvious problems, like the single mail list, etc)
And please don't misrepresent my positions - I don't wonder why anyone should "bother" with A-C - I just don't see anyone naturally doing it besides you and Justin w/ surf. It sounds like A-C is the proper solution for serf if it didn't fit in httpd, but wonder why there hasn't been any natural migration of other code from Apache projects.
In a way, A-C is what I hoped DB could be, language-neutral (or any-language) tools and utilities, just w/o the db focus. And to that end, I'm surprised too that there wasn't more that moved into db. I thought that there would be a bunch of ETL-ish things in languages other than Java that would want to find a home there, but there hasn't been any interest. I guess either because of poor advertising, or established projects don't see the value of moving.
To me, I read the resistance to A-C and some of the responses as feeling
threatened by a loss of community, self-rule, etc. So... that's how I read
it; if that doesn't represent your thinking, then I apologize, and will
blame email as an imperfect medium. What may simply be "devil's advocate"
discussion or similar can easily be construed as reading a position or
belief, which can be quite incorrect. And, of course, the easiest way to
correct it is to call out the issue as you did :-)
I think that 'loss of community' isn't a threat - it's a mistake, but that's just my opinion. And 'self-rule' wouldn't change - the people that should be in charge, the committers, will probably be in charge in A-C. So no real change from those perspectives.
I think I'm just confused. There is a 'Jakarta must die' flavor to all of this. (That's a quote, btw). Call it paranoia on my part, but when I see Robert, who is usually very factual and correct, on the J-C list saying things like :
"commons-maths will still be part of jakarta-commons :)
it'll only be managed by the apache-commons pmc.
best of both worlds :)"
http://marc.theaimsgroup.com/?l=jakarta-commons- dev&m=106841721301207&w=2
I'm just simply baffled by what's going on. How could it be that a J-C project is managed by A-C's PMC? Is that something the board has mandated? That A-C will managed codebases in other Apache projects?
Not to put to fine a point on it, but if I were "threatened" by the
possibility of jakarta-commons components moving to apache-commons, then I
probably wouldn't be offering suggestions about how best to encourage j-c
components to move to a-c, as you see me doing at
<http://article.gmane.org/gmane.comp.apache.commons.general/151> and
<http://article.gmane.org/gmane.comp.apache.commons.general/158>, to name
two examples.
And re-reading those two posts also leads me to say, "hunh. there *have*
been nuggets of 'here is info about J-C'". Mebbe too thick-headed to see
them :-). I might suggest that the answer is to simply start asking J-C
components if they want to move. If they say "yes", then fine. If they say
"no", then we ask "why?". That provides very concrete info rather than
continual discussion and theorizing...
The math component is doing that, and there are some interesting possibilities.
Cheers, -g
-- Greg Stein, http://www.lyra.org/
--------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
-- Geir Magnusson Jr 203-247-1713(m) [EMAIL PROTECTED]
