Re: A BCEL future? Was: [POLL] Committer attendence

On Tue, 2004-12-14 at 16:35, Henri Yandell wrote:

> BCEL has 6 committers at the moment. 3 are on the Jakarta PMC, but
> Markus Dahm and Enver Haase seem to be the main committers by a long
> way and are not; Markus has committed a couple of fixes last month,
> but that's it for 2004. Last website deploy was in 2003.
> 
> 372 java files, depends on jakarta-regexp, though only for one class.
> Otherwise it's nicely stand-alone. My vague understanding of BCEL is
> that ObjectWeb's ASM is in more use nowadays, but vague is the correct
> term for me there :)


I investigated BCEL and ASM when starting work on CLIRR
(clirr.sourceforge.net) about six months ago (clirr currently uses
BCEL). ASM had a fair level of activity around it, while BCEL was *very*
quiet.

> 36 bugs in bugzilla since March 2003, so definitely work to be done,
> with Cocoon being a user.
> 
> Some ideas:
> 
> * Wait for Markus and Enver to wake up, and see how we can help them.
> * Someone find time to apply patches and consider bugs. Always hard to
> do and yet still maintain a good sense of where the codebase is I
> think. But at least it lets the community keep things moving.
> * Move into Commons. I think it's probably too big, but I feel the
> Commons community is good at such things and the scope of BCEL doesn't
> seem that huge I might be underestimating it).
> * Move to another Apache/Jakarta project. Seems unnecessary unless it
> turns out that only Cocoon, say, uses BCEL.

ASM (asm.objectweb.org) is, it would seem to me, smaller, faster,
cleaner and has more active development. And it's under a BSD license. 

ASM seems to simply have found a better approach than BCEL. No criticism
intended to the BCEL developers; ASM is a more recent project that may
well have learnt from earlier projects like BCEL.

It would seem sensible to me for projects to migrate from BCEL to ASM,
and for apache to retire BCEL. Of course it would be nice for one
significant BCEL-using project (eg cocoon) to look at whether ASM has
the necessary features before doing so; certainly ASM had all the
features that CLIRR needed but that is a fairly tightly focussed
project.

Or a response from a BCEL developer to indicate what BCEL offers that
ASM lacks would be useful...

Regards,

Simon