Peter B. West wrote:

> Such a move would, obviously, have little or no impact on the 
> main project.  The situation with FOray is more complicated.  
> I don't know whether it is Victor's intention to fork from 
> HEAD and continue the development along the lines he has 
> previously discussed, or to attempt to integrate HEAD and the 
> maintenance branch in some way.  In any case, what Victor is 
> doing will closely parallel the HEAD development, and this, 
> combined with the possibility of some financial support, has 
> a great potential to de-stabilise FOP.  I'm not saying this 
> as a criticism of Victor, but as a bald statement of the reality.

How, if FOray "will closely parallel the HEAD development" could it possibly
"ha[ve] a great potential to de-stabilise FOP." If A = B, there could be no
reason to switch from A to B. There seems to be implicit in your comment a
nascent fear that maybe FOP really would benefit from being modularized.

The initial, relatively small, fork will be from the maintenance branch.
While it is being modularized, I will keep it in sync with my local copy of
the FOP maintenance branch code, and hopefully submit a patch to that branch
for your consideration. It is my hope that the benefits of this approach
will be compelling enough that someone will want to roll it into HEAD as
well, but that is your call, not mine. Depending on the success of this
approach and its reception, I'll decide whether forking other pieces is
needed (there are more extensive comments about this on the FOray home
page.) Some of those forks may well be started with code from HEAD.

WRT FOray's effect on FOP's stability, I don't see a need for any concern:
1. FOray's work is available to FOP (Apache 2 License), and I hope that FOP
(among others) will use it.
2. Most importantly, if I thought that there were FOP developers who shared
my zeal for modularizing FOP, I would a) not have left the project, and b)
would even now be trying to reenter FOP to do the work there. The very brief
review I took of the archives last week led me to believe that instead, the
baby steps I had taken toward my goals are being dismantled. (I'm not
offended, but merely reminded that I'm out of sync with you all.) Now, it is
possible that there are some developers not currently involved with FOP that
will be attracted to FOray. If there are compelling advantages to FOray's
approach, then that is as it should be. Further, if there are people who
see, for example, a benefit in having a freely-distributable alternative to
JAI, who want it for a totally non-FOP related purpose, and who contribute
to FOray, this seems like a net benefit to FOP. A similar situation would
exist for someone who wants to build an FOTree for a voice-related
application, use a well-crafted library of PDF-building routines, or write a
layout engine that was optimized for some particular axis.

There is no explicit intention to reconcile HEAD and maint. However, there
are several sets of circumstances that could have that outcome as a
byproduct. The biggest unknowns are the timing of LayoutManager's
completion, and the speed of progress in FOray.

WRT to the opportunity for funding, that had no part in my decision to use
sourceforge. (I did opt-in to accept contributions -- why not?). And I doubt
that serious funding will materialize through this channel. I would vastly
have preferred to do this within FOP for many reasons. My second best choice
was to find some other home on Apache for the independent modules, but I
frankly don't have time to jump through all of their hoops. The great
benefit to using sourceforge was the easy access. I can still get and
receive the benefits of an open-source approach and let you guys stay on
track with your work.

Victor Mote

Reply via email to