Re: [Libreoffice] Proposal to join Apache OpenOffice

Allen,

On Fri, 2011-06-03 at 22:06 -0400, Allen Pulsifer wrote:
>  
> So here is my suggestion: I propose the everyone here head over to the
> Apache Incubator and join the proposal as an initial member.

Just so you know, I've been following that thread on the Apache list by
reading the archives, pretty much fully, so I have a pretty good idea of
what's going on over there.

I'll keep this short (but will probably end up being long anyway).  It
makes zero sense, for whatever reason, to join an entirely different
project that someone else (in this case IBM) cares about when we have a
perfectly live, vibrant, passionate and fun project to nurture, and that
project already exists today.  Just because the two projects used to
share the same code base in the past doesn't mean anything here.  We are
different projects now, plain and simple.  Suggesting that we somehow
owe anything to them just because of the past is, to put it mildly
insane, and in some way insulting.

And let me put this in practical terms.  Managing a project is a big
chore.  I have hard time keeping up with all these bug reports, patch
reviews, helping other hackers, many releases, while at the same time
trying to clean up and refactor the code base to modernize the code.
Doing that in two projects would basically force me to cut back on that
for this project.  That would be a terrible disservice to those who
believed in TDF and LibreOffice, not to mention that would be against my
own will.

And we have done some amazing things in the past 8 months, none of which
is in the Apache OOo code base (whereever it may be right now).  Telling
us to forget all that and start from scratch is simply insane.

Plus, I've been burned (as you probably know) by the stupid corporate
bureaucracy trying to control the project in the past, with OOo project.
So I am very cynical about a corporate participant promising to help a
project in the future, promising to donate code at some unspecified time
in the future.

You may say "but Apache is all about individuals".  That's probably
true, but where are those individuals who are working on IBM's Symphony
product right now?  Rob claims that they'll be working on Apache's OOo,
but none of them are in the discussion right now.  I don't doubt Rob's
qualification, but I really doubt that Rob will be the one working on
the code base.  I'm sure he is too busy for that.  I would be more
comfortable if the individual hackers from IBM were openly speaking up
in honest terms.  But so far that's not happening.

Also, the difficulty working on both projects also applies to the IBM
Symphony team.  I doubt they'll be working 100% on Apache OOo.  So the
obvious question is what percentage of their time is spent maintaining
the Apache OOo code base?  They claimed they'll allocate 35 engineers,
but how much time they'll invest remains to be seen.  Plus, managing a
project is an on-going process; if they just dump some code from
Symphony to Apache OOo every now and then without follow-up clean-ups
and bug fixes, that wouldn't be called managing & maintaining a product.
That's just code dump.

And lastly, folks over at Apache seems to underestimate the difficulty
on bootstrapping this massive code base, setting up the repository and
managing different branches and releases.  Very little of that seems to
be discussed.  Talking from my experience, you can't manage this code
base using Subversion unless you want to cut back heavily on
productivity and efficiency.  I believe we have managed to set up, and
continue to improve our infrastructure over here.  And I don't see a
reason why we need to move to an imaginary infrastructure that's not
even set up & much less adequate.

The only reason I would join over there would be if my employer wants to
tie my hand and put a leech around my neck to drag me over there to the
Apache OOo land.  Then I would be working over there, with great
reluctance.

Kohei

-- 
Kohei Yoshida, LibreOffice hacker, Calc
<kyoshida@novell.com>