OpenOffice.org: Breaking up is hard to do
The birth of Apache OpenOffice.org is not without some pain
The proposal to shift OpenOffice.org into the Apache Foundation as an incubator project was approved in a vote over the weekend by a wide margin of 41-5. But the shiny new Apache OpenOffice.org "podling" won't come without birthing pains.
Rob Weir, ODF Architect for IBM, praised the vote outcome in his Monday blog entry, but then took the time to skewer critics of the move to Apache.
Weir's comments were directed at a "small but vocal minority of non-Apache members who disagreed with the proposal attempt to derail it." Weir stated that all opinions were welcome, but didn't seem too thrilled with the idea that some dissenting opinions were actually lobbying against the Apache vote: "The day open source advocates decide to [smother] a new project in its crib, because they personally favor a slightly more established project is the day that FOSS dies."
That statement is rife with hyperbole. Of course a dissenting opinion would want to change the vote's outcome. Why else hold the opinion? Weir seems to want to vilify the opposing side, and appears to be taking this personally. He then goes after what he believes is an "unstated assumption" in the reasoning of the objectors to the Apache OpenOffice.org move.
"I believe that one unstated assumption in their reasoning was that there is a scarcity of developers and a scarcity of users in the personal productivity application area, and that the success of a new project can only come at the expense of another project, in this case at the expense of LibreOffice. The assumption was that we're playing a zero-sum game, and like junk yard dogs we're fighting to the death over scraps."
Though Weir interjected this objection himself, this is a concern that has been voiced at other times by members of LibreOffice, the OpenOffice.org-based project that forked away from Oracle and OpenOffice.org last year and is now held by The Document Foundation.
Weir argues that this scarcity is a fallacy, because as Apache OpenOffice succeeds, it will promote growth within the OpenDocument Format (ODF) ecosystem, which will ultimately benefit the LibreOffice project.
"OpenOffice and LibreOffice can both win. OpenOffice and LibreOffice and Calligra Suite and AbiWord and Gnumeric can all gain users at the same time. And this can happen at the same time that mixed-source applications based upon OpenOffice, like Lotus Symphony, also grow and gain users."
Weir's use of the term "mixed source" and his mention of Lotus Symphony confirms, by the way, a theory I had as to why Apache was the ultimate home of OpenOffice.org and not the Document Foundation. Weir's employer IBM has heavily influenced moving the codebase to Apache because IBM much prefers the permissive Apache Software License (ASL) as opposed to the General Public License (GPL), under which LibreOffice is licensed.
IBM's goals were clarified in Weir's blog. With the OpenOffice.org code under the ASL, IBM will be free to apply that code directly to whatever proprietary software code they want and distribute it in Lotus Symphony under a proprietary license, something the ASL allows and the GPL does not. This is what he means by "mixed source."
Weir's argument that the developer community is not a zero-sum game may be, in theory, correct. In practice, however, I cannot believe that personnel shifts between one project and the other won't affect LibreOffice and OpenOffice.org. This community is (for now) just not that big. Indeed, if Weir is so unconcerned about the number of developers, why harp on the 87 members of the Apache OpenOffice.org project? Perhaps a little jab at the hype around the 33 OpenOffice.org defectors to LibreOffice back in November?
Interestingly, the opposing side's arguments aren't much better. Keith Curtis, who claims to be an "unofficial spokesman" for LibreOffice, was among the more vocal objectors to the Apache OpenOffice.org incubator vote. I outlined Curtis' objections myself last week.
Curtis' objections were made public on his blog, but they were originally sent to the Apache general incubator mailing list. If you follow the thread in which his post appears, it quickly becomes evident that Curtis's understanding of licensing or history is poor at best.
Curtis accused IBM and Apache of forking OpenOffice.org with this vote, disingenuously avoiding the fact that LibreOffice itself was a fork from the original OpenOffice.org. Curtis also advocated simply stripping the ASL from OpenOffice.org code (using grep) in order to use OpenOffice.org code indiscriminately.
Such an action would completely violate the ASL (or any other license for that matter) and its copyright. And it's almost not necessary, because (at least in one direction) code can be shared between the two projects.
LibreOffice, which is dual-licensed under the GPL v3 and the Mozilla Public License can benefit already from Apache OpenOffice.org. The Apache Software Foundation explicitly states that "Apache 2 software can therefore be included in GPLv3 projects, because the GPLv3 license accepts our software into GPLv3 works. However, GPLv3 software cannot be included in Apache projects. The licenses are incompatible in one direction only, and it is a result of ASF's licensing philosophy and the GPLv3 authors' interpretation of copyright law."
So, LibreOffice can use OpenOffice.org code, but OpenOffice.org cannot use GPL'd LibreOffice code.
Curtis' misunderstanding aside, he seems to have entered into a one-man crusade against the Apache OpenOffice.org project, though it is not clear what is his connection to LibreOffice. He claims to have contributed to the OpenOffice.org codebase at some point, though he is not listed in the LibreOffice or OpenOffice.org credit pages. (Curtis did file one OpenOffice.org bug in 2009.) Curtis has also not applied for membership in the Document Foundation, according to sources there.
So for whatever motivation, Curtis has taken up the campaign with a vengeance. In a comment on Weir's blog, Curtis attacked the initial 87 contributors Weir touted.
"The 87 people signed up because you created a fork and many people are naive. They don't understand what is going on, but they see the trademark, Apache, and IBM and assume this must be a good idea."
To which Weir replied,
"...[S]o the 87 people who signed up with Apache OpenOffice are 'naive' and 'don't understand what is going on'? I'm intrigued. They just participated in a two week discussion on the Apache list, a discussion in which you posted your long list of reasons to reject the project proposal. I didn't see anyone take their name off the wiki after reading your note. Did you? So if you think that these 87 initial [committers] are 'naive' and 'don't understand what is going on', then maybe you're to blame, for being unpersuasive in your role as 'unofficial spokesman' for LibreOffice. Just a thought."
What's that line from Megamind? Oh, yes:
"Girls! Girls! You're both pretty! Can I go home now?"
Luckily, things are moving forward, despite all the yelling.
While the community barbs continue to be traded, Apache OpenOffice.org is settling into its new status as a podling project. And the Document Foundation just announced the membership of their first Advisory Board, a group that includes members from SUSE, Red Hat, Google, and the Free Software Foundation.
Time will tell how the split between these two projects will ultimately affect the health of each project, and the ODF ecosystem as a whole.
One thing is for sure, they need to get their collective acts together soon. For instance, neither platform seems to have a cloud-based offering in the works. With Google Docs already making inroads in user space and even Office 365 on the way, it seems these two projects have a lot more to do than yell at each other.