June 08, 2011, 1:12 PM —
Former Microsoft developer-turned-free-software-advocate Keith Curtis has outlined his thoughts on why it would be a bad idea to make OpenOffice.org an incubator project stewarded by the Apache Software Foundation.
The missive, drafted to the Apache Incubator mailing list and sent as an open letter to the community at large, details a myriad of reasons why Curtis does not think Oracle's proposal to donate the code to Apache should happen.
Curtis sees the question of failure for an Apache OpenOffice.org incubator as not just simple failure, but one that will potentially damage the entire Open Document Format ecosystem.
"If this podling fails, it could hurt the value of the OpenOffice brand, LibreOffice, waste resources (these emails are just the start), hurt Apache's reputation, etc. Some think this could finally end the GPL debate for this codebase. It has always had a proprietary extra clause. That is the clause that is being used to create this license," Curtis wrote.
So what are some of the objections?
There are technical considerations at issue. Curtis cited LibreOffice, the fork from the OpenOffice.org code now maintained by The Document Foundation, and its move towards more Python-based tools. OpenOffice.org, will remain focused on Java-based code, especially given the preferences of the ASF, Oracle, and the other very interested advocate of this proposal, IBM.
And there's the licensing. Under the proposal, the OpenOffice.org codebase will fall under the Apache Software License v2 (ASL). That non-copyleft license will create practical concerns for sharing with LibreOffice code. It also, in Curtis' opinion, could put future OpenOffice.org in a bit of a bind.
"Copyleft is compelling to small LibreOffice contributors. Do you really want to write [ASL] so that IBM can sell it?", Curtis asked.
That could be a likely scenario, as IBM continues to sell its ODF-using product, Lotus Symphony. Having OpenOffice.org under the cover of the ASL is a much better option for IBM, which historically prefers non-copyleft licenses.
"Copyleft," for those new to the term, refers to software licenses, such as the GNU General Public License (GPL), which require source code to be released with binary software code. Non-copyleft licenses, like the ASL or the BSD licenses, have no such requirement. This is why the latter class of licenses are often labeled as permissive licenses.
Curtis also sees a big problem in manpower, speculating that already the amount of contributors working on the OpenOffice.org code now has been inflated, and he can't envision more than just a few volunteers trickling in to OpenOffice.org. Volunteers the LibreOffice team could use right now.
According to Curtis, "LibreOffice is a young community, easily confused and frightened. They barely know this name 'LibreOffice.' Meanwhile LibreOffice needs and would love to have another 10-whatever people."
That impassioned plea has a whiff of condescension and seems counter to other statements Curtis made in his blog, where LibreOffice should easily be able to "recruit in bulk" from any OpenOffice.org community that forms.
As disjointed as some of Curtis' letter seemed to be, he has codified a lot of the reasons why many people are objecting to the incubator proposal. Community sentiment seems to be with him. Curtis cited a recent statement from a Document Foundation member that reported LibreOffice and Document Foundation contributions had shot up since Oracle's proposal announcement last week.
It will be interesting to see how this plays out. Apache and the projects it oversees have typically been held in high regard within the broader community, and I'm sure the Apache reputation was a nice bonus for Oracle and IBM when they thought of this plan.
But community unrest is building, and not even the Apache seal of approval may be enough to prevent more bad blood between OpenOffice.org and LibreOffice.