It was immediately obvious at the time, I remember it as such. The devs jumped ship to LibreOffice. All Oracle was left with was a brand and a moribund codebase, and for some reason Apache accepted it. Why didn't Oracle give it to LibreOffice? Because they were sore at being bypassed, it looked like. The correct thing for Apache to do would have been to say, nope mate, we don't want to be part of your childish feud, keep it, trash it, or give it to the people who are actually doing the work.
If Apache had not accepted it, Oracle was NOT going to donate it to TDF. Instead, Oracle would simply have shuttered it, or sold it to IBM. There was no workable plan ever that would have resulted in the OO code going to TDF.
Even if AOO had never done a single release, LO still benefited enormously from OO going to Apache, since once at Apache, there was known and public IP provenance, AND the code was relicensed to ALv2. This allowed LO to inherit that provenance, relicense their stuff on AOO as needed, and close what had been a possible nasty IP hole.
The Apache Foundation could simply have accepted the OpenOffice.org codebase+trademarks from Oracle, then donate to LibreOffice.
But they did not do so. The Apache Foundation does not have experience with software that end-users use. They have experience with server software. That was their undoing.
As I recall Sun and IBM had a contract cause OOo was used for Lotus Notes. Then Oracle bought Sun. Oracle had obligations to IBM from that contract. Oracle could not figure-out how to make money on OOo so wanted to get rid of it. Oracle offered it to IBM but IBM did not want it. Instead IBM suggested ASF take it on. So Oracle relicensed under ASL and transferred everything hence Apache OpenOffice was born. IBM provided most of the devs to the project at that point, even hired some Star and Sun folks for it.
While this was all happening LibreOffice formed from... what was it... Goo go or something like that, the folks that were behind the build system and distro patches that most people outside of Solaris were actually running. I have a suspicion that there were terms agreed to between IBM, Oracle, and ASF that prevented ASF from donating anything to anyone.
Now IBM has scaled back support and at this point it's an LGPL v APL religious war.
> I have a suspicion that there were terms agreed to between IBM, Oracle, and ASF that prevented ASF from donating anything to anyone.
There were no secret deals. The Apache Incubator PMC voted on a public proposal (<http://wiki.apache.org/incubator/OpenOfficeProposal>) which included both a community and the promise of assets. Those assets were later delivered as promised. The incubation process then proceeded as it has for scores of other projects. It would have been completely inappropriate, unthinkable, for the Incubator (which oversees incubating projects) or the ASF Board (which oversees all projects) to take the assets away from that nascent community.
At the time of the donation, The Document Foundation had an (understandably) antagonistic relationship with Oracle and legal foundations which were firming up but not yet solid. The assets were not going to TDF from Oracle directly.
In accepting the proposal for a new Apache OpenOffice project, the ASF accepted not only the assets from Oracle but also a community, distinct from the LibreOffice community, that wanted to work with those assets. That community had to be given a chance. Without all those things, the deal could not have been completed.
Even at the time of donation, LibreOffice had already gained a substantial set of the former Sun/Oracle developers: Stephan Bergmann (to Red Hat), Bjoern Michaelsen (to Canonical), Eike Rathke (to RedHat), Michael Stahl (to Red Hat).
Together with other former Sun developers like e.g. Thorsten Behrens and Caolán McNamara, LibreOffice had more of the old Sun developers than Apache OpenOffice had at any point in time. Incidentally, they also made a lot of contributions that allowed LibreOffice to strife past other derivatives: like build system cleanup, static code analysis and fuzzing, bibisecting.
What community? The only community around OpenOffice.org were the Sun/Oracle devs + the Go-OO devs getting it to work on Linux.
The Go-OO devs forked it into LibreOffice and then Oracle shuttered their own work and reassigned their devs to other projects. Only a tiny handful of devs not part of Go-OO remained, and they had not been instrumental in actually maintaining the codebase as these other two groups had.
In fact, Oracle didn't even have many committers on OOo at the time - they'd reassigned most of the StarDivision devs to the abortive CloudOffice. At the time of the fork, LO had more dev resources on the codebase than Oracle did.
Just getting the code-base released under the ALv2 was a win for the world at large. Even if we never pushed out a single release, at least the code is out there now, under a permissive license, where anybody can use it. That alone makes the whole exercise worthwhile, no matter what happens next.
It's not impossible that the terms Apache got it under prohibited them divesting themselves of it for some length of time.
I don't recall thinking at the time this landed that it was so cut-and-dry who would end up the superior solution.
In the worst case, though, would it be better for Apache to have refused holding the OOo branding, given that there was no chance it would have otherwise made it out into someone like TDF or otherwise's hands?
