OpenOffice.Org and the LibreOffice Imperative

As expected, the Apache Software Foundation took the first steps to admitting the project to the Apache community, following Oracle's IBM-designed proposal. It now faces a time of maturing and proving in Apache's Incubator, a...


As expected, the Apache Software Foundation took the first steps to admitting the project to the Apache community, following Oracle's IBM-designed proposal. It now faces a time of maturing and proving in Apache's Incubator, a period when user-facing development will most likely grind to a halt.

I've avoided publishing articles here during the Apache discussion as I have both a history (see below) and strong views (although I speak only for myself this time). But with the end of voting, it's time to document the story so far. First some links to set the scene:

  • Vote on accepting for incubation at Apache
    After the heaviest traffic in recent memory on the Apache Incubator list, with many impassioned messages on both sides of the argument (along with some heavy-handed slapdowns here and there by proposers of the action), the vote ran for 72 hours and as expected gained approval by a comfortable majority.
  • My +1 Vote
    I voted in favour of starting an incubator podling. I did this because I believe there is a strong future for a new project drawing code from and maintaining it for use by multiple projects, such as LibreOffice, RedOffice and Symphony. I believe it would be hard and misdirected work to attempt to build a full competitor to any one of those projects at Apache - a good explanation is on the mailing list.
  • An Invitation to Apache OpenOffice
    While the official statement on IBM's behalf is interesting, the discussion in the comments is far more enlightening. The statement itself completely ignores the existing ecosystem and speaks as if Apache is moving into a green field, but the comments reveal how the existing community feels about that as well as showing the intense and inexplicable antipathy IBM feels towards The Document Foundation in general and LibreOffice in particular.  Given even Apple ships GPL code in the Apple TV and elsewhere, I can't believe it's just antipathy to the licensing. Note especially the wise and balanced comments from Jeremy Allison (of the Samba project).
  • The FSF's statement is pretty balanced. I can't help thinking they were forced to comment after misinterpretations of their advice on licensing were forcefully asserted by the proposers of the Apache podling despite correction from an FSF Board member.
  • The Decline and Fall of
    There's much to agree with in this analysis of the move by Oracle - at IBM's direction - to offload the assets they own onto Apache. While I don't agree with the complete negativity of the final paragraph, as you'll see below, I do agree with most of the other things Bruce says here.
[ Disclosures: was in the portfolio I worked on at Sun. I am a member of The Document Foundation but have no responsibilities there. I am a member of (meaning I have an e-mail address). I'm a huge fan of Apache. I used to work at IBM. Everything written here is entirely my own, unsponsored opinion.]

Code Is Not Community

The best summary of the story so far that I can come up with here is that the move by IBM (enacted by Oracle) fails to realise that the assets - which Oracle inherited from Sun, where I used to work - and the Community are not the same thing.

While IBM was able to ask Oracle to assign the assets to Apache, it was impossible for the community to be assigned over as well since almost all of the people who didn't work for Sun - with one or two exceptions - had already made a decision where to work. The conflict the corporations have manufactured isn't primarily about licenses, or about the maturity of Foundations, or even about personalities. It's that the corporations just tried to make decisions about something they don't own - people.

Despite the best efforts of many, Sun never really allowed the community to fully spread its wings. Just looking at what that community did achieve - especially the amazing wealth of localised versions for languages spanning the entire globe - shows how much potential was also being held back. Documentation, localisation, marketing, extensions and more flowed from that stifled community in a way that would have been remarkable even for a fully liberated community.

When Oracle took over the assets of the project along with Sun - the copyrights and the trademarks - the community finally burst free and started a second project with the same code, called LibreOffice. The LibreOffice project took some time to self-organise, but is now cranking out great code at regular intervals. If you use GNU/Linux, there's a decent chance it's their code that you have on your desktop as it absorbed the Go-OO fork Novell started. Its host, The Document Foundation, is in the final stages of incorporating as an independent Foundation from its (non-profit) sponsor and creator, Freies Office Deutschland e.V., having raised ‚¬100,000 in just a few days from individuals to fund the move. The whole rehosting which the community has done looks pretty much a textbook success.

Positive Planning

What annoyed most of the people I've asked about IBM's play at Apache was the fact it ignored the existing community and that success. While there are a few people who wish LibreOffice dead, and a few more who would rather not say so but clearly would dance on the grave, the vast majority of community members have been pleasantly surprised by just how effective the project turned out to be. It's thus important when being constructive to take that into account.

While my vote counted for nothing officially, I voted in favour of the podling being established, apparently to the surprise of some. I hope that it will now secure the full source for from Oracle, irreversibly transfer the trademarks and then set to work creating common code that can be used by a wide range of downstream projects. Even if the podling were to fail to graduate, that achievement would be a big plus for us all.

If a re-usable reference implementation of ODF editors for each ODF sub-format can then be created from the code Oracle is relicensing and maintained at Apache, it would be immeasurably positive for everyone. Over time I'd hope LibreOffice, Symphony and the rest could incorporate that new work, since the strongest path to interoperability is by way of a clear and open specification with a shared open source reference implementation.  It's time to rise above the divisions and work together at the community level.

If on the other hand the Apache podling just turns into an opportunity for the known opponents of LibreOffice to attempt to compete just for the sake of it using the "" name, that will be a dark development for software freedom and I've expressed my disapproval strongly elsewhere. If that were the case, despite voting "+1" for the podling to be started, I would vote against it being promoted to full top-level status in Apache.

Practical Conclusion!

What does that mean for you? If you use GNU/Linux in almost any flavour, you're may already be using LibreOffice since some distributions already switched to it.  If not, or if you use Windows or a Mac, you should just ignore all this noise and leave things to sort themselves out at Apache. Then, in the mean time, you'll find the best code and the most innovations happening at LibreOffice - go there right now and download the latest version

Follow Simon as @webmink on Twitter and Identi.Ca

"Recommended For You"

Apache OpenOffice - Can it still compete with LibreOffice? LibreOffice 3.5 released by The Document Foundation