LibreOffice: "It is wrong to blame Oracle"

Novell's Michael Meeks talks about the reasons for the fork, the first few weeks of the new project and plans for the future

derStandard.at: The criticism around the way Sun / Oracle has been handling OpenOffice.org is not entirely new, so why the split, why now?

Michael Meeks: First let me say: A lot of people blame Oracle and I don't think that's fair. Those decisions are actually taken inside of Star Division and if you see the history of Star Division and OpenOffice.org from Day 1 when they promised the foundation - well we've been waiting for that and the frustration got more intense over time. It was always like "We're going to do the foundation in a year", but it never happened.

Also there was always this idea that to make the foundation independent, you'd have to have this huge pile of money and this incredibly complicated organization to pay hundreds of developers and then every company would pay money into it. And this model makes no sense - it's no good for the companies, it's no good for the project to have a monolithic entity doing it all. And that colored the discussion and took the topic off the table every time this was discussed.

And then there was this great hope that when Oracle acquired Sun - because Oracle historically engaged well in lot's of open source communities like with Apache or the Linux kernel - that this expertise would be brought to Star Division and that we'd get a better product. But sadly that expectation - as yet - has not been fulfilled. They more or less left it alone and in this case it would have been better if they'd shown a more hands-on approach.

derStandard.at: But there's been nothing specific that Oracle did that triggered the split?

Michael Meeks: Not not at all. And I'm optimistic, the offer we gave to Oracle is genuine, we'd love to have them involved in a vendor-neutral way. It's not an attempt to attack them. And the success that we already see - it should be obvious to any business man that this is the right way to go. We don't want to hurt Oracles business, we just want to grow the project.

derStandard.at: You mentioned the early success for LibreOffice, could you put this in numbers?

Michael Meeks: So - it's been only three weeks since we started and we've seen a lot of new contributors popping up on our mailing list, people we've never even heard of before. At the moment we have 65 new contributors, two or three coming along every day and that's incredible. And in addition to that, we have, the 15 Novell people, the two Red Hat people, the Debian guys - all of those that have contributed continuously over the years.

derStandard.at: Those new contributors are those mostly developers or translators?

Michael Meeks: There are some translators in these numbers, but most of them are actually coders. I think what happened is: We require no copyright assignment to add your code and we also made the project vendor neutral. And that's attracting to a lot of developers - I mean it's not like figuring this out would be rocket science. Those are often people that had high hopes for OpenOffice.org and then their code died a slow death in a dark corner of the issue tracking system, in process and problems - and now it's easy.

We've had 2,2 million lines of patch since the beginning of the project, sure - most of them are not substantial changes, a lot of this is cleaning up the code, like removing german comments that have been lingering in there forever. Or removing libraries that have been needlessly duplicated. But if only 10 percent of those people stay and go on working on other stuff - that would be simply great.

derStandard.at: In the initial announcement for LibreOffice you were very cautious to not mention the word "fork", but I guess with Oracle declining to work together in the Document foundation, it's a real fork now, right?

Michael Meeks: Simon Phipps did a great quote on this, to loosely paraphrase him: If the community goes this way and the company stands still - who forked?". But remember: There's still a lot of sharing still, we're taking code from Oracle, there's a lot of mention of them in release notes.

derStandard.at: As the projects advance it's safe to assume that the code-bases of OpenOffice.org and LibreOffice are going to divert more an more which makes it increasingly difficult to backport Oracles code, do you have any plans how to handle it.

Michael Meeks: There's some truth in there, for the moment this hasn't been a problem, and I think in one year we'll just simply have enough developers ourselves, so that this won't be a problem.

derStandard.at: So you're going to go separate ways in this likely future?

Michael Meeks: I didn't say that. But if you look at the current situation, we're already going separate ways. If you look at something like the GStreamer-integration (which Novell initially did) it was rewritten by the Oracle guys - nominally from scratch though there were some striking similarities to our work.

derStandard.at: At the moment OpenOffice.org is licensed under the LGPL-v3 are you going to change something about this?

Michael Meeks: The license unfortunately is very difficult to change. One of the problems here is having an LGPL-v3 only license and the Free Software Foundation and others are recommending a Plus-license. So for new code we will use LGPL-v3+ with a Mozilla Public License dual-licensed - pretty much like the Firefox / Mozilla model.

derStandard.at: How are you going to handle OpenOffice.org extensions?

Michael Meeks: There are a bunch of problems here. For instance if we want to be compatible with OpenOffice.org extensions, we can't break the ABI, that's one thing I'm not a big fan of. I'm also not a big fan of bundling very useful core functionality in extensions, like it's now with the presentation viewer. If you look at the statistics nearly nobody downloaded this in relation to the people downloading OpenOffice.org, so it seems silly to me to take something that is useful to everyone giving a presentation and shoving it somewhere else. Stuff like this should just be in the general build - and that's what we are going to do.

derStandard.at: At the moment OpenOffice.org / LibreOffice uses its own toolkit with VCL, is this something you want to change?

Michael Meeks: There are lot of things that you also can do inside that toolkit, getting the layout to work better and such stuff. So I think VCL serves a purpose there. I think we need to improve the visual look but I don't think it makes sense to do a major toolkit switch right now.

derStandard.at: Why should someone who uses OpenOffice.org up until now switch to LibreOffice?

Michael Meeks: I think if you now use OpenOffice.org it's a logical step to move to LibreOffice - more features, more function and more freedom.

derStandard.at: Are you going to keep the dependency on Java, considering that this is a pretty heavy dependency and not a lot of the code is written in it?

Michael Meeks: Well there are some good things about Java, it's a cross-platform managed runtime and you can intercept it, so we can do interesting things with that and still ship one binary for every platform. On the other hand it adds compile time problems, it's bigger and it's certainly slower. So - I don't know yet. It obviously makes sense to keep the Java functionality around for people that use it, but maybe we'll do some more native code, use some tools to do migration stuff. We'll see.

derStandard.at: When are you going to deliver your first stable release?

Michael Meeks: In the next few weeks. This really depends on code quality and what our QA people say.

derStandard.at: I guess you are not going to follow OpenOffice.org release cycles forever, did you think about adopting a fixed release cycle?

Michael Meeks: Well - I think there is a huge benefit of having a fixed and predictable release cycle - at least for major releases it is very helpful when you are forced to stop working on features and concentrate on bug fixing. You have something like a six-month-cycle in OpenOffice.org already, but that's still delayed a lot, so this needs to be "more fixed".

(Andreas Proschofsky, derStandard.at, 16.11.10)

Share if you care