The office suite has occupied a very strange position in the world of open source. As a key software tool used by practically everyone on a daily basis, it was vital for free software to be able to offer one. And yet what came to be the leading office suite – OpenOffice.org – was widely recognised as deeply unsatisfactory. Its early versions were barely usable, and even in its later incarnations it was hard to get enthusiastic about it.
That was largely a function of the way that it had come into being, starting as the closed-source application StarOffice, and then being open-sourced by Sun, which had bought the product, largely in an attempt to irritate Microsoft. Licensing issues meant that OpenOffice.org never really became a true community project. As a result, there was no real passion behind its development, and it showed.
Things were made even worse when Oracle bought Sun. It soon became clear that the former cared even less than the latter about making OpenOffice.org a vibrant and successful open source project, and the announcement of The Document Foundation and the new LibreOffice – effectively a fork – was probably inevitable at this point. Here's its stated mission:
The Document Foundation has the mission of facilitating the evolution of the LibreOffice Community into a new, open, independent, and meritocratic organization over the next few months. An independent foundation is a better reflection of the values of our contributors, users and supporters, and will enable a more effective, efficient and transparent community. TDF will protect past investments by building on the achievements of the first decade with OpenOffice.org, will encourage wide participation within the community, and will co-ordinate activity across the community.
Oracle's unhelpfulness was further emphasised when, instead of letting LibreOffice become the official incarnation of OpenOffice.org, it gave the code's name and copyright to The Apache Software Foundation. That meant it would be released under the Apache licence, which implies a rather different dynamic from the LGPL that LibreOffice had chosen
, not least because Apache-licensed code requires copyright assignment non-exclusive licence from coders, and LibreOffice doesn't.
A couple of weeks ago, The Document Foundation celebrated its first anniversary. A blog post summarised some of the things that had been happening in that time:
"we have 136 members who have been nominated for their contributions to the project; we have some 270 developers and 270 localizers (although we always want to attract more), many of whom are also members; we have over 100 mailing lists, with over 15,000 subscribers, half of whom receive all our announcements; and there have been thousands of articles in the media worldwide".
LibreOffice is the result of the combined activity of 330 contributors – including former OpenOffice.org developers – having made more than 25,000 commits. The developer community is well balanced between company-sponsored contributors and independent community volunteers: SUSE and community volunteers new to the project have provided around 25% each of the commits, with a further 20% coming from RedHat and another 20% coming from the OpenOffice.org code base. The remaining commits came from pre-TDF contributors, Canonical developers, and organizations like Bobiciel, CodeThink, Lanedo, SIL, and Tata Consultancy Services.
"Thanks to a very welcoming attitude to newcomers, to the copyleft license, and to the fact that it is not requesting any copyright assignment, The Document Foundation has attracted more developers with commits in the first year than the OpenOffice.org project in the first decade", says Norbert Thiebaud, a first-day hacker who jumped on LibreOffice code on September 28, 2010, and is now a member of TDF Engineering Steering Committee.
Downloads since January 25, 2011, the day of availability of the first stable release, have just exceeded 6 million from 81 TDF mirrors, and amount to 7.5 million when you add external sites (like Softpedia) offering the same package. In addition, there are many more users who install LibreOffice from a CD burned from the ISO images available online or bundled with a magazine. TDF estimates that there are 10 million users worldwide having installed from downloads and CDs. Over 90% of those are on Windows, with another 5% on MacOS.
That suggests the LibreOffice project is coming along nicely, but leaves open the question about future developments – a crucial issue if it is to move on from its past. Those were addressed at the recently-concluded LibreOffice Conference, which I attended in part (disclosure: The Document Foundation paid for my travel and accommodation expenses.)
As well as a general air of enthusiasm that was evident among conference participants, there were some important plans revealed:
LibreOffice Online Prototype: you can watch a demo video at the following address:
http://people.gnome.org/~michael/data/2011-10-10-lool-demo.webm. LibreOffice Online is based on GTK+ framework and HTML5€²s canvas, and has been developed by SUSE's Michael Meeks, built on GTK+ broadway from RedHat's Alex Laarson.
LibreOffice port project to Android and iOS, based on the voluntary work of Tor Lillqvist, a SUSE finnish developer know for having ported GIMP to Windows. The LibreOffice Android and iOS port has the objective of bringing the office suite to iPads and Android tablets, and eventually smaller devices. The user interface work has yet to start in earnest but the bulk of the code is compiling.
It was emphasised that both of these are at very early stages, and that users should not be expecting to use them any time soon. But I don't think that really matters. What impressed me most about these plans were the fact that they existed at all. They bespeak an ambition that was sadly lacking during the old OpenOffice.org days. Certainly, there's no guarantee that the LibreOffice teams will manage to bring all three projects to quick and successful fruition, but at least they're trying – and that's a hugely important sign.
There were also some more concrete announcements at the conference:
500.000 desktops, mostly Windows, at several French Government entities switching from OpenOffice to LibreOffice (this increases the Windows installed base of LibreOffice by 5% in a single move)
800.000 USB keys with LibreOffice and other free software distributed to students of the Paris Region (ÃŽle-de-France)
Region ÃŽle-de-France becoming a member of TDF Advisory Board
As this shows, the ÃŽle-de-France is a big supporter of free software – I do wish someone in the UK (Boris Johnson?) would take up the great idea of giving away USB sticks with free software to encourage young people.
LibreOffice is evidently thriving; so what about OpenOffice.org? Not much has happened yet, but here's a very interesting statement from The Apache Software Foundation on the subject:
More than 70 project Committers are actively collaborating to ensure that the future of the OpenOffice.org code base and community are in alignment with The Apache Way. The project's extensive plans include assessing the elements necessary to update a product that hasn't had an official release in nearly a year; parts of the product's functionality encumbered by non-Apache-Licensed components; and a code base that has been forked and maintained by a community pursuing market dominance. As such, it is critical that we remain pragmatic about the project's next steps during this transition phase.
We understand that stakeholders of a project with a 10+ year history – be they former product managers or casual users – may be unfamiliar with The Apache Way and question its methods. Those following the project's migration to process and culture unique to the Apache community may challenge the future sustainability of the project.
Such concerns are not atypical with the incubation of Open Source projects with well-established communities – the successful graduation of Apache Subversion and Apache SpamAssassin, among others, are proof that The Apache Way works.
What's fascinating here is the repeated emphasis of "The Apache Way". This underlines the fact that the split between LibreOffice and OpenOffice.org is now ideological. On the one hand, the LGPL-licensed LibreOffice code is about preserving the freedom to share by ensuring that the code is never enclosed; while the Apache-licensed OpenOffice.org believes that the freedom to enclose is also important, and ultimately leads to a healthier, more vibrant ecosystem.
This means that the we will have a perfect opportunity to observe these contrasting models evolve starting from the same code base. It will be interesting to see what differences and similarities there are in terms of the people and companies that lend a hand.
But this does not mean that the two projects will be at each other's throats. As The Apache Software Foundation statement goes on to say:
We congratulate the LibreOffice community on their success over their inaugural year and wish them luck in their future endeavors. We look forward to opening up the dialogue between Open Document Format-oriented communities to deepen understanding and cease the unwarranted spread of misinformation.
We welcome input and participation in the form of constructive contributions to Apache OpenOffice.org. There are myriad ways to help, from code development and documentation to community relations and "help desk" forums support to licensing and localization, and more.
The way to move this forward is via the ASF, which owns the OpenOffice.org trademark and official code base. This is our chance to be able to pull together our talents towards a cohesive goal and protect the project's ecosystem.
At a minimum, we owe that to the hundreds of millions of users of OpenOffice.org.
Leaving aside the rather catty remark about "the unwarranted spread of misinformation", there is a crucially important point here. At the heart of both LibreOffice and Apache OpenOffice lies the Open Document Format (ODF). Whatever other divergences there may be between the projects, ODF will always be the common thread that binds them. And that's important, because open source is partly about choice, and the ability to move seamlessly between LibreOffice and Apache OpenOffice is absolutely key. Without it, both of them would necessarily be weaker and less important; with it, they can be friendly rivals working towards a common goal: widespread use of truly open standards for documents.