An attic for LibreOffice Online
The Document Foundation is, of course, the entity charged with furthering
the development of LibreOffice and related software. LOOL, which allows
collaborative editing of documents, has been under development for a decade
or so; that work picked up in 2015 when TDF announced that LOOL would be "developed
into a state of the art cloud application
". Over the years, though,
the relationship between TDF and Collabora, the company doing the bulk of
the development work for LOOL, began to sour. In mid-2020, the company
complained that it could not bring in
revenue to support LibreOffice developers when everything was available for
free; among other things, Collabora wanted a clearer path toward making
money on LOOL. Despite efforts on all sides to find a solution, Collabora stopped working on LOOL later that year,
choosing instead to continue to develop that code outside the TDF under its
"Collabora Online" brand.
The elephant in the attic
Since then, development on LOOL has come to a halt; without Collabora's contributions, there wasn't much to sustain an ongoing project. That left TDF with a body of abandoned code with no maintenance or credible prospects for future development. Not wanting to emulate Apache OpenOffice, TDF started looking for a solution to this problem; the answer was the attic proposal. Code that TDF cannot support would be "atticized" — moved to a read-only repository — available for anybody who wanted to do something with it, but not a target for ongoing development. The proposal also included a process for "un-atticization" should enough developers demonstrate interest (in the form of patches) in carrying the project forward.
The response to the proposal was swift, but few participants wanted to talk about the attic itself; instead, there was a fair amount of rehashing of the LOOL story and discussion of what should be done with that code now. Not everybody feels that LOOL should be moved into any sort of attic. Marco Marinello, for example, replied that what was really needed was a new set of agreements with TDF's member companies:
I have already said this many times but I want to repeat it: it has to be clear (and hopefully stated by legal contracts) to the companies working in the LibreOffice ecosystem that they cannot wake up one day and bring their development outside the LibreOffice project. They cannot stay with one foot inside the ecosystem, contributing to it, and with the other one bringing their development effort outside.
Just-elected TDF board member (and Collabora employee) Jan Holesovsky answered that a different approach was called for:
From my point of view, rather than legal contracts, a much better strategy is to listen to what the ecosystem companies or other contributors are telling you; work with them, instead of against them; treat them as partners, not as enemies. If you do that, there is no reason for anybody to leave the community, move the code away, or fork.
The fork of LOOL, Holesovsky continued, was the result of a failure to listen to contributors (and Collabora in particular), especially when it came to distributing a binary version of LOOL under the LibreOffice brand. Board member Paolo Vecchi responded with a somewhat different view of the events surrounding LOOL. The proposal to publish binary LOOL builds, he said, was meant to help schools and nonprofits dealing with COVID-related disruptions, which is well within TDF's public-benefit mission. He, too, called for agreements to be made with companies working on TDF projects:
It is totally fine if someone wants to start his/her own projects based on LibreOffice and host them under his/her rules. However, I don't think it is fine to benefit from TDF and the work of its community for years, and then change the rules and walk away.
He later added
that, in his opinion, companies making major contributions to TDF projects
should be made to agree to a "backporting agreement
" for a
one-year period; Behrens disagreed,
saying that such a requirement would kill any willingness to bring projects
into TDF at all.
On January 9, Marinello posted
a counterproposal to putting LOOL in an attic. Under this plan, TDF
would claim
that Collabora Online was really "always LibreOffice Online
"
and provide daily builds of LibreOffice integrated with the Collabora Online
code. Others
would be given the right to redistribute these builds under the LOOL name.
Unsurprisingly, outgoing board member (and Collabora employee) Michael
Meeks disliked
this idea, characterizing it as "let's de-fund the
developers
". Focusing on Collabora Online, he said, has led to
"a new and better model for everyone
". While it would be a
good thing if TDF could drive sales for its members, that "has been
repeatedly shown to be structurally impossible for TDF
"; he
suggested that was not an entirely bad thing. Meanwhile others, including
Vecchi
and Behrens
expressed opposition to redistributing Collabora Online under the TDF umbrella.
Keeping this from happening again
The discussion went on for quite some time, with numerous people trying to place the blame for the "loss" of LOOL according to their view of the events. For all that, though, it was a surprisingly polite conversation, most of the time. The participants seem to feel that, regardless who is to blame for what happened with LOOL, the important thing is to figure out how to keep TDF viable and avoid similar problems in the future.
From the point of view of at least one large commercial member of TDF, the solution is to make sure that member companies don't feel like they are funding an organization that is then sabotaging their business plans. So TDF would need to avoid taking actions that reduce the value of commercial offerings — actions like offering binary versions of LOOL. Otherwise, it looks like TDF is using the members' own donations to compete with them. Industry consortia often need to walk this sort of tightrope, but many TDF members don't seem to think of it as an industry consortium.
The opposing point of view is that member companies should not be able to use TDF's resources and support to build a product, only to remove it from TDF once it can stand alone. Supporters of this viewpoint would like to see additional agreements put in place with member companies to prevent this from happening, especially for projects where a single company dominates the development effort. TDF, to some members, is meant to provide a public benefit rather than services to corporate members, and they are not happy when they feel that this benefit has been compromised.
As it happens, TDF has just elected
a new board including many of the participants in this discussion. It
would seem that this board should have, at the top of its list of
priorities, coming up with a solution to this disagreement that can keep
all of the stakeholders, if not happy, at least out of a state of complete
disgruntlement. The January 14
board meeting, with both the new and old members, discussed the attic
proposal (and punted any action pending further discussion) but did not
address the bigger question; the agenda
for the January 28 meeting does not mention it at all. TDF is an
important institution in our community;
its governance needs to come up with a solution to this question.