The Novell-Microsoft Deal



[I originally drafted this as a short--short!--blog right after the Novell announcement; and then it was to be an OpenOffice.org response to the announcement. The tone was cautious; you can probably gather as much from the content. Since then, a lot of others have weighed in and Novell's position has evolved. In fact, it continues to evolve, and seemingly in a positive direction. See, for example, Groklaw of 15 November.]

The agreement announced Thursday night 2 Nov. in the US between Novell and Microsoft has been widely greeted by journalists as astonishing. It even had the (intended?) effect of chasing Oracle's announcements into the shadows. No one expected the Novell announcement and the wildest metaphors have been flying. Some have argued that Novell has "sold out" (Pamela Jones in Groklaw) or has made a bargain that may help it in the short run but not in the long and that can only help Microsoft (Asay) or has been foolish and seriously harmed the Linux community by playing the patent card (Phipps). Few outside of Novell or Microsoft have been sanguine about the deal, though Stuart Cohen of OSDL did give it a thumbsup .

I can't speak for the entire OpenOffice.org community, and Novell is part of that community, though it maintains its own, separate mail lists and repository; these are details. It does contribute the code it works on to the main repository and has promoted the project at events, like our recent OOoCon, and Michael Meeks, Novell's lead developer on OpenOffice.org, is a passionate believer in FOSS, community, and OpenOffice.org. He is also a member of the OpenOffice.org Engineering Steering Committee. All of which is to say that I respect Novell and Michael and when I disagree with their strategy, I do so respectfully. But there are points about the deal that make me cautious in seeing this as a "win-win." To be sure, I can see Microsoft winning from this deal, which benefits it rather obviously. It will be able to take its virtualization solution to market faster, as well as occupy a strategic position in enterprise Linux development and deployment--and I'm not even counting the ramifications to OpenDocument and OpenOffice.org development, which I touch on below.

Who can doubt that the patent deal will have a chilling effect on the development of LInux and OOo, which is also seemingly covered? Novell customers are indemnified, as are Novell developers and unpaid contributors; but what if, as Simon points out, the customer changes her mind and opts for another distribution? And what about developers or contributors for other Linux distributors? Those outside of the agreement's scope are effectively threatened. And it's not even that good for Novell, as the point of Linux development and of FOSS development in general has been to work intermurally, free from divisive threats, implied or not. This agreement seriously threatens that freedom.

The community, after all, chooses its own allegiances and finds patents anathema. It's thus hardly surprising that the Samba Team, which develops tools enabling Linux to work with Windows file and print servers, and which is leading the effort for real interoperability between the two systems, denounced the agreement on the 13th. Their point: "using patents as competitive tools is not acceptable," and they called upon Novell to work with Moglen's Software Freedom Law Center. As Steven J. Vaughan-Nichols reported , Novell is already doing so, though it's unclear how those discussions will affect things.

But my chief concern lies with the implications for OpenOffice.org and the ODF. According to the agreement and to Novell's Michael Meeks' own accounts, in his blog, the agreement is only positive, especially for OpenOffice.org. As he writes, "my hope is over the long haul: better interop, more bodies hacking on OO.o, wider penetration of (Novell's) OpenOffice into the enterprise, and more individuals able to boldly hack on Free software." As he sees it, under this agreement, OOo will flourish as interoperability improves. People will choose OOo because it is free software, presumably and may even have nice attributes in addition to its file format, but he doesn't specify. Of course, under the agreement, customers will preferentially choose Novell's OpenOffice.org because it is indemnified against (patent) suit; other flavours of OpenOffice.org will be at risk. And what about the ODF? ODF will benefit less because it is less important ("Open-Standard' of ~700 pages written by a small team over a short period"); the important thing here is "staggering value that is found in Free software," in particular, OpenOffice.org, which could use just about any file format--even, though Michael doesn't say it, OpenXML.

One the one hand, I agree with Michael: I see OpenOffice.og as more important than the file format, though we would probably differ in why it is important. But on the other, I disagree with his characterization of the ODF--he makes it seem hacked together an unlike Microsoft's massive OpenXML--and with the emphasis on making OpenXML the privileged format. ODF has been in development since 1999 and represents thousands of man-hours of work; it has been built from the ground up with the idea of openness. ODF has been furthermore a group effort from the beginning and is supported by numerous bodies, both private and public. It is not a file created by and for a single company now desirous of getting the validation a standardizing body provides in order to retain its customer base. Indeed, OpenXML is so completely specified it would seem to limit implementations. I thus agree with Bob Sutor here: it's only a matter of time, in this logic, before we see OpenXML as the default format for Novell's OpenOffice.org.

Of course, such an eventuality would constitute a real divergence and probably be a fork, which, I have to believe, no one, especially Michael, wants. So let's imagine that won't occur. And what about OpenXML? I am hardly opposed to interoperability and have long touted OpenOffice.org's current excellent interoperability, and believe we must have equally good interoperability with OpenXML But that implies to me bidirectional interoperability (which Michael does mention) and implies equally that Microsoft should work with the community in general and not with a specific company to accomplish real, standards-based, interoperability. It does not mean that OpenOffice.org, though it is still a minority application, must slave itself to MS Office or that interoperability must ignore standards. Doing so limits the sustainability of the effort; using standards makes it easier to sustain the effort. What is at stake--our intellectual wealth--demands real sustainability.

Unlike OpenXML, the ODF is relatively easily implementable; it is also, I have been told, more open to emendation and addition. There are no encumbrances attached to it, either technological or political. It is a format that lends itself the the creativity of both the proprietary and open community. The office suite, the office productivity suite--the whatever you want to call it--is a kind of legacy artifact. It is merely a way of producing, as opposed to consuming, knowledge and more generally information. The world is turning to more interesting ways of producing and capturing and representing information, and the Web and all it offers is central. OpenOffice.org will evolve to take advantage of this future and to lead it; to weave something that goes well beyond what we imagine now (who would have thought of mashups five years ago?) and to create the tools that will allow users to produce freely and easily--and be able to consume their and others' productions well into the future, using any of a number of standards-based applications. That's the promise of open standards: no vendor lock in, no hidden surprises, and the freedom to focus on what's important: what you and others are doing. And the ODF is up to the task. But by privileging OpenXML, much of that exciting future is foreclosed.

Michael expressed in his blog the expectation that more developers would come to OpenOffice.org because of this deal and that it would be overall good for FOSS. But FOSS depends on freedom--from patent anxiety, to be sure--but also from vendor lock in. Novell should reconsider the deal; as we have seen, it is not too late.


Posted: Mon - November 13, 2006 at 09:21 PM          

©