OSBC: The Microsoft/Novell panel [LWN.net]

The 2007 Open Source Business Conference featured a panel discussion on the question of whether the Microsoft/Novell agreement is good for open source or not. Your editor was asked to sit on this panel and try to represent the community's point of view - as if the community has a single point of view. The event is a bit of a blur - only partially caused by the beer your editor sought out to help with the recovery afterward - so this will not be a complete report. Hopefully it will still be useful.

Other members of the panel were Justin Steinman (Novell), Sam Ramji (Microsoft), and Alison Randal (O'Reilly). The panel was moderated by Doug Levin, the CEO of Black Duck Software. Mr. Levin did an admirable job of keeping this standing-room-only event on track and inclusive.

After the introductions, each panelist got to make some opening remarks. Your editor had worried considerably over this stage of the event, and had prepared the following statement. The actual words were not read from this text, however; the intent was substantially the same but the wording was generally different.

So is this agreement good for open source? For much of the agreement the answer is simple. More money for open source companies, used to develop more open source software, can only be a good thing. And our community has always been about interoperability with everybody; no complaints there.

The situation changes when we look at the patent side of the deal, however. Even if you accept that algorithms expressed in software are patentable - something much of the world does not accept - the current software patent situation in the U.S. is impossible to deal with. There are thousands of software patents covering the most basic techniques. You cannot write any non-trivial program without infringing upon an unknown number of patents, without having ever seen those patents. There is no way to know where they are until the tax collector shows up at the door.

It is hard not to see a certain amount of sincerity in Microsoft's recent statement that it will not go out filing patent infringement suits. Microsoft is arguably the largest victim of the U.S. software patent regime, having literally been hit up for billions of dollars. The company says that it is typically defending two dozen or so patent suits at any given time. But when Microsoft's CEO starts talking about how Linux users are carrying undisclosed liabilities on their balance sheets, it makes that sincerity hard to believe. That is a clear fear, uncertainty, and doubt attack. As a platform for this sort of FUD, the agreement for Novell is not a good thing for open source.

In the free software community, we are most careful about the provenance of our code. We do our own work; that is the only thing which lets us give that work away. Novell has now come and said that the free software it is distributing is not our own work, that it owes a debt to Microsoft, which wrote none of that software. The company's protestations that it has acknowledged no infringement ring hollow; Novell is paying Microsoft for something, and unless the company is willing to come out and say that its payments are simple protection money, it is paying for perceived patent infringements.

When a company in our community makes a statement that taxes are due to Microsoft for the use of our work, it makes it harder for others to resist that demand. It weakens our defense.

But the real reason why this agreement has taken such criticism from the community is deeper. We in the community are proud of our work. We have done it ourselves, and have not stolen anything from anybody. When a company like Novell tells me that my work was stolen from Microsoft, and that anybody using my work owes taxes to Microsoft, I cannot help but be deeply offended. When such a statement comes from inside our community, it's even worse. It feels like a betrayal of the trust which holds the development community together, it's a divisive thing. In that way, I think this agreement is not good for open source.

Justin Steinman's opening remarks mostly highlighted how the agreement has been good for Novell. Microsoft is now selling SUSE Linux into places it has never been before; in fact, Microsoft is now Novell's biggest single sales channel. Increasing Linux adoption is a good thing, he says, and this agreement has certainly served that end.

Allison Randal took the position that the agreement is mostly irrelevant for open source. It's just another boring industry partnership, with the usual sort of joint ventures. Had it been between Novell and IBM, she says, nobody would have noticed. Microsoft's patents are a problem, like so many other patents held by many others in the industry, many of which would come to our defense if Microsoft were to decide to start suing Linux users and start the "patent armageddon." Still, it would be nice, she suggested, if Microsoft were to join the Open Invention Network and help bring an end to this problem.

Sam Ramji talked mostly about interoperability. Microsoft sees the computing network of the future as being entirely heterogeneous and it wants to be a proper player in that environment. He reiterated the "we would rather license than litigate" line, noting that Microsoft spends about $100 million per year defending patent suits.

One of the questions from the audience had to do with the effect that GPLv3 will have on this agreement and potential patent litigation in general. Justin and Sam both declined to comment, saying that they saw no point in talking about a license which is still in a draft state. Allison pointed out that GPLv3 is unlikely to change much from the current discussion draft, but no more information was to be had from Microsoft or Novell.

There was also a comment to the effect that open source users are better served by adherence to standards than interoperability agreements. Adding support for Microsoft's OOXML format seems to be a particularly sensitive point. Justin responded that if he were to go into a company trying to sell a solution based on OpenOffice.org, and that solution could not handle Office 2007 files, he would be laughed out the door. So supporting Microsoft's formats is important, even if Novell's policy remains that OpenDocument is the format of choice. Sam noted that standards are great, but true interoperability requires a great deal of work and testing; this agreement is about getting Microsoft and Novell engineers together to do that work.

Perhaps the most surreal moment came in response to questions about why Microsoft is going out trumpeting its 235 patents if it does not intend to sue, and why it does not reveal the actual patents. Sam made the statement that the 235 patents came out as a response to requests for greater transparency in Microsoft's dealings on intellectual property issues - a response which did not achieve universal respect among the members of the audience. He did not want to address the question of revealing the patents, though - and he did not have to. Microsoft's lawyer in charge of open source issues just happened to be sitting in the front row; he sprung up and claimed that Microsoft doesn't reveal the patents because the administrative burden of doing so would be too high - a statement that The Register had much fun with the next day.

Allison made the point that the community really does not want to know about these patents. Once we have been put on notice that we are alleged to be infringing upon a specific patent, we must respond in one way or another.

A theme that came out a few times in the discussion is that there are voices within Microsoft arguing for greater participation with the open source community. There are people within Microsoft who understand the power of free software and who want the company to be a constructive force in the industry as it heads in that direction. Like any large company, Microsoft suffers from a certain amount of schizophrenia, with the result that different messages will be heard from different groups. But there is reason to hope that rational and friendly (though always competitive) thought will prevail.

The session ended with a show-of-hands vote on whether the audience thought the agreement was a good thing or not. There were approximately equal numbers of yes and no votes - but the bulk of the audience did not vote at all. In many minds, it seems, the jury is still out on whether this deal is good for open source or not.

(Log in to post comments)

Yes, listing 235 numbers is turr'ble hard work

Posted May 31, 2007 22:00 UTC (Thu) by giraffedata (subscriber, #1954) [Link]

He didn't say it would be too hard to publish the list. He said it would be too hard to respond to all the questions, challenges, lawsuits, etc. that would result.

He also said it is customary in the industry not to give specifics of the infringement you expect until you're ready to actually demand compensation.

Yes, listing 235 numbers is turr'ble hard work

Posted May 31, 2007 22:38 UTC (Thu) by felixfix (subscriber, #242) [Link]

It's as simple as a press release or even just a web page listing the patent numbers. Requests for licensing, if any, could be turned into a profit center, I am sure Microsoft knows how to do that.

Other requests; what would they be? Clarification of the patent? Refer them to the press release or web page. It's a very simple process.

Challenges, lawsuits, etc.; what would those be? Seems to me a tacit admission they have weak patents.

Industry custom; sorry. Physical products have patent numbers listed on the packaging and on the product itself. What makes software so special? Could it be that their patents are worthless for anything more than an initial stock jump from the announcement?

Yes, listing 235 numbers is turr'ble hard work

Posted Jun 1, 2007 3:25 UTC (Fri) by giraffedata (subscriber, #1954) [Link]

Industry custom; sorry. Physical products have patent numbers listed on the packaging and on the product itself. What makes software so special?

They don't label software with the patents it uses like they do physical products?

It doesn't matter, though, because this is not the industry custom the Microsoft lawyer referred to. He referred to a custom of not publicly announcing specifically which products you suspect of infringing which of your patents (presumably in spite of the fact that you publicly announce that you have such suspicions) until you're ready to do something about it. (I don't think anyone cares which Microsoft products use which Microsoft patents).