LWN.net Weekly Edition for October 14, 2004

Novell waves its patent weapons

Novell got off to a bit of an awkward start with the free software community; since then, the company has missed few opportunities to state its support for the community's goals - and to back up those words with actions. Releasing iFolder and Ximian Connector and jumping into the SCO fray are a few examples of note. Now Novell has posted a policy statement describing how it will respond to patent attacks on free software. This policy may not be all that the community might ask for, but, if Novell lives up to its words, the community may have just gained a new, potent ally in the patent battle.

So what is Novell saying? The company makes its purpose clear at the beginning:

In other words, Novell wants to make the world safe for Novell products - and their customers. Yes, this is a selfish motivation, but one should not forget that this is a corporation we are talking about here. The important point is that Novell sees litigious patent holders as a threat to its interests, and is responding in the hope of heading them off.

Here is the stick intended to deter possible attackers:

It is a sort of intellectual property mutual assured destruction policy: if you deploy your patent weapons in a way which threatens Novell's interests, Novell will respond with "highly relevant" weapons of its own.

This promise is worth something, for a couple of reasons. The first is that it is credible: Novell has truly committed itself to Linux, and is indisputably threatened by anybody who brings threats against Linux or its users. The company's own interests will compel it to respond to such threats.

The other notable point here is that a threat against almost any package shipped in the SUSE Linux distribution is a threat against Novell. The announcement for SUSE Linux Professional 9.2 claims over 3500 packages. So, while Novell has not committed itself to defending any free software project, especially if Novell customers have not been directly threatened, the fact remains that the company must be prepared to step in and defend a large number of projects if its promise to its own customers is to remain credible. Anybody who considers launching an attack against any of those 3500 packages will have to include a possible response from Novell in their calculations. The patent threat, while still very real, has just gotten a little bit less scary.

There is one thing which Novell did not say, however: nothing in the posted policy commits the company to not using its own patents to attack a competing free software project. We asked Novell about whether the company would make an IBM-style "no first use" declaration; we got this response back from company PR Director Bruce Lowry:

That is good stuff, and what one would have expected to hear. But it would have been nice if Novell's patent policy contained an explicit promise not to attack free software with patents.

This point leads into another thing which is absent from Novell's patent policy: any sort of commitment to work toward reform of the patent system. The simple fact is that Novell, like IBM and others, appears to be happy with the patent system itself. Novell has acquired enough "highly relevant" patents to be confident in its ability to fend off attacks from others. Having gotten into a position where just about anybody in the industry is probably infringing upon at least one of its patents, Novell has no particular motivation to drop its weapons. Such is the nature of the U.S. patent system; at least those weapons are, for now, deployed in the defense of free software.

Comments (5 posted)

Reports from the GNOME summit

The distributed development model works very well for the open source community, but sometimes there's just no substitute for putting people together in a room to work on a project. The GNOME Summit held this past weekend in Boston did just that with 50 to 60 GNOME developers.

Since we were unable to attend in person, we did the next best thing and got the skinny on the Summit from two of the attendees, Luis Villa and Owen Taylor, both members of the GNOME Foundation Board. Villa said that about half of the scheduled time at the Summit was devoted to hacking and that a big focus of the Summit was to "get the juices flowing again, not listen to someone pound through PowerPoint slides."

Despite the heavy developer attendance, Taylor told us that the topic that drew the most interest was marketing. Villa said that there were three sessions on marketing, and that the group had come up with good ideas on what kinds of people they should be marketing to, and how to talk to those target markets. Villa mentioned that it was very important to market not just to users, but also to ISVs and developers to try to get those groups to develop products using the GNOME platform. Villa mentioned that GNOME hasn't always done the best at marketing its product, noting that other projects have gotten more press coverage for the same features:

For those interested, Villa's notes on marketing are posted to his website. It looks like the marketing discussions at the Summit have also spurred interest in reviving the GNOME marketing list.

Taylor led a session at the Summit on next generation rendering for GNOME based on Cairo and new technologies coming out of X.org

Villa said some of the discussions covered usability, integration with X.org, and "administrative stuff" including a possible move away from CVS for the GNOME project. Taylor said there were also good discussions on hardware integration, control center reorganization and D-BUS. Since only a small number of GNOME developers were at the Summit, Villa said there was "a lot of discussion about the directions the project will be taking" but concrete decisions will be deferred to until the discussions can be taken to the GNOME lists.

We were hoping that the Summit would provide a clear picture of what to expect in the next release of GNOME, but Taylor said it's really too early to say what features will be in GNOME 2.10:

Villa also said it would be hard to predict exactly what would be in the next release, but did throw out a few hints:

One feature that was heavily discussed at the conference that might be in the next release is Beagle. The Beagle project, not yet officially part of GNOME, is a tool for indexing various forms of data, including mail, web pages, Instant Messaging, and integrating search into the desktop.

Villa compared Beagle to Apple's Spotlight and the search technology that is reported to be in Microsoft's "Longhorn" release. Villa says the name doesn't have any specific significance, except that "it's about sniffing out things, finding things." Villa also told LWN that Beagle isn't tied to "official" GNOME applications, and will work with a variety of applications. "If you only talk to the official GNOME browser, mail client, you're locking out a lot of people. This approach is a little more flexible."

Readers interested in following Beagle development can turn to the Planet Beagle blog.

Both Taylor and Villa said that the Summit was a success. Taylor noted that he was happy to be able to pull in 50 or 60 developers when the Summit was announced just a few weeks in advance of the event:

Villa also mentioned that the Stata building where the Summit was held was "an incredible place to gather," and the photos from the event certainly support that. Links to photos from the conference can be found on the Summit website.

Comments (none posted)

Buying the kernel

Long-time LWN readers have seen their share of, um, "interesting" Jeff Merkey quotes over the years. Mr. Merkey worked at Novell, but left to form the Timpanogas Research Group, which, at times, intended to sell "virtual network disk" technology, the Ute-Linux distribution, and a Netware-like kernel called MANOS. The company spent vast amounts of money in litigation with Novell, and was ahead of the curve in the indemnification game:

Mr. Merkey claimed to have disposed of the Novell issue by means of having filed a sexual harassment suit against the company, but life was not to be so easy. The closure of Timpanogas was announced in 2001:

One would think that Mr. Merkey would have had enough intellectual property litigation for one life, but that appears to not be the case. He recently resurfaced on linux-kernel with this interesting offer:

The offer has spawned a number of side conversations on what an insultingly inadequate offer $50,000 really is. Certainly any number of companies would jump at the chance to pick up a non-GPL version of the kernel at that price. But such discussions - and the offer itself - miss the real point.

Unlike many other large free software projects, the kernel does not require any sort of copyright assignment from contributors. Those who get code merged into the kernel retain their copyrights on that code. As a result, the kernel has hundreds - if not thousands - of copyright holders. Getting them all to agree on a licensing change would be a challenging task. Simply finding them all is likely to be beyond just about anybody's capabilities.

Critics of the kernel's organization claim that the lack of copyright assignment exposes the kernel to legal claims. They also state that the absence of a single copyright holder makes it difficult to enforce the GPL against those who fail to respect its terms. In response, one can point out that a copyright assignment would have been unlikely to deter the SCO Group from its campaign against IBM, and that the Netfilter team has been doing an admirable job of copyright enforcement.

What widely distributed copyright ownership does do, however, is to make a relicensing of the code impractical, if not impossible. We need not worry that Linus will someday succumb to temptation and sell out the kernel. Some developers are suspicious of OSDL, but none fear that it will start selling off private versions of the kernel to well-heeled companies. For all that some people like to compare certain distributors with Microsoft, those distributors will never get into a position where they are shipping proprietary Linux kernels.

Given this context, one wonders what Mr. Merkey thought he would be able to accomplish. There is no risk of him being able to buy himself a GPL exception for the kernel. The structure of the kernel's ownership is such that taking it private is not a practical possibility. This discussion is done; we must confess, however, to a certain curiosity about what Mr. Merkey's next scheme will be.

Comments (34 posted)

Page editor: Jonathan Corbet

Inside this week's LWN.net Weekly Edition

• Security: Security technologies which could be deployed now; New vulnerabilities in cyrus-sasl, mysql, squid, ...
• Kernel: Approaches to realtime Linux; Four-level page tables; InfiniBand.
• Distributions: A First Look at Specifix Linux; StartCom MultiMedia Edition; Ubuntu 4.10 RC
• Development: The Nvu Web Authoring System, new versions of ZODB, GLib, liboggz, Quixote, QjackCtl, KDE, jPOS, wxWidgets, GStreamer, Gnumeric, Epiphany, Dowser, Latex Beamer Class, Parrot, GHex, Treebeard.
• Press: Benefits of open-sourcing, the other side of software patents, Linus interview, Netscape DevEdge shut down, Linux vs Mac OS X, INDUCE act is dead, the Open Publication License, OfB Choice Awards.
• Announcements: SGI's Linux graphics system, Eclipse BIRT, Australian OSDC, CodeCon 2005 CFP, LAC05 CFP, YAPC::AU::2004 registration, Tuxme.com, cost estimate for Linux.
• Letters: Popularity and security

Next page: Security>>