NEW YORK -- A leader of the Open Source movement recently offered sharp words about legal battles in the community, ranging from SCO's fight for code control to what he calls patent abuse. Bruce Perens also gave an update on several of his prime projects: the Desktop Linux Consortium and UserLinux.
Perens, who very publicly left Hewlett-Packard in 2002 to work in Open Source, gave his second biannual "Open Source State of the Union" talk at LinuxWorld Expo here last week. Much of the talk focused on the ongoing SCO legal case, which Perens calls laughable. He also discussed the danger that software patents present to Free Software and Open Source.
A Battle for Linux?
The SCO case looms large over the Open Source community, Perens acknowledges. In a nutshell: SCO, which owns intellectual property relating to Unix, has filed a multi-billion-dollar lawsuit against IBM, alleging that the firm contributed to Linux some proprietary Unix code belonging to SCO (and licensed to IBM). The claim has met with derision from nearly everyone involved in the development of Linux, including Linux author Linus Torvalds himself.
For its part, SCO has been largely unwilling to detail exactly which parts of the Linux source code allegedly infringe on its copyright. The case began nearly a year ago, and yet, Perens says, the community has seen only a list of 13 code files, though SCO promises more.
One of those files defines error messages that the Linux kernel can return. Both the SCO version and the Linux version of the files follow a decades-old Unix approach for error messages that has been made public in open standards at least three times over the years, Perens says. He notes that Torvalds would not have needed access to SCO's code to write the Linux files.
Perens calls SCO's other public claims equally laughable. He suspects the company is pursuing the legal action because it needs the money, pointing out that before SCO filed suit against IBM, its stock was trading at around the 50-cent level. It now trades at around $15.
"SCO is engaging in fraud," Perens says. "The particular kind of fraud is software piracy, which is claiming as your own the copyrighted work of another. This is all intended to kite the value of their stock. They will take their profit and exit before they lose the court case."
Even if SCO wins in court, the Open Source community can respond, Perens adds.
"If SCO wins, we'll have a list of files that the court found infringement in. We would take that list of files, and we would say to our families, girlfriends, etc., 'We'll see you in a month.' And in that month, we would just replace--cleanly--whatever facilities we have to replace. There are thousands of us...we can do all of this in a month. Where would this leave SCO? With a claim on software we would not be using any longer."
Software Patents: Danger?
But the next, larger trouble for Open Source is patents, Perens warns.
He calls patent-granting out of control, saying half of all software patents should not have been granted, because they do not detail actual inventions. Frequently, they only detail a specific way to make a computer do something entirely obvious using well-known methods, Perens says.
This is not the aim of a patent, but it occurs because patent office researchers lack the time to conduct a proper search for prior art, Perens says. Too often, companies obtain patents and wait for another developer to trip over their wire. "It costs $2.5 million on each side for a typical patent fight. A settlement is cheaper," he says.
Dodgy patents are even a thorn in Microsoft's side, Perens notes.
"Take the Eolas case," he says. "Eolas claims to have invented a technique [that is] obviously older than the date of their patent. Microsoft is stuck having to pay damages."
He expects Microsoft will appeal, but wonders whether Microsoft might launch a patent offensive against Linux someday if it finds itself in trouble competing on a level playing field. The company certainly has the war chest to buy up all sorts of patents that would cause trouble for the Open Source community, Perens warns.
Two projects Perens helped launch in 2003 are geared toward encouraging new Linux deployments.
The Desktop Linux Consortium comprises all firms with a stake in the success of Linux on the desktop. Members include distributors like Mandrake and Xandros, as well as TransGaming Technologies, which creates software enabling Linux machines to run popular Windows-based games. The Consortium first met in November, and plans to ramp up its desktop Linux efforts in 2004.
UserLinux is a different sort of project. Its goal is to create a new, enterprise-ready distribution of Linux for distribution by independent service providers around the globe.
As a sample application, Perens describes a dentist office that wants to replace four PCs with networked Linux boxes running software specifically tailored to the office's needs. The operation is too small to draw the attention of major Linux vendors such as Red Hat or IBM; but under UserLinux, the dentist could call on a smart geek in the area who designs such systems.
"I want to have a UserLinux provider who specializes in dentists in Minnesota," Perens says with a grin, describing the granularity of the project. Ultimately, the strategy behind UserLinux, as Perens describes it, is identical to that of the Open Source movement: It's all about a distributed community using shared resources (such as the UserLinux software) to accomplish its mission. In this case, the mission is to provide superior, supported tech solutions--and to make money doing it.