Microsoft's accusation that the open-source software industry has infringed 235 Microsoft patents has spotlighted a difficult issue: how aggressively should a company police itself for patent violations?
Microsoft said it released the tally--though not the 235 specific patents--in an effort to bring open-source companies to the table to hammer out intellectual property licensing deals similar to the one struck by Linux seller Novell in 2006. But industry experts said the declaration's implicit demand--that companies with open-source software should figure out what Microsoft patents they're infringing and come to the negotiating table--is unrealistic at best.
In general, searching for potential software patent violations isn't practical, given the number, breadth and opacity of patents in the United States. Not only that, but it's at odds with Microsoft's own policy to wait for patent holders to get in touch with it if they think there's infringement.
In fact, searching for potential patent problems can actually leave a company financially exposed: if a lawsuit concludes a patent was infringed, a company or individual that knew about the potential infringement must pay triple the financial damages compared with an unknowing infringement.
"The fear of willfulness is so great that often firms instruct their engineers not to look at patents," said Matthew Schruers, senior counsel to a tech lobbying firm called the Computer and Communications Industry Association. Because of the willfulness issue, the expense of searching patents and the difficulty of actually understanding them, "It's gotten to the point where most software application developers cannot plausibly say they've conducted complete patent searches," he said. (The CCIA, although counting Microsoft among its clients, scorned its patent move on Monday as "very troubling.")
Though Microsoft has paid $1.4 billion in three years to license others' patents, the company indicated that it takes a more passive role when it comes to licensing others' patents.
"If a company believes we are using its intellectual property, they should come talk to us," the company said in a statement. It didn't respond Tuesday to questions about whether it has notified specific parties such as Red Hat or Linux kernel leader Linus Torvalds of any of the alleged infringements.
Microsoft said the Linux kernel infringes 42 Microsoft patents, but Torvalds is among those who refuses to investigate whether he's violating any.
"There are several reasons why engineers should not read other people's patents, only their own. And it's not a 'hide your head in the sand' issue, it's a very practical issue of it being a waste of time," Torvalds said.
"The fear of willfulness is so great that often firms instruct their engineers not to look at patents."
Computing and Communications Industry Association
For one thing, developing technology without looking at patents lets a person honestly say they developed that technology independently, which helps show that the patent in question doesn't meet the requirement of a technology not being obvious, he said. And engineers aren't likely to comprehend patents in the first place: "Unless you have a patent attorney at your side, patent language usually makes no sense."
He derided Microsoft for spreading FUD (fear, uncertainty and doubt) rather than tackling the issue forthrightly. "If Microsoft were to actually tell people what patents they claim we violate, we could either laugh in their face and show prior art, or just show them to be obvious, or we could do things differently," he said.
Some, including David Jenkins, an intellectual property attorney with Eckert Seamans, believe in a more active patent-hunt approach. For example, Motorola should probably look for patent issues in Linux before shipping a Linux-powered mobile phone, he said.
But doing so isn't easy, and "Most people should not attempt to perform a search," he said.
"Finding a patent, especially a software patent, on the U.S. Patent and Trademark Office Web site is very difficult," Jenkins said. "Almost nobody calls everything the same thing...Either you use a (search) term so broad that you get a return of 1,000 patents, or if you narrow it down, it's likely you're going to miss a lot."
Jenkins' firm charges about $1,000 for a patent infringement search, but the prices go up in cases where a search yields more patents that must be scrutinized.
And even Jenkins could think of only a single instance when a client went back to refresh an existing search with up-to-date results. The frequency of checking "depends on how litigious the patent owner is going to be, and are they going to catch you," he said.
Another issue complicating Microsoft's case is the widespread belief that patent infringement is the rule rather than the exception. "People are infringing other people's patents all the time and don't pay for it," said Mark Radcliffe, an intellectual property attorney with DLA Piper.
"By sitting in my chair right now I'm probably violating somebody's patent," adds Matt Asay, vice president of business development for open-source document-management firm Alfresco and a competitor with Microsoft's SharePoint software. Asay, also a lawyer, doesn't want to violate Microsoft's intellectual property rights, but he, too, said Microsoft needs to take the initiative of describing what patents Alfresco might be violating. "Until we know--and IBM knows Red Hat knows--what can we do?"
"By sitting in my chair right now I'm probably violating somebody's patent."
Microsoft wouldn't say whether it believes any of its products infringe patents held by other companies with which it doesn't have a licensing agreement. It also wouldn't say whether it requires its own programmers to check if their software infringes others' patents, whether the company routinely checks to see if its products infringe or whether it ensures its products don't infringe before shipping them.
Horatio Gutierrez, Microsoft's vice president of licensing, said in an interview that the alleged open-source infringement is "not accidental." As evidence that the infringement is intentional, the company points to a 2006 speech by Richard Stallman, who single-handedly built much of the intellectual and legal framework of the free and open-source software movements.
Stallman didn't come close to a detailed analysis of where problems might lie or even a definitive admission, though. In it, Stallman refers to a 2004 study funded by Open Source Risk Management, a start-up selling insurance in open-source intellectual-property matters. "Two years ago, a thorough study found that the kernel Linux infringed 283 different software patents, and that's just in the U.S. Of course, by now the number is probably different and might be higher," Stallman said.
The quality factor
There's a wide gap between being accused of infringing a patent and being found in a civil lawsuit to actually infringe. And a recent Supreme Court decision means the gap likely will be getting wider.
In a unanimous April decision, the court sought to set a higher standard for weeding out patents for obvious technology. "Granting patent protection to advances that would occur in the ordinary course without real innovation retards progress and may, in the case of patents combining previously known elements, deprive prior inventions of their value or utility," the court said in its opinion, a decision that could make it easier to challenge patents' validity or harder to obtain them in the first place.
Microsoft believes its patents are solid. "Our patent portfolio scores very high on patent quality and science linkage," the company said, pointing to its second-place ranking among computing companies by the Patent Board, a patent analysis firm.
The court's decision was welcome news to those who complain about the quality of software patents. Improving patent quality to ensure patents are "truly novel, truly not obvious and truly useful" is one major part of the CCIA's patent reform agenda.
Mike Dillon, Sun Microsystems' top lawyer, said his company was "hit with two new patent troll cases" in April but is cautiously optimistic about the effects of the Supreme Court's move. He called spurious patent suits a "tax on innovation."
Those sentiments were echoed by Dillon's colleague, Sun Chief Technology Officer Greg Papadopoulos. "Are software patents useful? My answer is, mostly no. Copyright appears to be mostly better for maximizing innovation while giving individual copyright holders the ability to modulate compensation and derived works."
Torvalds was more direct. "The bulk of all patents are crap," he said. "Spending time reading them is stupid. It's up to the patent owner to do so, and to enforce them."