Issue No. 03 - May/June (2005 vol. 22)
DOI Bookmark: http://doi.ieeecomputersociety.org/10.1109/MS.2005.65
Inside the Software Patents Debate: Some Good News for Open Source Developers
Several high-profile court fights, such as the SCO Group's pursuit of IBM regarding disputed Unix copyrights, have made software patents a top debate topic in the open source community. Open source companies have had to address fears raised by detractors who argue that a developer could inadvertently use a piece of open source code protected by a third-party patent, exposing his or her company and customers to intellectual-property suits. The debate has been heated and time consuming—but not all bad, say some members of the Linux community.
"Three years ago, SCO was affecting open source in a negative way," says Jason Wacha, general counsel for MontaVista Software, a Linux development tool maker, and cofounder of OpenBar, a nonprofit group that educates about open source and intellectual-property law. "My answer now is the SCO lawsuit was a fantastic thing for open source. It caused people from engineers to lawyers to really address this issue. It prompted this huge desire to know more," he says. "In the long run, this helps open source." (For the OpenBar URL and other URLs related to this debate, see the sidebar.)
For open source developers and customers seeking advice as they navigate the complicated waters of intellectual-property law, more information is now available, according to Wacha. Also, a few long-simmering worries for open source seem to be cooling. SCO's chances of a lawsuit victory appear to be decreasing. IBM and Sun are sharing patents with the open source community, a move designed to spur innovation and offer protection. And a long-discussed, controversial software patent law in Europe has stalled.
Many people's education in the subject of intellectual-property law and open source started when SCO launched legal salvos in 2003. SCO mailed letters to its customers and other Fortune 1000 corporations using Linux technologies, referencing its intellectual property relating to Unix source code and requesting that the companies pay licensing fees to SCO.
In March 2004, SCO also sued customers DaimlerChrysler and AutoZone. The DaimlerChrysler case, based on contract issues, was dismissed. The AutoZone case was based on the theory that AutoZone infringed SCO's copyrights just by its use of Linux. The portion of the case relating to whether the Linux kernel violates SCO's copyright has been stayed. The judge permitted limited discovery relating to whether AutoZone infringed any SCO copyrights in the process of transitioning from Unix to Linux. That discovery was continuing as of April.
SCO might not be making much headway in its battle against IBM, either. The US$5 billion suit against IBM took a recent hit. US District Judge Dale Kimball wrote in a February ruling that it was "astonishing that SCO has not offered any competent evidence to create a disputed fact regarding whether IBM has infringed SCO's alleged copyrights through IBM's Linux activities."
The court cases continue, but intellectual-property lawyers watching the case believe that SCO's licensing prospects with regard to Linux seem to be diminishing. "Based on the tenor of the court rulings recently, they're fading as a threat," says Wacha.
It's important to note that SCO is making copyright and contract claims in these cases, not patent claims. Also, a dispute with Novell still exists over what Unix copyrights SCO actually owns, says Dan Ravicher, who heads the Public Patent Foundation, an advocacy group fighting for reform of the US patent system.
"Very few people are saying SCO has legitimate copyright claims against Linux," Ravicher says. "One problem we have in the US system is people without meritorious claims can harass you significantly, and those claims trigger large legal and financial obligations."
Litigation risks often misunderstood
The risks of intellectual-property suits related to the Linux kernel, which could affect everything from software to devices such as cell phones that run on Linux, are easily misunderstood, according to Ravicher. For instance, Ravicher says an August 2004 study that he conducted for Open Source Risk Management (OSRM), a consulting and insurance firm for companies concerned with intellectual-property claims, has been taken out of context at times.
The much-cited study found 283 software patents related to the Linux kernel that could potentially be used in suits. However, Ravicher says many people miss a key point: No software patent has been found by the US courts to be valid and infringing upon the Linux kernel. When a patent is challenged in the US court system, the court finds it invalid about 35 to 46 percent of the time, he notes.
A patent's validity is tested only if the defendant being sued challenges its validity. "Only about one percent of issued patents are litigated," Ravicher says. Another often-missed point: "If we were to study any other operating system, including Windows, we'd come up with at least as many patents, if not more," he claims.
OSRM, which oversaw the study, won't specifically list the 283 patents because of an aspect of US patent law. If you examine a software patent and are sued later, and a judge decides you have willfully violated that patent, you can be hit with triple damages and attorneys' fees. If you didn't know about the patent, you can't have willfully violated it.
One key to protecting yourself as a developer is to learn about intellectual-property law, even if you dislike the notion of software patents, advises James Harvey, a lawyer at Alston & Bird, the firm representing AutoZone in the SCO suit. You also need to sweat the details—for example, examining and really understanding the language in the licenses of software you're using, he says. "Don't just stick your head in the sand, and don't just copy things like license agreements and contribution agreements from SourceForge," Harvey recommends.
At large companies working with open source, developers can consult corporate lawyers when questions regarding intellectual property arise. Independent developers should use common sense, shouldn't ignore patents, and should read and comply with licenses in developing derivative code, Wacha advocates.
Some companies are also making a big deal about indemnifying their customers. This means the company promises to protect you if a third party sues you for unknowingly using software that infringes its intellectual property. HP and Novell are among the big Linux players who have made indemnification promises, although each company's promise details vary.
"Novell, HP, and others are confident in their intellectual-property situations and are offering protections to customers," Harvey says. "Where the open source community and where open source products have to get to is where they're no different from proprietary products. One way is offering intellectual-property protection competitive to proprietary products."
Proprietary software (even Microsoft Office) includes indemnification to the user as part of the price, Ravicher points out. The software vendor, rather than the user, shoulders the responsibility of fighting intellectual-property claims.
A January 2005 study by market research firm International Data Corporation found that 4 percent of US companies "have been faced with the threat of vendor legal action over intellectual-property infringement." The figure rises to 8 percent at firms with more than US$750 million in annual revenues.
Indemnification will soon become a matter of course for open source companies because customers are demanding it, says IDC's Stephen Graham.
But one problem is that customers are confused by the varying indemnification promises, Graham adds. "Apples-to-apples" comparisons of the indemnification protections are difficult to do, he says. Protections, for example, might require a maintenance contract or might mandate use of a specific company's hardware. Linux distributor Red Hat has addressed the issue by promising that if any of its code is found to be infringing intellectual property, it will replace that code for customers.
Indeed, the number of licenses continues to multiply beyond the well-known GNU General Public License. The Open Source Initiative, a nonprofit group that approves open source licenses, has certified more than 50 licenses to date. "License proliferation is a serious problem," says Ravicher.
"We're seeing more 'major' licenses, like Sun's, all the time," Wacha says. "That's confusing for people. In the proprietary world, the code is so controlled, you don't have the problem of developers trying to understand all the different licenses."
Even the GNU GPL is getting a makeover. The existing version, circa 1991, needs a complicated facelift to tackle patent and licensing questions, plus the problem of differing patent laws between countries. No target date is available for the new license's release, but it isn't expected in 2005.
"It'd be a nicer world if there were only 10 licenses," says Wacha. Whether the open source community will reduce, or at least halt, the number of license flavors remains to be seen.
Developers can take advantage of new resources that can help clarify intellectual-property law and license issues. For instance, the Open Source Development Labs recently put $4 million toward funding a new group, the Software Freedom Law Center. The center, which will give legal help and education to open source developers, is run by Eben Moglen, who also serves as general counsel to the Free Software Foundation.
The IBM effect
Some developers also took comfort in IBM's announcement that it will share some 500 of its patents with the open source community. What might the move really mean? Will it significantly impact the use of open source code?
"It's a reaffirmation that there's a community out there for you as an open source developer," Wacha says.
"It impacts the legal climate at a macro level quite positively," says Harvey, "because it lets open source developers know that IBM is committed in certain respects to not using patents to put open source developers out of business."
In Europe, where the open source community has been fighting a proposed software patent law for several years, the IBM move got mixed reviews, says Till Jaeger, a lawyer who specializes in software law and open source issues for JBB Rechtsanwälte. (In 1999, Jaeger cofounded the Institute for Legal Questions on Free and Open Source Software.) Did open source developers in Europe think IBM's grant was significant?
"This is rather disputed," Jaeger says. "Some see it as a good first step, while others judge the grant as a worthless marketing campaign to appease those who criticize the practice of registration of software patents."
Developments in Europe and Asia
The progress of the controversial law that sought to allow certain types of software patents in Europe had come to a halt as of February. The law, known as the European Union Directive on the Patentability of Computer-Implemented Inventions, has been winding its way though the European Parliament and the Council of the European Union for some time; the two groups have been unable to agree on provisions regarding software patents.
Open source groups such as the Foundation for a Free Information Infrastructure and Free Software Foundation Europe have been publicly fighting the law, saying it would hurt open source innovation. And a new effort called NoSoftwarePatents (NSP), being funded by Red Hat, the Swedish firm MySQL, and Germany's 1&1 software firm, is pushing the use of copyright law as opposed to patent law for software intellectual-property protection in Europe. But some large European firms have backed the proposed law, saying some software patents are necessary.
The law has trudged through Parliament, with its most recent opposition coming from Poland. In February, the European Parliament rejected it. At this point, it's not clear when or if a revised bill will surface, Jaeger says, but many small and medium-sized businesses are also fighting this law or any other software patent law expansion in Europe.
Meanwhile, in India, Minister of Industry Shri Kamal Nath grabbed the open source world's interest when he issued in early 2005 an ordinance modifying the Indian Patent Act in a way that opens the door to software patents in that country. The ordinance was prompted by patent law requirements from the World Trade Organization, with concern centering on intellectual-property protection for pharmaceutical companies. However, the ordinance could also affect technology firms.
Previously, the law stated that companies couldn't patent "a mathematical method or a business method or a computer program per se or algorithms." The modified law changed that language to "a computer program per se other than its technical application to industry or a combination with hardware, a mathematical or business method or algorithms."
Some groups such as the Delhi Science Forum, an advocacy group working with technology issues, fear this language could be interpreted broadly. However, the ordinance will expire in July 2005 unless India's parliament approves it, so the matter isn't cemented.
Japan's patent system is more like that of the US than those of other Asian countries, but in China, patents can be hard to find and sometimes conflict, says Wacha, who often works with Asian clients. Developers and software companies can be more confident about intellectual-property searches done in the US than in China, he says.
On the other hand, he says, Asian open source developers have had less to worry about. "People who have been legal threats to open source have focused on the United States," Wacha says. "In some sense, people are outside the bull's-eye if they're doing open source outside the US."
Cataloging History:Museum Database Helps Remember the Holocaust
The victims database at Yad Vashem, Jerusalem's Holocaust History Museum, exemplifies how advancements in software architecture can provide for cross-border and cross-cultural information exchange.
Launched in November 2004 at an estimated cost of US$22 million, the Central Database of Shoah Victims' Names ( www.yadvashem.org) currently holds information about more than three million Holocaust victims.
Three mechanisms are helping to grow the database. Survivors and family members are submitting information online. Staff members are inputting an array of historical information, originally in many different European languages. Finally, dozens of local initiatives involve gathering the names of Jews from a specific region, country, or camp. Standard searches on the database will cull information from all the sources.
Unique technical challenges
According to Michael Lieber, Yad Vashem's CIO, one of the biggest hurdles in developing the database was the lack of a common identification system such as social security or passport numbers for Holocaust victims. So, searches will often start with information about a victim that might be written in several different ways.
To account for these differences, the database's indexes include many versions of the same name in English and Hebrew and a series of variants related through phonetics, semantics, or association—also in both English and Hebrew.
"The name Isaac, for example, appears in our documentation in over 1,500 variations," Lieber explains. "Similarly, a search for 'Alexander' will also retrieve records containing the name Sasha, the Russian nickname [for Alexander]."
Because indexes aren't based on set algorithms but on careful screening of the variants, the database requires substantial human intervention to maximize its technological capabilities. Currently, 30 experts manually decipher the database's Pages of Testimony. They perform updates monthly, but plans are underway to have weekly updates.
Lieber admits that false matches have presented a problem. "The user interface has to deal with queries by people with varying levels of knowledge," he explains. "Some people have very accurate information while others have only vague details. As a result, some matches may be 'false' because not enough specific data has been entered. That said, many users have discovered information related to unknown relatives within the 'false' records."
In addition to coping with the multitude of variations in names and spelling, another challenge was providing for massive traffic. In the first month of operation alone, over three million users from more than 160 countries accessed the site. The average user performed seven queries per visit, which resulted in more than 21 million queries.
Because the database was intended to be accessible to as many users worldwide as possible, it was designed to be compatible with as many browsers as possible as well as with current accessibility standards.
"Our users range from high school students to people over 80," adds Lieber. "All texts may be enlarged, so that people with eyesight problems can gain access to all data; surprisingly, this is not a feature of the main accessibility standards."
"The user interface was designed to reflect a modern system, keeping in mind the subject matter," he says. "We decided to opt for a clean and simple design with a minimum of graphics. … This allowed us to deal relatively easily with the complex issues of browser compatibility and accessibility."
Lieber also says users can toggle between languages on individual pages.
"It is refreshing to hear about an application that has put user-friendliness above functionality," notes J.S.G. Wright, a UK-based software engineer and CEO of SymTech.
"The technology to do this has been around for some time. However, the typical implementation is for the user to select the language per visit, not per page. Being able to select the language on a page-by-page basis is a really neat solution," he remarks.