A better way to set Web standards

Here's how to set the standards for Web services and yet allow inventors to be paid for their patents.
Written by David Coursey, Contributor
commentary How should industry standards be set? Should the companies that develop those standards be allowed to profit from them? If so, who'll determine what those profits should be?

Those are some of the fundamental questions underlying the issue I raised, reacting to a report by my colleague David Berlind of TechUpdate. David and I may disagree on the particulars (and he knows the nitty-gritty details of this stuff much better than I do), but we share the same basic concern about patented "standards" being used to block innovation or control the marketplace.

STANDARDS come in two flavors: "official," meaning some organization has endorsed it, and "de facto," which means the marketplace made it a standard without the imprimatur of a standards body.

The HTTP protocol used by Web browsers is an official industry standard, sanctioned by the World Wide Web Consortium. Windows and Java are de facto industry standards, sanctioned by success in the marketplace.

Official standards, at least on the Internet, are almost always available for free, available to anyone to use without payment of royalties. But standards bodies can be too democratic, giving your competitors a chance to water down your technology to protect their own, and way too slow.

That's why IBM and Microsoft want some of their key Web-services protocols to become official standards without putting them into the public domain. The two companies want to assert their patent rights on these protocols and promise to make them available to other companies on the basis of "reasonable and non-discriminatory" (or RAND) licensing.

Most people recognize "non-discriminatory" when they see it, but "reasonable" can be harder to define. Especially when IBM makes a huge amount of money ($2 billion annually is the estimate I've heard) licensing its patents, and a federal judge has found Microsoft guilty of jerking around its OEMs over its licensing of Windows. Both are reasons to be wary of adopting any IBM/Microsoft "standards" that have potential licensing fees attached.

OF COURSE, Microsoft says there are many reasons why RAND licensing makes sense. (I haven't spoken to IBM.) First, it helps prevent other companies from patenting the same technology. Second, it gives the licensee leverage: You want to use my technology? Then give me a license to use yours. There are other reasons, none of them related to using patents for competitive advantage. The point being that choosing RAND licensing over royalty-free licensing isn't evidence that IBM and Microsoft have ill intent.

Given the highly competitive marketplace, it's hard to blame IBM and Microsoft for not wanting to expose their new standards to abuse by Sun and others--especially when whatever IBM and Microsoft both support is likely to become a de facto standard.

I think IBM and Microsoft should go ahead with their RAND licensing plans. But I also think they need to be responsive to a marketplace that has fair reason to question their intentions. Microsoft and IBM should explain exactly what they mean by "reasonable and non-discriminatory." They might also allow an official standards body to serve as the licensing custodian--setting terms and prices--while the two companies retain actual ownership and control over the technology.

That compromise approach could be messy. But it could also help move Web services more quickly into the marketplace, while making sure the marketplace isn't seriously hurt by two of its largest participants.

In fact, this approach could form the basis for a new licensing model, in which the standards-making process is protected, but companies still profit from and retain substantial control over the technologies they invent. But before we accept a new standard or standards process, let's get "reasonable and non-discriminatory" defined.

Do you think IBM and Microsoft should be allowed to set Web-services standards on their own? Or should they have to submit their standards to the Web's usual standards-setting process? TalkBack to me.

Editorial standards