Daniel Weitzner is Chairperson of the World Wide Web Consortium's Patent Policy Working Group,the body that is deciding if it's okay to charge "Reasonable and Non-Discriminatory" (RAND) license fees for use of W3C-endorsed standards that are covered by patents or other trade restrictions. We sent him the questions you asked back on October 8. Here, at long last, are his answers.
Why should standards be for sale?
by Fleet Admiral Ackbar
My question is - There have traditionally been two types of standards. The first type is an agreed standard, such as the RFCs. The 'market' has no say, but there is a presumed compensation in the availability and usability of said standard. The second type is a 'de facto', or 'market' standard. This standard is decided by people voting with their checkbooks.
So, "we" get what "we" want, but we have no guarantees of availability, usability, or definability.
Doesn't the idea of charging to use the standards combine the worst features of both? Doesn't doing so severely compromise the respectability of the process?
Internet and Web standards (some might say specifications, but I won't quibble) have tended to be a mixed product of market-driven and agreement-driven selection. In both W3C and the IETF, specifications only advance when they show indications of being widely-deployed (there are exceptions) and generally agreed-to. Adding a fee for implementers would certainly have some impact on this process, but I'm not sure it's the one you point to.
You suggest that the potential to earn license fees on a particular standard would drive its adoption. It seems equally possible that standards for which one has to pay a fee would see low levels of deployment, which could prevent a spec from receving full standard status. The IETF, which has some experience with standards that have license fees associated with them, won't advance a spec too far on the standards track unless it's clear that license fees are reasonable 'enough' to allow broad implementation. After, a number of commenters on Slashdot and firstname.lastname@example.org have suggested that there would be no open source implementation of RAND standards at all. That would certainly have an impact on the level of deployment.
The Patent Policy WG draft proposed included two possible licensing modes:
(1) Royalty-Free: specs produced in this mode would have as a requirement that they be implementable royalty-free. If information about patents came to light that indicated the spec could not, in fact, be implemented RF, then the WG responsible would either have to take some action (design around, secure RF licenses, etc.) to make it RF, or switch status to a RAND group.
(2) RAND: RAND specs would have to be implementable on a RAND (which could include a zero royalty) basis.
The Patent Policy WG (PPWG) included RAND as an one of the two options for several reasons:
(a) There is evidence that in other technology arenas RAND standards work perfectly well. For example, CD players, TVs, phones, all come with significant patent royalty obligations for their producers, but these devices are nevertheless widely available at seemingly affordable prices.
(b) Disallowing technologies which may only be available for a fee (RAND terms) would deny the Web access to the best technology available.
(c) If W3C adopts an RF-only model, then many important Web technology standards may end up being developed at other standards bodies or industry consortia.
W3C is now in the process of assessing whether the RAND-RF balance we struck in the Last Call draft is the policy we should adopt. As chair of the PPWG, I can't predict the outcome, but can tell you that this issue is being seriously debated. When we put out another draft policy, the public will also have a chance to comment more (if that's possible :-) ) on this subject.
abuse of fees and rights?
How do you propose to police, if at all, the potential of abuse from the imposition of fees, so that the interests of all parties, especially simple users and developers are protected.
Examples of abuse that I would want to avoid include the Fiasco with the licensing fees for implementing .GIF files many years AFTER they became a standard, and the recent action by MS to enforce a two year upgrade cycle for use of their software in businesses, when many companies are much more comfortable with a three or four year upgrade cycle. (ZDNet has had several articles critical in the extreme of that action)
I would very much prefer the situation to be similar, say, to domain name registration, where many parties can provide that service.
One approach to abusive or unfair licensing practices would be to rescind the Recommendation. It's clear that having a W3C Recommendation that would require patented technology increases the value of the patent to the patent holder substantially. So, if the patent holder is not willing to license its patent in some terms (RF or RAND depending on the policy we adopt) then we should consider rescinding the Recommendation as it cannot promote interoperability. W3C is discussing a mechanism to do this.
I don't understand the DNS analogy. Yes, there is competition being introduced into the domain name registration space, but patents are, by their nature a monopoly. Requring either RF or RAND licensing, which is what the proposed policy does, is a mechanism to require that particular patents held by particular entities are available on at least some 'reasonable' terms. (See the answer to question #5 and #10, too.)
standards vs patents
It's always been my opinion that standards are something developed for the general free use of the public. Standards are often decided upon by organizations such as NIST, IEEE, academia, etc.
Patents on the other hand aren't standards in my opinion. Commercial companies have traditionally been the ones pushing for patents and when they obtain one, they charge users of it a license fee.
A lot of patents are very useful and deserve to be standards, but I feel they should have the fees removed before it can be called a 'standard.' The patent can still stand after something has been standardized, but I dont think people should be charged to use it.
What's your call on opinions like mine that are carried by many individuals here at slashdot?
To me, the fundamental question is: what approach to patents and standards will best serve the goals of building, maintaining and expanding the Web as a universal information space? That Web standards have been implemented royalty-free until how. I believe it's clear that the Web would not have the reach or the accessibility it does if we had to rely on the more traditional, patent-driven standards to work, but it's ultimately hard to predict what would happen if the world were different.
Some have asked, 'if it ain't broke why fix it?' But the relationship between patents and Web standards is increasingly broken. At W3C we've had several specs with threats of license fee demands. And, there are an increasing number of patents surfacing where the licensing intentions are not clear. So we have to take steps to provide a licensing environment that meets our goals rather than just relying on a series of historical accidents and well-placed good will that has worked in the early days of the Web. This means that we can't do nothing, we can't refuse to acknowledge the existence of software patents. We have to find ways to make sure that where there are patents that essential to implement a given Recommendation, that they are known and available on terms that make widespread, diverse implementations possible.
A bit naive?
In it's Response to Public Comments [on RAND], W3C states W3C takes no position on the public policy questions surrounding software patents. Isn't that statement at best naive? The Internet and Web were originally designed with the idea of free and open communication. Today, there are powerful forces that would like to see open communication closed down and the Web turned over entirely to commercial pursuits. If a RAND policy is adopted for Web standards, won't the next move by those commercial entites be to create as many propriatry standards as possible and force them on the entire Web community (using hammers such as DMCA), like it or not?
Here are the considerations that went into the scope of the W3C Patent Policy and why we don't take up the software patent policy debate:
First, we are a technical standards organization, not lobbyists or, for the most part, lawyers. We think that we are pretty good at developing standards, but don't view it as our mission to lobby to have laws changed. Therefore, we need a patent policy that deals with the present-day reality of software patents, which exist in a variety of forms in a variety of countries (not just the US) around the world.
Second, we are an international organization, so even if we succeeded at changing US patent law, or the law in another country, as long as any country has software patents we have to deal with them.
Your question seems to suggest that we are being 'naively' dragged along in a larger campaign to make the Web into a more closed, proprietary environment. I think that you are right to point out that the Net is subject to a greater range of regulation than it was five or ten years ago, but you're mistaken to suggest that the patent policy we proposed is in support of that trend. To the contrary, as I've said in question (3), even the policy as proposed in August would go a long way to make it more certain that core infrastructure standards can be implemented on a royalty-free basis. I agree that any Web standards that are legally required for all Web software (or even just key tools such as servers or browsers) should be implementable on an RF basis.
[See my answer to #12, too.]
Who will define 'reasonable'?
What's reasonable for me, or for any other small, independant developer, is probably not what's reasonable for, say, Adobe, Microsoft or Macromedia. So will they be allowed to pay "unreasonabley" low liscensing fees, or will I be asked to pay "unreasonably" high fees? It seems that "reasonable" is in the eye of the beholder.
Reasonable is measured in light of prevailing license rates paid in the market. Question number 10 does a nice job of illustrating how this might work.
Why follow other standards bodies?
In its Response to Public Comments, W3C made the following statment:
One of the strongest criticisms of organizations such as ISO and ITU is that they charge exhorbitant fees to even read, much less implement, their standards. Some government entities have even gone so far as to incorporate proprietary standards into laws, such that individuals have to right to read the laws that they can be jailed for violating (e.g. Veeck vs. City of Austin).4. Is RAND licensing common for bodies like W3C? Yes. A RAND license is common among standards organizations.
The Web and Internet have always been unique in having their standards freely available for review and implementation. Is it a good thing for W3C to seek to emulate the "closed source" standards bodies?
You're not exactly correct about all Internet standards (See my point about the role of RAND licensing in the IETF in #12) but I do think it's important to preserve the openness of W3C and IETF. Even the proposed policy, which would allow RF or RAND standards to be developed at W3C, is far more supportive of the RF tradition than ANY other standards body. The proposed policy provides a way for W3C to decide to do an standard on an RF basis and to hold ourselves to that goal, forcing the spec off the Recommendation track if non-RF patents are discovered and there is no way to avoid paying a fee.
Fork in Standards?
If the RAND patent proposal is adopted by the W3C, there has been much discussion in the open source/free software community about forking the standards process, to preserve patent-free web standards. This fork would create a new standards body as a competitor to the W3C.
1.Do you think such a fork is likely if the proposal passes?
2.If a fork does happen, what are the consequences for the web?
3.Should the W3C take the likelihood of a fork into account when considering this proposal?
1. If the proposed policy passes we would do some standards as Royalty-Free standards and some as RAND. I can certain imagine the free software folk would seek to develop alternate standards in those areas where we had RAND standards.
2. A complete fork, that is a competing set of standards for all W3C Recommendations whether RAND or RF, would be very destructive. I'm also not sure that it would be successful in the RF case, because there wouldn't be a practical reason to do it, for most open source projects, except on a boycott basis.
3. Under the proposed policy, I would say that we would have to take the possibility of a fork on particular Recommendation into account before we decided to go RAND.
Finally, you can have as many standards organizations as you like and there will still be patents out there. Forking Web standards development doesn't answer the question of what to do, given that patents are out there. W3C has been burned by this. The only responsible course we feel we can take is to create clarity through an explicit policy.
Can you define "non-discriminatory"?
Can you please define exactly what it means for licensing terms to be "non-discriminatory", since the definition has a huge bearing on open source software?
For example which, if any, of the following examples would be non-descriminatory?
A) Fixed per-unit licensing costs. Let's say, for example, $1.00 per unit shipped. While this would be non-descriminatory in some senses, it would also prohibit free software and hence be descriminatory in other ways.
B) Licensing fees only for non-free software. While this method would not descriminate against open source, it could be seen as "descriminating" against closed-source software since they would be the only ones paying licensing fees.
C) Percent-of-price fees. For example, 5% of the price of shipped units. This would seem to be fair since the same licensing terms would apply to everyone, but in practice it would be similar to B) and hence seen as descriminatory.
D) Free for everyone to use. This would seem to be the only license which cannot be viewed as descriminating against someone. On the other hand, I'm fairly certain that this is not what is intended when the W3C uses the term "RAND".
Any insight into this would be greatly appreciated, I'm sure
The general understanding of non-discrimination in US law is that 'similarly situated parties are treated similarly.' Consider an example from a different legal realm, shipping. Most common carriers (like railroads and truckers) were historically required to charge non-discriminatory shipping rates. So, if you charge Customer A $1000 to ship 2 tons of freight 500 miles, then you must charge Customer B who comes along $1000, too.
If Customer B's load is actually harder to ship for some reason (it takes up more space in your truck, or requires special handling) then you can charge B more without being accused of discriminating. But if you charge B more than A simply because you like A better, or because B might be competing against you in some other business, then that's discriminatory.
The Patent Policy Working Group has on its agenda to consider the precise interaction between both the RF and RAND license models in the draft Patent Policy on the various flavors of Open Source licenses. So, I don't have an answer to that question yet, but it will be answered by the time we propose another draft policy. (I'll take your examples to the PPWG, too.)
The W3C must realize that ANY fee no, matter how reasonable they may seem, would have the effect of not allowing free and open source developers to participate.
I am sure that you are aware of the fact that the internet as we all know and love/hate it was developed by free and open source software.
Certain commercial developers, have been trying to convert the internet from, "...the greatest democratizing medium..." (ACLU vs. Reno) into yet another commercial morass where they can reap the maximum profit from the unsuspecting populous.
Given that, why would the W3C even concieve of a proposition that favors corporate purse strings over the greater good of the common man, and the internet itself?
We have seen governments try to regulate what we can say (CDA, CDA II, COPA), corporations what we can do (DMCA, SSSCA, DRM) and even ICANN with the warped sense of justice embodied in its UDRP. Does the common internet user now have to be leery of standards bodies like the W3C, which appears to favor corporate interests?
You're inviting an answer here of the form 'trust us'. That's really not for me to say, though, it's for the community at-large to decide through individual actions. Will independent developers continue to implement our standards, comment on our draft specs, etc?
These questions are going to be answered over time, based on the way our final patent policy looks and how it actually works. I do hope that the independent developer community has reason to continue to work with us, if not trust us. I hope that in the wake of the 2500+ comments we've received that people have recognized that we have responded in a way that seeks more input (by extending the comment period), involved recognized open source leaders (Bruce Perens and Eben Moglen were added to the group charged with developing the policy, and more may be added, too).
Most important, the W3C Process requires that we put out regular public drafts for review and that there will be another Last Call draft before we finalize a policy.
Your reference to Reno v. ACLU does touch a nerve for me, though. I was one of the lawyers who organized the First Amendment challenge to the CDA, resulting in the US Supreme Court overturning the law and enshrining wonderful quotes such as the one about democracy that you cited, into US law.
I'm reminded of 1991 when the US Congress allowed the Internet (at the time the NSFNet) to be opened to commercial traffic. Many advocates of the status quo (research, education, and mostly-benign Usenet chatter) predicted terrible results from allowing commercial content and commercial providers on to the Net. They, those academics and government folk who had the run of the Net in the 80s, felt that the Net should remain a public, non-commercial resource, run by the (US) government.
The switch to commercial use did happen and it's clear that we are all the better for it. Without that, most of us would not be online and there would be no Web, no Freenet, no Napster, and only closed, proprietary email systems like the old Prodigy. The point of this historical digression is NOT that we should accept RAND standards as part of the inevitable commercialization of the Web. Rather, we should recognize that the Net that we know and love will change and can survive change if we handle it wisely.
I chair a JEDEC committee. JEDEC's legal office reserves the right to define "reasonable" in our RAND clause.
Since all of our standards apply to hardware with well-understood costs of manufacture, the royalty structure most often adopted is "fixed percent of OEM price." This is a necessity in a market where semiconductor prices drop with Moore's Law, and a trivial royalty one day dominates the price the next.
In software terms, the same terms would be "fixed percent of sale price." The astute reader will observe that this allows continued distribution of free (gratis) software w/o royalty encumberance, even if it causes RMS to break out in a rash.
The alternatives, as we've already seen, can be as extreme as one-time charges in the millions of dollars. For Microsoft, able to amortize a flat-fee across hundreds of millions of units, the royalty would indeed be "reasonable." For Joe Startup, it would be prohibitive.
So, my question: what prospects do we have that RAND will be clarified, preferably with guidelines such as "percent of ASP," but at least with exclusion of abominations such as flat-fee?
So far we've decided that we do not have a good mechanism within W3C for assessing or defining reasonableness. Generally that's done in private negotiations or sorted out by a court. We've come to that conclusion because there do not seem to be well-established models for Web software royalties, as there are in your hardware examples. If we adopt a policy that includes a RAND component, and if we discover that the lack of definition is a problem, I am sure that we would reconsider.
What about patents that only exist in the US
If the w3c want to use patented technology in standards, what are they going to do about patents that only exist in the promiscuous US patent system and would never be granted in other countries. Or even the other way around. Won't that turn the world wide web into a us only web or at least split it into lots of separate entities?
The problem you bring up will be one for the patent holder (not W3C, which does not hold, nor plans to hold any patents). Patents can generally only be enforced over implementations sold or distributed in the country where the patent is issued. So, if a company has a US patent that is essential to a W3C Recommendation, then it could only seek royalties for products/services distributed in the US.
International and Third World ramifications
Many parts of the Third World are attempting to join the information age. Typically they are strapped for resources, and so far the ability to get on the internet with minimum barriers to entry has been key for them. The availability of free software has been essential. Not just free for use, but free for them to enter the development process, both to meet local needs and to build their technical base.
Doesn't patent encumbrance of W3C standards constitute another barrier to entry for poorer nations?
Even if we are willing to give them all copies of some proprietary web browser, aren't we still standing in the way of their developing their own technical expertise, because they won't have access to and use of the source code.
Doesn't it come across just a little like (Pardon the inflammatory language, but it exaggerates the argument appropriately.) fat Western capitalist pigs trying to keep the Third World down?
W3C has worked very hard to be sure that our Recommendations meet internationalization requirements to that the World Wide Web lives up to its name. And, we put a lot of energy into reaching out to countries whose technology sectors are in early stages of development. So while I don't necessarily accept that nefarious intentions in the flaming part of your question, I do think that we have to pay special attention to the result that you predict.
We need open standards
People from the W3C have "acknowledged" that the Internet's growth has been due to open standards.
This isn't not even half the story. The Internet would not exist without open standards promoted by bodies like the IETF and W3C (until now).
In the 1980's and early 1990's there were a number of network protocols in use: DECNet, VINES, NETBIOS/NETBEUI (shudder), IPX/SPX, SNA, and more. None of them initially would have been as scalable as TCP/IP. However, if any of them had been truly open, it might have been possible to fix them.
But none of those other protocols were open... and where are they now? Nowhere.
It's the same situation for hypertext protocols. People and companies have proposed substantial improvments onto existing protocols. A notable example of this is Hyper-G, which was then commercialized by Hyperwave.com. It fixes a lot of problems with navigation, and stuff like broken links. However, there was never a free and open implementation, and so it has languished in obscurity for the last 5 years.
My question to the W3C is this: Do they have any evidence that proprietary protocols will foster continued growth of the Internet and the applications that run on top of it?
Here are two answers:
[snide] What evidence? The IETF, since at least October 1996 when RFC 2026 went into effect, explicitly allowed standards track specs to have RAND encumbrances. The question really is: has the IETF has been irrelevant, useless, and closed since 1996? I don't think that it has, but you may have another view. [/snide]
[more_serious] There's no question that core protocols and document formats have a far greater chance of being widely deployed and well-designed if they can be widely implemented, without payment of royalties. The question is whether that means that every spec W3C does has to be RF? [/more_serious]
Make up of the RAND committee?
I would like to ask why this committee is so heavily biased towards the legal side in its makeup. Of course there should be lawyers there, since it deals with legal issues. But these issues have a tremendous impact on technology --in particular on its free exchange. Yet the members come from the IP departments of huge corporations, with little visible representation of technical people from the free software world.
The W3C is setting up a standard which will potentially harm many free software projects, yet is not giving that side any true voice in the drafting committee. I find it hard to believe that this was not a deliberate decision driven by corporate interests.
This attitude is sad, short sighted, and at the very least ungrateful. It avoids acknowledging the extent to which the free software world has contributed to the very existence and success of the internet as we know it.
In early October we announced that we have invited Bruce Perens (OSI co-founder) and Eben Moglen (FSF General Counsel) to join the Patent Policy Working Group as 'invited experts' with full participation privileges. They've already begun working with the group and I'm really pleased with what they've been able to add so far.
We should have done this earlier. We will, in the process going forward be sure to learn from this lesson and the process going forward will be more open. What's important though is that the final policy will certainly reflect their participation.
Documents for reference:
- W3C Patent Policy Framework Working Draft - 16 August 2001
- Backgrounder for W3C Patent Policy Framework - 20 August 2001
- Response to Public Comments on the W3C PPF WD
- Patent Policy Working Group Face-to-Face Meeting Summary