How can something be a de jure standard and still not be open and freely available? When it's license isn't Royalty Free/Reciprocal.
Thanks to ZDNET for inspiring me to write up my thoughts on this subject.
SOAP, WSDL and UDDI are all RAND standards which require companies to get multiple licenses from multiple sources in order to use them. In each case these specifications are standards for ancient technologies that have no real current value outside of being a standard. Apparently taking old ideas and encoding them in XML somehow makes them 'new' again.
Two or three mega corporations are pushing RAND in their standards because they are big enough that they will likely receive more money from RAND then they will pay out. Much more importantly then the money, RAND will place these companies in the position of having everyone, everywhere come to them for permission to implement (but they promise to be fair and non-discriminatory).
The way to stop RAND is for companies to commit themselves to Royalty Free/Reciprocal (RFR) licensing for standards. These licenses allow the use of a standard in return for not suing to enforce patents related to the standard. But to make RFR real companies must be willing to create alternate versions of RAND standards. This is a relatively small price to pay to keep RAND standards from taking over.
Working group charter allows for use of RAND. Needs to be watched carefully to ensure RFR.
W3C will probably handle, needs a strict RFR charter.
W3C probably won't touch, perhaps OASIS?
The key is to not target any particular company or companies. We have to leave an easy path for RAND companies to become RFR without losing face. We must make sure not to repeat the Liberty experience.
Everyone is excited about the new web service technologies like XML, SOAP, WSDL, UDDI, etc. What many people don't realize is that SOAP, WSDL and UDDI are all only available through Reasonable and Non-Discriminatory (RAND) licenses. They are not royalty free as XML, HTTP, TCP, IP, FTP, SMTP, etc. are. If you are using any of these technologies today and have not arranged for licenses with the appropriate companies then you are in violation of their intellectual property rights. It is unclear what sort of terms/royalties these companies will demand but at minimum you have to get their permission to use the technology on their terms. But hey, they promise to be reasonable and non-discriminatory.
In the table below I list the specific companies you must get licenses from in order to 'safely' use SOAP or WSDL.
Companies that claim RAND
UDDI licenses everything as RAND but doesn't make it clear if you have to get a license from UDDI directly or from each member company individually.
What makes RAND especially galling is that every one of these specifications use ancient, well understood technologies. Apparently taking old ideas and recycling them in XML somehow makes them new. This isn't to dispute that the various companies have patents on the various technologies, the world patent system is infamous for its willingness to allow just about anything to be patented. But until now the gentlemen's agreement has been that software patents are only to be used for defensive purposes. That is, if someone sues you then you can sue them back. Think of it as a 'no first use' policy. RAND violates that agreement.
Introducing RAND starts an ugly arms race in which companies fight to get on as many standards as possible in order to collect more RAND then they have to pay out. Of course only two or three companies are big enough to have a reasonable shot at winning that race. Strangely enough, these companies are on just about every new web standard out there.
For everyone else RAND is a fool's game. Paying real money for old technology is just silly. But worse, it weakens everyone to these two or three company's benefit.
The key to victory is embodied in three words: Royalty Free/Reciprocal (RFR). RFR means that the technology is available royalty free but only if the user accepts that they can not sue the issuer for patent violations in relation to the technology. This is not a blanket disarmament such asproposed by the League for Programming Freedom. Rather this is a targeted disarmament for a particular technology. A good example of a RFR license is HP's license terms in WSDL:
- HP will make available to anyone implementing the Recommendation a royalty-free license to any HP intellectual property rights that are essential to using the technology described in this contribution, for the purpose of implementing the Recommendation. One condition of this license shall be the party's agreement to not assert patent rights against HP and other companies for their implementation of the W3C Recommendation.
To make RFR real it is necessary for companies to be willing to do two things:
Publicly declare their adherence to RFR.
Dedicate real resources to writing alternate standards for SOAP, WSDL, UDDI, etc. that are RFR.
Simply complaining about SOAP/WSDL/UDDI/etc. won't achieve anything. That is why point #2 is so important. What is needed are alternate standards. Given how old and crusty the technologies used in these standards are it isn't technically difficult to write alternative standards with the same features but without RAND terms.
The W3C is attempting something along these lines in the SOAP working group which has declared itself to be royalty free. But the actual charter language allows for the use of RAND if the working group so chooses. Careful watch must be kept to ensure the final result is RFR.
No similar movements exists for WSDL/UDDI. Now is the time for companies to step up and put forward real resources. The short term costs are relatively small but the long term consequences are enormous.
It is true that patent holders could still sue. But for the companies currently demanding RAND this is a game of mutually assured destruction. Even the biggest companies with their enormous patent portfolios don't have patents on everything so they are in violation of innumerable patents held by others. If they should start suing they will set off a shooting war that will hurt them as much as anyone else.
As such there is a road to victory but only if a critical mass of companies are willing to commit resources to RFR alternatives to RAND standards.
The Liberty Alliance Project was formed to respond to Microsoft and AOL's efforts to create unified identity systems. While Microsoft and AOL were both willing to share their technologies neither was willing to make it open and licensing terms were still a question mark. A number of companies with something to lose if identity became the sole property of one or two companies decided to get together and create their own standard. The entire effort was created, funded and pushed by Sun. But very quickly the effort took off on its own. Today Liberty is funded solely by its members and Sun has just one vote like all the other board members.
Unfortunately Sun's CEO Scott McNealy and Jonathan Schwartz, chief strategy officer for Sun and Sun's board member to Liberty, did their best to anger Microsoft. While this may be a fine strategy (or not) for Sun's business, doing it in the context of Liberty was just bad for Liberty. The result was a war of words between Sun and Microsoft that has made it impossible for Microsoft to join Liberty, even if they wanted to.
The overwhelming majority of Liberty members want Microsoft to join Liberty. They are happy to make room for all the players, as demonstrated by Liberty's easy acceptance of AOL as a member. But thanks to Sun's efforts it will be a long time, if ever, before Microsoft will join Liberty.
This is how not to push forward the RFR movement. RFR must not be about punishing or excluding anyone. So long as a company is willing to accept RFR they should be welcomed and until they are willing to accept RFR they must not be insulted or attacked. This is a small industry and we must all work together. So let's make sure we leave the door open for everyone to eventually become part of RFR.
 W3C standards tend have IP declarations of the form "Before the W3C standardizes this we claim X but if the W3C standardizes this then we claim Y". Generally speaking X is almost always RAND while Y may be RAND or royalty free. The table shows the claims made for Y.
 These companies put in language that reads something along the lines of "…on a royalty-free basis or on other reasonable and non-discriminatory terms…" (taken from Arbia's WSDL declaration) Near as I can tell this means that their contribution is royalty free unless they decide it should be RAND.