Evangelists believe Web services will one day enable computers everywhere to talk to each other without any human intervention. But before that communication can take place it needs to have a context, and that's what Dave Hollander is trying to create. Hollander -- who co-created XML, leads the World Wide Web Consortium (W3C)'s XML Schemas Working Group, and participates in its Web Services Architecture Working Group -- is the chief technology officer at Contivo Inc., a Mountain View, Calif.-based company that looks to reduce integration costs using XML. On his way to this week's W3C meetings in Boston, Hollander spoke with SearchWebServices.com about royalty-based standards, the W3C's agenda and the evolution of XML.
Hollander: That's not my working group, so I can't speak for them, but right now my understanding is that it's a matter of the exact wording, as opposed to the intent. But sometimes that can be just as troublesome.
SearchWebServices.com: Can Web services succeed in the long term without royalty-free standards?
Hollander: I think the further we go from the core standards -- SOAP and WSDL, but maybe not UDDI -- the more likely we are to see some royalty structures introduced. I don't know how that is going to happen. When adding a lot of features and functionalities, you always have to make the tradeoff between what will impede the industry versus benefit the industry. If that means changing some perceptions around what we do with royalty-free standards, then that may be better than having meaningless competition.
SearchWebServices.com: Can Web services succeed if some standards are royalty-based?
Hollander: It depends which ones. The capability of Web services is ubiquitous data transport, and achieving the same kind of ubiquitous data access that the Web gives you. Going beyond that, I suspect there's another whole layer that should be royalty-free, but if you get much beyond that you're paying for a lot of R&D for things that haven't happened already. But the CD-ROM is a good example of licensed technology that's not royalty-free that's still a standard and ubiquitous, and there are lots of examples of licensed technology that are very successful in the marketplace.
SearchWebServices.com: What issues are at the top of the agenda for the W3C's Web Services Architecture Working Group?
Hollander: We really have two goals. One is to finish flushing out the base architecture specification, such as how to use SOAP, WSDL, UDDI registries, discovery, and to build on the fundamental foundation of how those things work together. Then we'll do the introduction to the next whole layer of capabilities that are needed, like reliable messaging, choreography and transactionality. Second, we're helping paint the picture of what choreography is -- the issues and constraints -- so they can be considered by the W3C as a whole next week, so we can decide whether to do choreography as a working group.
SearchWebServices.com: When the W3C is done with its Web services architecture blueprint, what will it mean to companies that want to work with or implement Web services?
Hollander: It'll mean a cleaner understanding of what capabilities are going to be functionally interoperable, and what capabilities are going to take extensions. From my viewpoint, the bigger goal of the architecture is less for the vendor companies and more for the industry as a whole. There's so much more work to do with Web services, and so much potential, [but] there are some heavy concepts out there for people to try to understand. We're trying to put those into some context so intelligent decisions can be made about what to work on next.
SearchWebServices.com: The W3C has been advocating its vision of the Semantic Web for a while now. Why does the W3C believe the Semantic Web is important?
Hollander: Clearly there's a gap in getting computers to understand computers, instead of having humans do the moderating. That would be a huge step forward. With the Web today, we use computers to facilitate human-to-human interaction. With XML and Web services, we can now get data moving on its own, but we still don't know what it means. That's clearly the next big step.
Hollander: No, and anybody who can is lying. The Semantic Web is as unknowable today as today's Web would be to someone in 1990. Back then, nobody could predict what we have in our browsers today and how they work. Sitting here today and trying to figure out what the Semantic Web will be in 2020 is impossible.
SearchWebServices.com: As one of the creators of XML, are you pleased with how it's evolving and how it's being extended?
Hollander: I think they took a simple concept, and now there's a whole new layer of junk on top of it like schemas -- which I'm responsible for -- and XQuery and XPath. In a few more years, we'll be ready to filter that back down. But they're very capable and they do the job well; they're just a little bit awkward to get your head around. XQuery is going to be a large, difficult spec, but it does a large difficult job and it's going to take the industry a while to figure out how to manage it.
What people don't remember is that with XML -- among the nine or ten of us who were on that original committee from day one -- we had well over 100 years of shared experience in dealing with markup languages. So we had a lot of shared experience to know if what we were creating met the minimum for success. So I think in general we're going a pretty good job, but it's a lot harder than any one of us would like to think.