Last week, the W3C held its Workshop on Frameworks for Semantics in Web Services in Innsbruck, Austria. People who have been working on all the major approaches to semantically enhanced Web services were there, along with other interested parties from industry and the public sector. The major questions were when to go to the W3C to form a Working Group to standardize this area and how to decide what should be standardized. There are several competing (in a friendly way) proposals on the table. I was there as one of the co-authors of the WSDL-S proposal. Our position is that it is too early to start standardization of semantic Web services. We need to gain experience with this technology attacking real-world problems to quantify the benefits and the impediments. This is why we proposed a simple approach like WSDL-S. It allows us to cover the basic needs in a way that we can use in proof-of-concept projects. That is also why last week we release on alphaWorks the Semantic Tools for Web Services, that support WSDL-S and can be used to implement such projects.
The workshop participants agreed that some practical business implementation projects need to be done as the next step and some proposed that one or more challenge problems should be defined and solutions implemented using the major semantic Web services proposals. After that, the relative benefits can be evaluated. Not everyone agreed that we need to wait for this step. Some felt that there was a window of opportunity that is open now. I think we need to have more people trying the semantic technology before we proceed. The possibilities it presents to speed integration projects, promote higher quality solutions and enable better service reuse need to be tested and understood.
Semantics and Web Services: To Standardize or Not