- From: Bijan Parsia <bparsia@isr.umd.edu>
- Date: Mon, 21 Nov 2005 18:03:46 -0500
- To: drew.mcdermott@yale.edu
- Cc: public-sws-ig@w3.org
On Nov 21, 2005, at 10:52 AM, Drew McDermott wrote: >> [jeff@inf.ed.ac.uk] >> >> BTW, why is it said that "the current WSDL standard operates at the >> syntactic level"? What is any more semantic about the things that >> are labelled "semantic"? > > By old and well established usage, "semantic" means "complex, > expressive, insightful, ours," contrasted with "syntactic," which > means "simple, weak, error-prone, theirs." My fav example is "semantic description" vs. "syntactic description". > It would be nice to avoid this term completely, but then we'd have to > change the name "Semantic Web." Call it the SWeb where "S" stands for "super". Aside from this linguistic reflections, it is the case that WSDL descriptions are by design, intent, and happenstance rather thin. Even in the realm of the evil syntaticalist roaders, you see that (hence the panoply of WS-* specs and, of course, BPEL and WS-CDL). So, core WSDL does *not* describe many interesting aspects of a service. What is somewhat distinctive about various semantic web approaches is that 1) the use of preconditions and effects (conditional or otherwise) derived from planning and 2) logic based formalisms for categorization of services. Service taxonomies are present in things like UDDI, but tend not to be based on descriptions but on tagging (i.e., you always must asset what category you are in, perhaps with inheritance of supercategories). I do think that grounding stuff on formal models can be useful (and, if you look at, e.g., XQuery (which has a formal semantics) or WSDL (which has a Z specification...much to the consternation of a chunk of the group :)), you'll see that there is some movement in this direction). Extant semantic web languages are also interesting because they are differently expressive than what, I believe, a lot of folks are used to. Sometimes this can provide insight (e.g., <http://www.mindswap.org/2005/services-policies/>). Cheers, Bijan.
Received on Monday, 21 November 2005 23:04:05 UTC