- From: Drew McDermott <drew.mcdermott@yale.edu>
- Date: Wed, 18 Aug 2004 17:16:27 -0400 (EDT)
- To: public-sws-ig@w3.org
> [Martin G\"ulich]
> How are web services supposed to be found by semantic matching
> (capability-based matching) of descriptions written in OWL-S? All
> descriptions describe what a web service really does implicitly, so you > are
> supposed to be able to infer what a service does by looking at the > inputs
> and outputs, right? But what happens if you have a trivial calculator
> service that adds two numbers into a sum. Inputs can then be two > Integers
> and output can also be an Integer. But what distinguished this > service
> from another calculator service that subtracts, multiplies or divides
> instead? Or from a coordinate transformation service that maps from two
> dimensions to one? Are preconditions and effects supposed to enable this
> kind of distinguishing? Or do you also have to annotate each service
> explicitly by calling it Calculator_Add, Calculator_Subtract > etc? It
> seems to me that somehow a service has to be distinguished by explicitly
> annotating the service itself, the conditions and effects or the inputs > and
> outputs.
In addition to a description of the output, we also need a description
of the effect, which is, to boil it down to its essence --
(know-value client (+ input1 input2))
--
-- Drew McDermott
Yale University CS Dept.
Received on Wednesday, 18 August 2004 21:16:28 UTC