- From: Frank Manola <fmanola@acm.org>
- Date: Mon, 19 Dec 2005 17:04:14 -0500
- To: Joshua Allen <joshuaa@microsoft.com>
- CC: Henry Story <henry.story@bblfish.net>, semantic-web@w3.org
Joshua Allen wrote: >>they can't be used in conjunction with machine-learning. It seems to >>me, for example, that a useful "hybrid" approach would be to apply >>machine-learning techniques to allow such schemas to evolve from their >>initial definitions with use, hence becoming less static and manually >>maintained. > > > That's reasonable. I worked for a time with some technology from MSR > which takes two different relational or XSD schemas and suggests > mappings between them. This was based purely on analysis of the two > schemas and some instance data -- it is easy to imagine this being > augmented with feedback from thousands of users. Sounds like Phil Bernstein's model management work, right? Good stuff. > > You're right that schema itself is not a bad thing; although I feel that > rdf:type is a stumbling block to RDF adoption right now. > What do you see as the problem with rdf:type? The need to identify a specific type at all (which implies a set of types somewhere which somehow constrains the instance)? That doesn't mean that the instance can't also be of some other type, or even many other types, say from sets of types defined by different people. --Frank
Received on Monday, 19 December 2005 22:03:34 UTC