- From: LYNN,JAMES (HP-USA,ex1) <james.lynn@hp.com>
- Date: Tue, 23 Sep 2003 20:07:27 -0400
- To: Jim Hendler <hendler@cs.umd.edu>, "Peter F. Patel-Schneider" <pfps@research.bell-labs.com>
- Cc: public-sw-meaning@w3.org
I really liked the "Strict Tim" approach for its simplicity, but have slowly been changing to a more social approach. Let me throw out a few ideas and wait for comments. I think I see parts of what I am thinking in both sides of the debate, but let's see. It seems that it would be useful to embrace the concept of vocabulary (ontology) as "definition by usage". Nothing very revolutionary here, just following what we do in everyday life. In everyday assertions, disputes will often be settled by referring to a dictionary, generally any commonly recognized dictionary such as Webster's or Oxford (assuming all conversations are in English here). The point is, the publisher of the dictionary doesn't "own" the words and definitions, but rather reports on commonly accepted usage. Wouldn't it would be beneficial to allow definitions of names (URIs) to evolve in a similar manner? Sort of a genetic approach? I realize this brings up issues such as "who is the keeper of the URI http://www.common-usage-dictionary.org" (there doesn't have to be just one) and the versioning issue as usage evolves, but let's save that for another thread. Secondly, it is very useful to have special meanings in various disciplines. I think that this is much more easily solved, and to some extent is already being done - Dublin Core. Perhaps in some areas there could exist specialized ontologies a la "technical dictionaries" http://www.math-logic.org , but even in their absence one could reference definitions from another ontology. This is what is done in technical writings all the time, as in the LBase doc which references Enderton's book on logic regarding a definition of first-order logic. Nothing prevents me from writing a paper on logic stating that first-order logic is really second-order logic, while citing Enderton's definition, but why would anyone do this. What use would there be to have a spec which prevents this? Eagerly awaiting rebuttal, James Lynn
Received on Tuesday, 23 September 2003 20:08:29 UTC