- From: Drew McDermott <drew.mcdermott@yale.edu>
- Date: Mon, 9 Jun 2003 16:07:07 -0400 (EDT)
- To: sambrosz@ipipan.waw.pl
- CC: www-ws@w3.org
Sorry for the delay in replying to this message; I've had some deadlines to cope with. Date: Thu, 22 May 2003 16:32:11 +0200 (CEST) X-PH: V4.4@mr2 From: Stanislaw Ambroszkiewicz <sambrosz@ipipan.waw.pl> ... Does the following scenario follow your line of reasoning? There was a protocol (plan) such that once performed successfully by X, "X has received an invoice saying \ldots about Y." Then, suppose that X performed the protocol, and X got Y. X gave the following name to this situation: "X owns Y". It seems to me that if X already knows it "got Y," then he already knows he owns it. Or perhaps you mean that "owns" is to be used in a technical sense: "X obtained Y by performing plan P." Later on, X applied so called "data abstraction principle" to this situation, and got the following formula: "?x owns ?y". X realized that there were other protocols that lead to the same situation. Okay, then "owns" _can't_ mean what I suggested, because by definition protocol P must be involved in "ownership." X wanted to speak to others in terms of this formula. So that, X published the formula "?x owns ?y" (specifying also the protocols) to "the committee in charge of the ontology". If "owns" really has the meaning it has in natural language, then X already knows the meaning before he starts dabbling in protocols and plans. He or the committee can issue information about how the word translates into different natural languages, clarifications of important borderline cases, and so forth. What's puzzling about the story you're telling is that it's trying to solve a problem we shouldn't be trying to solve: the problem of how words get their meanings and what we mean by the "meaning of a word." As I argued before, we can formalize protocols and their effects using predicates derived from natural language, without thereby incurring an obligation to explain natural language. Just one remark: "The committee" does not seem to be a good solution in heterogeneous and distributed environment. You're right that there won't be one committee that decides everything, but there might be lots of committees, each of which standardizes the vocabulary in a particular area. The trick is to make sure everyone agrees on which committee is deciding what. I don't think that's so hard in a heterogeneous distributed environment; sometimes this planet seems like a small town, where everyone knows who is interested in what. -- -- Drew McDermott
Received on Monday, 9 June 2003 16:07:10 UTC