From: Peter F. Patel-Schneider (pfps@research.bell-labs.com)
Date: 05/22/01
I've been looking at Stefan's Layers document, and have a number of questions about it. Most of these have to do with undefined or underdefined terms and phrases. I have some suggestions on what needs to be changed, but I would prefer to make a general plea for such documents: If anyone wants to make a language proposal, they should have a formal definition of the proposal. I do not view this as too onerous, as there need not be a semantic document for most things, as the current semantics can be used. The formal document thus boils down mostly to a firm definition of the syntax of the language. My comments on Stefans proposal: Level 0: I have no problems with this, except that it is very under-defined. I think that we could easily provide a layer of DAML+OIL that is essentially frames done right. Level 1: The terms ``simple class'' and ``simplified property restriction'' are not defined. This makes the proposal very hard to evaluate. Even with a very restrictive definition of these phrases I have a very strong feeling that reasoning will be just as hard in this level as in the entire language, as level 1 includes qualified number restrictions and enumerations. Unique Name Assumption: We discussed this a lot. I think that a UNA would be very useful, but it needs to be carefully crafted so that no non-monotonicity leaks into the formalism. peter
This archive was generated by hypermail 2.1.4 : 04/02/02 EST