From: Peter F. Patel-Schneider (pfps@research.bell-labs.com)
Date: 02/07/01
From: Dan Connolly <connolly@w3.org> Subject: Re: added diagrams to "Using XML Schema Data Types..." Date: Wed, 07 Feb 2001 11:23:18 -0600 > "Peter F. Patel-Schneider" wrote: > > > > It appears to me that Dan's proposal is to turn every property (that has a > > datatype as its type) into two properties, one that maps into the string > > and one that maps into the datum corresponding to the string. (Maybe the > > proposal is even to have multiple properties, one that maps into the > > string, and one that maps into a datatype reading of that string, for every > > possible datatype.) > > > > I would be much happier with a solution that had a single property, namely > > size, that mapped directly into a datatype. I don't see any benefits from > > having these multiple properties; only disadvantages. > > The advantage is that parsing formulas (RDF documents) > remains independent of other stuff, including trust issues. > > Otherwise, if I put the range(size, Decimal) information in > in file X, and then I <size>10</size> in file Y, > the interpretation of Y depends on whether I have seen > (or believed etc.) file X. That doesn't seem workable to me. Either you trust the definition of a property or you don't. If you don't, then you don't get to use the information therein, and lots of things may break. If you do, you do get to use the information therein, and everything should work. As far as timing goes, if you don't have information on how to parse some input, why not defer the parsing until you have more information? In particular, it seems to me that your proposal has exactly the same problem. You also depend on external information on how properties should work. > > > Further, it appears to me that Dan's proposal breaks RDF in a very > > significant fashion, requiring literals (or at least datatype values) to be > > the source of properties. > > I don't see this as breaking RDF. It's always been > the case that you could look at the string "xyz" > as the resource data:,xyz and use it as the subject > of an assertion. Where is this described/defined? How is the URI ``data:,xyz'' related to the RDF literal xyz? peter
This archive was generated by hypermail 2.1.4 : 04/02/02 EST