W3C home > Mailing lists > Public > w3c-rdfcore-wg@w3.org > October 2001

Re: big issue (2001-09-28#13) & changes in MT

From: Pat Hayes <phayes@ai.uwf.edu>
Date: Mon, 8 Oct 2001 17:11:40 -0500
Message-Id: <p05101014b7e7d20557d8@[205.160.76.193]>
To: Sergey Melnik <melnik@db.stanford.edu>
Cc: w3c-rdfcore-wg@w3.org
Sergey, sorry this slipped off my screen.

>I'm thinking of how composite literals may be potentially introduced in
>the MT document, and what the implications of that could be.
>
>A first shot at MT of datatyping could be this: assuming that literal
>constants are pairs of kind (URI, unicode string), let the URIs
>(resource constants) be mapped to resources as before using function IS.
>Additionally, one could define a mapping IDT from IR into the powerset
>of LV that maps each datatype-resource to its extension, namely, the set
>of all literals that correspond to this datatype. For example, if X001
>is an interpretation of resource constant
>[http://www.w3.org/2001/XMLSchema-datatypes, integer], so the extension
>of the datatype XOO1, i.e. IDT(X001), would include all "integer"
>literals.

If I follow you here, the literal itself would be rather like a pair 
of a URI and a special kind of property whose value is the class of 
literal values that this URI might have.  But what determines the 
actual literal value? Eg suppose I write

aaa foo <bbb, [http://www.w3.org/2001/XMLSchema-datatypes, integer]> .

then the interpretation I(bbb) isn't itself an integer, right(? Or is 
it? If so, what ensures that it is?) 
IDT([http://www.w3.org/2001/XMLSchema-datatypes, integer]) is the set 
of integers, OK. But what is the thing that is the value of the 
property IEXT(I(foo)) on I(aaa)?

Sorry if I'm  being dense.

>Some general comments on MT:
>
>I very much welcome your move towards a cleaner separation between RDF
>and RDFS. I think it maybe worth it to emphasize the point that other
>languages (besides RDFS) can be layered on top of RDF, and to illustrate
>how their semantics can be defined incrementally (e.g., using the
>I-mappings of the underlying languages). In this light, RDFS-MT could be
>presented as a proof-of-concept of how this layering is done.

That is how it is meant at present. I will try to state this more 
clearly, though I want to be careful not to claim the moon. Also, 
Peter P-S is re-writing the DAML MT to be a further extension in the 
same style, so that will help make the same point.

>A minor comment: could you make sure that all symbols are introduced in
>the same section (or are summarized in a table)? For example, most of
>the definitions are in 1.4, but LV and XL appear first in 1.3.

OK, I will fix this.

>Another
>issue, which I raised in my previous email, is a consistent naming of
>constants vs. individuals in the domain of discourse (e.g. resource
>constant and literal constant instead of URI and literal, or something
>like this).

We are still arguing about this on that other thread.

Pat

-- 
---------------------------------------------------------------------
IHMC					(850)434 8903   home
40 South Alcaniz St.			(850)202 4416   office
Pensacola,  FL 32501			(850)202 4440   fax
phayes@ai.uwf.edu 
http://www.coginst.uwf.edu/~phayes
Received on Monday, 8 October 2001 18:11:45 EDT

This archive was generated by hypermail pre-2.1.9 : Wednesday, 3 September 2003 09:40:58 EDT