W3C home > Mailing lists > Public > www-rdf-logic@w3.org > September 2001

RE: Literals (Re: model theory for RDF/S)

From: Pat Hayes <phayes@ai.uwf.edu>
Date: Thu, 27 Sep 2001 20:33:21 -0500
Message-Id: <p05101016b7d983a21096@[205.160.76.184]>
To: Patrick.Stickler@nokia.com
Cc: www-rdf-logic@w3.org
>  > -----Original Message-----
>>  From: ext Pierre-Antoine CHAMPIN [mailto:champin@bat710.univ-lyon1.fr]
>>  Sent: 27 September, 2001 12:03
>>  To: www-rdf-interest@w3.org; www-rdf-logic@w3.org
>>  Subject: Literals (Re: model theory for RDF/S)
>>
>>
>>  Sorry to bring back an old debate, probably set up for everyone except
>>  me, but their in curently another debate on rdf-logic about the model
>>  theory, and the special case of literals make it quite complicated.
>>
>>  So I ask: given the existence of a quite convinient URI
>>  scheme 'data:',
>>  is the Literal/Resource disctinction still necessary in RDF?
>>
>>  Couldn't we consider that any literal in the syntax should be
>>  converted
>>  by the parser into a 'data:' URI? This does not need to alter the
>>  syntax, and then keeps compatibility with existing RDF bits.
>
>I've been leaning towards this sort of view for awhile now. I
>also agree that it would make the conceptual model more consistent.
>
>>  Furthermore, this would offer a straightforward solution to some
>>  problems like expressing the xml:lang attribute in RDF.
>
>I don't think actually that it would provide a solution for
>that per se, as xml:lang is a property of the occurrence, not
>the literal. E.g. the literal byte sequence "pan" could be
>either English or Spanish (and possibly other languages as well)
>yet I would presume that <... xml:lang="en">pan</...> is
>a statement about the particular occurrence of the literal and
>not about the byte sequence "pan" in general.

Seems odd to me to call this a statement *about* the occurrence. Why 
not say that this entire expression, ie the string 'pan' surrounded 
by an xml:lang="en" tag, is a particular lexical entity (more highly 
structured than a character string, to be sure, but still a lexical 
item as far as the semantics is concerned) which differs from, eg 
<...xml:lang="fr">pan</...> simply by being, well, different? On this 
view, the things called 'literals' in the model theory might be quite 
complex entities which included xml-defined properties in their very 
lexical nature. This treats xml tags as markup, in the literal sense, 
ie as *part of* the very structure of the language, rather than a 
*description of* it. This isn't really a new idea; we might, a few 
years ago, have wanted to make lexical distinctions between 'pan' 
written in one font from the same character string written in a 
different font, say (the original KIF documentation used font 
distinctions to encode semantic categories, for example) , or even in 
a different style (boldface versus italic), all of which would appear 
in xml as different tags, so the idea of markup as included in syntax 
rather than about it seems quite natural to me.

That said, I would agree with what I bet would be your reply, which 
is that what one really wants is for it to be both at the same time, 
and to use the descriptive power of the semantic language to be able 
to re-describe its own syntax, right? (If I follow the rest of your 
message)

>You wouldn't
>(I don't think) want to have multiple statements attributing
>multiple languges to a given string -- at least insofar as
>the intended use of xml:lang is concerned, right?
>
>I would myself love to see a data type URI approach by which
>otherwise "literal" values could be defined as instances of a
>given data type URI. E.g.
>
>    dt:integer:5
>    dt:token:en
>    dt:date:2001-09-27
>    dt:time:2000-11-01T17:32:20Z
>    dt:float:38829.11883292
>    ...
>
>etc.
>
>Then, one can associate the syntax and semantics of the particular
>data types using RDF.
>
>Furthermore, one could say things about the data type scheme like
>
>    <rdf:Description rdf:about="dt:date">
>       <rdfs:isDefinedBy rdf:resource="http://www.w3.org/TR/NOTE-datetime"/>
>       <rdfs:seeAlso
>rdf:resource="http://www.iso.ch/iso/en/CatalogueDetailPage.CatalogueDetail?C
>SNUMBER=26780"/>
>
><dt:pattern>[0-9]{4}-((0[1-9])|(1[012]))-((0[1-9])|([12](0-9))|(3[01]))</dt:
>pattern>
>       <rdf:type rdf:resource="http://purl.org/dc/terms/W3CDTF"/>
>    </rdf:Description>
>
>etc.
>
>All the info you need to constrain property ranges, perform validation
>of input values, and what not. (Note that the URI scheme is combined
>with a namespace for defining properties of namespaces (instances) of
>that scheme).
>
>Eh?

Wow, that meta-description idea would be quite a challenge for a 
model-theoretic semantics, but I'd love to give it a try. In theory 
one can do things like this in KIF 3, using quasi-quotation, but I 
doubt if anyone ever has seriously tried to do it in a realistic 
setting.

Pat Hayes

-- 
---------------------------------------------------------------------
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 Thursday, 27 September 2001 21:33:28 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 2 March 2016 11:10:36 UTC