W3C home > Mailing lists > Public > www-rdf-rules@w3.org > October 2001

Re: Let's get the Literals out of the RDF Graph

From: Pat Hayes <phayes@ai.uwf.edu>
Date: Fri, 12 Oct 2001 22:44:42 -0500
Message-Id: <p0510101bb7ed68287bf1@[205.160.76.193]>
To: "Thomas B. Passin" <tpassin@home.com>
Cc: www-rdf-rules@w3.org
>[Sandro Hawke]
>
>>  > From Patrick.Stickler
>>  >
>[...]
>>  >
>>  > After all, a string *is* a sequence already. That's part of its
>definition.
>>  > No need to make explicit what can be left implicit and reliably obtained
>>  > as needed. It's all about which is more expensive. I.e., whether you'd
>>  > be dissecting the literals or concatenating the characters more often.
>>
>>  If we make it explicit, then computers can understand it, which I
>>  think is the point of the semantic web.  As I understand your
>>  proposal, every new type of literal to come along would require
>>  additions to all the deployed semantic web agents for them to
>>  understand data using the literal.  Not so with mine.  With mine, once
>>  an agent understands the very concept of Dates and Numbers (via an
>>  ontology like my String one), they get (for free) to understand all
>>  formats of dates, floating point numbers, numbers in various bases,
>>  etc -- because they never see that stuff, since it's not part of the
>>  graph.
>>
>
>I don't see that we have to make knowledge about a string "explicit".  Every
>programming system has some notion of "string" already.  What we have to
>have is a way to distinguish between properties (objects) that are intended
>to be strings (or string representations of types like integers) from URIs.
>
>In other words, give me a way to distinguish the difference between a
>property value that really is a URI (i.e., it really is a resource) from a
>property value that is a string that happens to look like a URI (i.e., it
>really is a literal).  One virtue of using something like the data: scheme
>for literal values is precisely that it makes that distinction possible.

Right. Also, it can be made explicitly accessible to an RDF reasoner 
by a slight extension to the current RDF to make it datatype-aware, 
eg by writing (in Ntriples):

something hasproperty lit:couldbeanything .
hasproperty rdfs:range xsd:string .

Pat Hayes (but the idea is Peter Patel-Schneider's)


-- 
---------------------------------------------------------------------
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 Friday, 12 October 2001 23:44:52 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:53:09 GMT