W3C home > Mailing lists > Public > w3c-rdfcore-wg@w3.org > April 2002

Re: RDF Datatyping MT *does* define Datatyped Literal Pairings

From: Pat Hayes <phayes@ai.uwf.edu>
Date: Fri, 19 Apr 2002 18:22:20 -0500
Message-Id: <p0510154eb8e6543b960f@[65.217.30.94]>
To: Patrick Stickler <patrick.stickler@nokia.com>
Cc: w3c-rdfcore-wg@w3.org
>On 2002-04-19 3:24, "ext Pat Hayes" <phayes@ai.uwf.edu> wrote:
>
>>
>>  A snippet which gets to the root:
>>
>>>
>>>>   To conclude that Jenny's ex:age is ten would be a mistake, an invalid
>>>>   inference. We should make this painfully clear to users, so they do
>>>>   not get their RDF in a muddle.
>>>
>>>  I didn't say Jenny's ex:age was ten. I said Jennys age was ten.
>>
>>  Fine. But what Jenny's (real) age is, is NOTHING TO DO WITH US. We
>>  are giving a spec for the RDF. The RDF uses the uriref <ex:age>, so
>>  the meaning of that is what we are concerned with. And in the
>>  example, the meaning of that is that IEXT(I(<ex:age>)) contains
>>  <I(Jenny), "10">, not <I(Jenny), 10>. So that is what we should say
>>  about it, clearly and unambiguously; so that if someone wants the
>>  relational extension to contain something else, they can know to
>>  write their RDF differently.
>
>I'll try this one more time...
>
>The meaning of the literal node in the inline idiom is always
>the string. It is crystal clear from the MT that an application
>should always consider the value of the ex:age property to be
>a string.
>
>However, the presence of an rdfd:datatype assertion for the
>ex:age property also communicates to the application that it
>is free (nay, *expected*) to interpret the literal as a lexical
>form of xsd:integer.

No. That is like saying that a Java application is expected to run in 
a way that isn't supported by the Java VM. It is *not* expected to 
interpret RDF in a way that is inconsistent with the MT. The MT isn't 
some arcane piece of mathematics just put there for window-dressing. 
It is the reference point for RDF meanings, and moreover it is the 
ONLY reference point that is accessible to developers of RDF 
inference engines. So any other interpretation placed on an RDF graph 
is liable to become invalid as soon as any conforming inference 
engine draws any valid RDF conclusions. That would makes RDF 
completely irrelevant: it would no longer be a bearer of 
machine-accessible information.

>Thus, the semantics of the RDF graph say that the property
>value of ex:age is a string. But the semantics of the inline
>idiom and the rdfd:datatype assertion together communicate
>the datatype value ten.

They may 'communicate' it, but they do not mean it.

>This latter interpretation does not
>change or override the semantics of the literal node in
>the graph. But it does capture the intent of the content
>producer in associating the datatype with the literal

NO. It should not, because there is no way to encode 'intent' other 
than by interpreting the RDF, and that is interpreted according to 
the MT. If I want to send you some content in RDF, all you get is the 
RDF. You don't get a meta-commentary telling you my intent. So if 
someone has the intent of saying one thing, and they use RDF in a way 
that says something else, then they are using the language not 
according to the spec. Now, some users might want to ignore the fine 
points of the spec, and use it in their own way, particularly 
'in-house' where they can be sure that the readers will understand 
their quirky ways. Fine; but that, to repeat, is not our problem, and 
I would submit is not something that we should endorse or encourage.

>, to
>say that the literal should be interpreted in terms of the
>specified datatype (as there is no other rational reason
>to associate a datatype with a literal).
>
>Yes, the datatype value is provided to the application "above"
>or "beyond" or "in addition to" the meaning of the explicit
>nodes in the graph, but it still provided (or should be
>provided) by the RDF Datatyping semantics/interpretation.

You just contradicted yourself. The meaning of the nodes on the graph 
IS the semantics/interpretation. That is what 'semantics' MEANS.

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 Friday, 19 April 2002 19:22:23 EDT

This archive was generated by hypermail pre-2.1.9 : Wednesday, 3 September 2003 09:47:34 EDT