W3C home > Mailing lists > Public > www-rdf-interest@w3.org > February 2002

Re: [xml-dev] URIs are simply names was:Re:[xml-dev]"Abstract"URIs

From: Patrick Stickler <patrick.stickler@nokia.com>
Date: Wed, 20 Feb 2002 23:16:35 +0200
To: ext Manos Batsis <m.batsis@bsnet.gr>, RDF Interest <www-rdf-interest@w3.org>
Message-ID: <B899DED3.F248%patrick.stickler@nokia.com>
On 2002-02-20 19:24, "ext Manos Batsis" <m.batsis@bsnet.gr> wrote:

> From: Patrick Stickler [mailto:patrick.stickler@nokia.com]
> 
> [...]
>> The present consensus in RDF core is that literals will
>> be tidy and thus context is not born by the literal node
>> itself in the graph.
> 
> But it's not a literal node; it's a resource xsd.int:22, not "22" or 22.
> I understand the limitations of literals as you present them; that's the
> reason to use a non-retreivable resource scheme to get rid of them.

I think perhaps you are equating literal with 'value'. "22" is not
the integer 22. It may in fact be a lexical representation in hex
or octal notation, for instance.

A literal is just a string. In the context of a datatype interpretation,
that string is taken as a lexical representation for some value, and
it is the datatype that determines which value that is.

> 
>>> Ok, from scratch. How about this:
>>> 
>>> :mySweater xx:label xsd.string:22
> 
>> This is a TDL URV, a typed data literal uniform resource value.
> 
>> C.f. http://ietf.org/internet-drafts/draft-pstickler-tdl-00.txt
> 
> Indeed. Wow I had no clue. Actually, I expected some serious flames for my
> post :-)

Everyone's flamethrowers have already run out of fuel on me...  ;-)

The URV concept is considered somewhat radical...

>> A literal must have context to have consistent interpretation.
> 
> Allow me to sumarise; My intention was to use real URIs, resources as RDF(S)
> anticipates them to ban literals from the model.
> 
> Now, if the WG does not agree with "resourcing" literals (which, IMHO is by
> far the best solution), I guess the next best thing is the
> "literals-in-context" thingy; still I don't think I like it... it sounds like
> complicating things for less; but that's just me.

Well, the challenge is that, in RDF we can never have a representation
for a value. We can only identify values indirectly because RDF needs
to be portable across disparate systems, platforms, programming languages,
etc.

So, a value such as the integer 5 may have different system-internal
representations in different contexts.

And that's why we're stuck with the literal-in-context situation.


>> We'll see (and real soon now, we're supposed to be done already ;-)
> 
> I'm glad I got close :-)
> 
> Kindest regards,
> 
> Manos
> 
> PS: To tell you the truth, I cannot understand why the literal integer 22 (for
> example) is not unique :-(
> The way I see it, it's just a sequence of characters. Anyone using 22 means 22
> and sees 22; now what interpretation such a literal may have in a closed
> system should not be a consern to RDF.
> I just don't get it...

It has to do with the datatype context. For instance, the lexical form "22"
for a decimalInteger represents the integer value 22, but the same lexical
form "22" for an octalInteger datatype (using octal notation) represents the
integer value 18.

So it's not the case that the literal string "22" always means the
same thing. It depends on its interpretation in the context of a
given datatype.

Cheers,

Patrick
 

--
               
Patrick Stickler              Phone: +358 50 483 9453
Senior Research Scientist     Fax:   +358 7180 35409
Nokia Research Center         Email: patrick.stickler@nokia.com
Received on Wednesday, 20 February 2002 16:15:03 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:51:53 GMT