W3C home > Mailing lists > Public > w3c-rdfcore-wg@w3.org > April 2003

RE: xmlsch-01 Typed Literal Structure proposal to close

From: <Patrick.Stickler@nokia.com>
Date: Wed, 23 Apr 2003 10:40:08 +0300
Message-ID: <A03E60B17132A84F9B4BB5EEDE57957B01B90D02@trebe006.europe.nokia.com>
To: <jjc@hpl.hp.com>, <w3c-rdfcore-wg@w3.org>

> -----Original Message-----
> From: ext Jeremy Carroll [mailto:jjc@hpl.hp.com]
> Sent: 22 April, 2003 14:07
> To: w3c-rdfcore-wg@w3.org
> Subject: xmlsch-01 Typed Literal Structure proposal to close
> Summary: reject.
> The comment is
> [[ 
>    The introduction of pairs consisting of a lexical form and 
> a type (or,
>     strictly speaking, a lexical form and a type label) seems at first
>     glance to complicate the RDF model somewhat. We have had the
>     impression that in other parts of RDF, typing is handled by adding
>     further arcs and nodes. If the type of a resource is identified by
>     having an arc labeled rdf:type from it to (the URI of) 
> its (RDF) type,
>     and if the type of an arc is similarly identified by an arc, then
>     surely a reason ought to be given for shifting to a 
> different method
>     for typing literal strings. It seems like a dramatic shift in the
>     infrastructure of RDF, from "everything is a node, an arc, or a
>     literal value" to "everything is a node, an arc, or a 
> typed literal
>     value". Perhaps not quite so dramatic, after all. But the 
> question of
>     design consistency remains: why not "everything is a typed node, a
>     typed arc, or a typed literal"?
> ]]
> The propose response is to do nothing.
> Here is a draft reply:
> [[
> We have considered your comment and rejected it.
> This aspect of the new design has been the one that
> we found the most difficult, and we are not surprised that
> aspects of it seem overly complicated.
> To answer your question as to why we use a different
> means to type literals than that used to type nodes and
> properties:
> - RDF nodes and properties are typically untryped.
>   Type information can be added monotonically,
>   potentially by a further document elsewhere on the
>   Web, or by an inference rule.
> - untyped literals, however, are character strings, or
>   pairs of character strings and language tags.
>   Adding type information, such as maling a speciifc
>   literal node into an xsd:int, is not monotonic, but
>   a destructive operation, in that previously true
>   information, and valid inferences may cease to be true.
> Because of this difference in the behaviour of
> literal nodes from other nodes in response to addiitonal
> type information, the WG decided to provide a different
> mechanism, operating syntactically, rather than the rdf:type
> mechanism which operates semantically.

Hmmm....  sorry, but I don't myself buy that explanation.

I don't see how asserting a datatype to a literal (presuming an
untidy treatment of course) is any different than asserting a 
type for any resource. If long range datatyping is non-monotonic
then all RDF typing is non-monotonic.

The *real* reason, IMO, why the proposal should be rejected is 
because to accept it would require literals to be subjects and
I think the WG has a fair bit of consensus that such a change
would not be allowed by the present charter.

(Though I fully agree that having literals as subjects and
allowing rdf:type assertions with datatype objects would have
been the most elegant solution overall)

I suggest that the response be changed to focus on the disallowance
of literals as subjects rather than positing non-montonicity of
long range datatyping (which I don't see is a problem).



> We remind you that the group was far from unanimous
> in its discussions concerning literal typing, with Mike Dean's
> formal objection being indicative of the minority position:
> http://lists.w3.org/Archives/Public/w3c-rdfcore-wg/2003Jan/0173.html
> ]]
> Jeremy
Received on Wednesday, 23 April 2003 03:40:11 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 20:24:22 UTC