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

Re: email straw poll: literal semantics proposals

From: Frank Manola <fmanola@mitre.org>
Date: Tue, 08 Oct 2002 10:13:37 -0400
Message-ID: <3DA2E811.6010501@mitre.org>
To: Brian McBride <bwm@hplb.hpl.hp.com>
CC: RDF Core <w3c-rdfcore-wg@w3.org>

Votes:

B: 0

It doesn't do datatyping, and backs out what I considered progress on 
the issue.  It also addresses only (at least, as currently written) XML 
Schema formats, which I consider unreasonably restrictive.  (If we're 
going to back out what we've already decided about datatyping, I think 
I'd consider backing out the whole thing, rather than adding last-minute 
formatting declarations.)

C: 1

D: 2

F: 5

Comments:

1.  As the options are presented, it seems to me we're mixing (a) what 
the semantics of otherwise undatatyped literals are, and (b) how to  get 
datatypes attached to otherwise undatatyped literals.  It seems to me 
that whatever scheme(s) we provide to do (b), we can't necessarily 
guarantee people will use it/them, so we still have to have a clear 
answer to (a).  (I think this means we need to intended entailments for 
the various proposals, as well as the proposed syntax).

2.  I have to admit that D is growing on me a little (particularly if 
that's the only option to completely punting.  I'm still a bit concerned 
with the scoping issues, though, which do not seem to me have been 
sufficiently explored.

3.  My vote for F is a 5 because (according to the stated rules) that's 
my "most preferred".  However, it's got a caveat:  I still don't know 
what F says about the semantics of untyped literals (see comment 1).

4.  Is this straw poll connected in any way with the communication to 
WebONT?

--Frank

Brian McBride wrote:

> 
> I would like to take an email straw poll of the WG regarding the options 
> for semantics of literals we currently have on the table.  For each of 
> the attached options, please could you give a score to indicate your 
> preferences.  I would like to have a strong indication of the sentiment 
> of the WG by Tuesday evening.
> 
>  o Unnacceptable proposals should be given a score of 0.  Please 
> indicate what absolute requirement(s) are not satisfied
> 
>  o Other proposals should be ranked from 1 to 5, with 5 being most 
> preferred.  Give options with equal preference equal scores.
> 
> Please note that proposal B includes retracting the decision to have 
> datatype values in the graph, i.e. retracts the
> 
>   <foo:age rdf:datatype="xsd;integer">10</foo:age>
> 
> construct.
> 
> 
> 
> Proposal B:
> 
>   <rdf:Description rdf:about="Jenny">
>     <foo:age>10</foo:age>
>   </rdf:Description>
>   <rdf:Description rdf:about="&foo;age>
>     <rdfs:format rdf:resource="&xsd;integer"/>
>   </rdf:Description>
> 
> The object of an age property denotes a string literal [*] from the 
> lexical space of xsd:integer.
> 
> Proposal C:
> 
>   <rdf:Description rdf:about="Jenny">
>     <foo:age rdf:datatype="&xsd;integer>10</foo:age>
>   </rdf:Description>
>   <rdf:Description rdf:about="Film">
>     <foo:title>10</foo:title>
>   </rdf:Description>
> 
> 
> The object of the age property denotes a member of the value space of 
> xsd:integer.  The object of the title property denotes a string literal[*].
> 
> Proposal D:
> 
>   <rdf:datatype rdf:property="&foo:age" rdf:datatype="&xsd;decimal"/>
>   <rdf:Description rdf:about="Jenny">
>     <foo:age>10</foo:age>
>   </rdf:Description>
>   <rdf:Description rdf:about="Johnny">
>     <foo:age>10</foo:age>
>   </rdf:Description>
> 
> Proposal D is proposal C with some syntactic sugar (the rdf:datatype 
> declaration) so that it is not necessary to put an rdf:datatype 
> attribute on every use of the age property.
> 
> Proposal F:
> 
>   <rdf:Description rdf:about="Jenny">
>     <foo:age>10</foo:age>
>   </rdf:Description>
>   <rdf:Description rdf:about="&foo;age>
>     <rdfs:range rdf:resource="&xsd;decimal"/>
>   </rdf:Description>
> 
> The object of an age property denotes an integer from the value space of 
> xsd:decimal.
> 
> [*] A string literal is an old style RDF literal which consists of a 
> unicode string and language identifier.
> 
> 


-- 
Frank Manola                   The MITRE Corporation
202 Burlington Road, MS A345   Bedford, MA 01730-1420
mailto:fmanola@mitre.org       voice: 781-271-8147   FAX: 781-271-875
Received on Tuesday, 8 October 2002 09:58:02 EDT

This archive was generated by hypermail pre-2.1.9 : Wednesday, 3 September 2003 09:52:22 EDT