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

Re: rdfs:member for Collections and collection syntax?

From: pat hayes <phayes@ai.uwf.edu>
Date: Mon, 24 Feb 2003 14:45:25 -0600
Message-Id: <p05111b52ba80336cc27a@[10.0.100.86]>
To: Dan Connolly <connolly@w3.org>
Cc: w3c-rdfcore-wg@w3.org

>In DAML+OIL, we had item, inferred from first/rest...
>
><rdf:Property rdf:ID="item">
>   <rdfs:comment>
>     for item(L, I) read: I is an item in L; either first(L, I)
>     or item(R, I) where rest(L, R).
>
>http://www.daml.org/2001/03/daml+oil#item
>http://www.w3.org/TR/2001/NOTE-daml+oil-reference-20011218
>
>In RDFS, member seems to have no relation to first/rest...
>http://www.w3.org/TR/2003/WD-rdf-schema-20030123/#ch_member
>http://www.w3.org/TR/rdf-schema/#ch_collectionvocab
>http://www.w3.org/TR/2003/WD-rdf-schema-20030123/#ch_collectionvocab
>
>I thought rdfs:member was specified ala daml:item
>to be inferred from first/rest.

My understanding is that rdfs:member is the superproperty of all the 
container membership properties rdf:_1 etc, and has nothing to do 
with collections. Our collections practically have no semantics, 
remember: we don't even require that rdf:first be functional.

>
>I did some implementation assuming those semantics...
>I was given a collection of a WG's documents,
>and I just wanted the documents. There's no single
>property that will get me from the WG to the documents,
>so I used rules to make up the difference...
>
>------
>{
>   :GROUP a org:WorkingGroup; con:homePage :PG.
>
>   :F is log:semantics of [ a :GroupKB].
>
>
>   (:F { this log:forAll :L, :X.
>          { :L l:first :X } log:implies { :L l:item :X }.
>          { :L l:rest [ l:item :X ] } log:implies { :L l:item :X }.
>        }) log:conjunction [ log:conclusion :FLists ].
>
>   :FLists log:includes {
>         [ a org:WorkingGroup;
>           con:homePage :PG;
>           wgi:hasDrafts [ l:item :WORK ]
>         ]
>   }.
>}
>   log:implies { :GROUP org:deliverable :WORK }.
>-------
>
>... but that seems like a big hassle when the RDF
>parser could have just spit out rdfs:member triples
>along with rdf:first and rdf:rest triples.
>
>In sum...
>
>  (1) pls specify that rdfs:member connects collections
>  to their members

First we gotta say what counts as a member. Then we gotta have a 
*recursive* definition of member. Then do we have to somehow justify 
that this definition always works? even on forked-tail lists and 
other exotica?

>
>  (2) specify that parsers spit out rdfs:member triples
>  when parsing Collection syntax.
>
>  (3) specify that parsers spit out rdfs:member triples
>  for old-style _n Containers too.

Why would a parser do it? Those are entailments (via rdfs:subPropertyOf).

Pat

>
>Hmm... I suppose there could be missing member triples
>if the RDF/XML syntax used explicit first/rest or
>explicit rdf:_n triples. But that seems acceptable
>to me.
>
>This idea is probably too late to be worth the trouble.
>But it comes from real implementation experience,
>and I figure I ought to report the experience,
>even if the WG decides it's not worth the change
>at this point.
>
>
>--
>Dan Connolly, W3C http://www.w3.org/People/Connolly/


-- 
---------------------------------------------------------------------
IHMC					(850)434 8903 or (650)494 3973   home
40 South Alcaniz St.			(850)202 4416   office
Pensacola              			(850)202 4440   fax
FL 32501           				(850)291 0667    cell
phayes@ai.uwf.edu	          http://www.coginst.uwf.edu/~phayes
s.pam@ai.uwf.edu   for spam
Received on Monday, 24 February 2003 15:46:03 EST

This archive was generated by hypermail pre-2.1.9 : Wednesday, 3 September 2003 09:55:52 EDT