[Fwd: Re: status of rdf:text] - for discussion on DTB tomorrow

As for rdf:text, see below my earlier mail to the chairs:

Reading instruction:
please read that mail below first, before you check the comment next ;-)

One small comment on it to be added:

> In RIF, you can formulate arbitrarily mor complex expressions (e.g. Y 
> being odd, etc. by calling respective built-ins)... Thus, I suggest to 
> keep facets separate.

that is not entirely true, particularly, I can't imagine at the moment 
whether we could emulate *negative guards* for facetted primitive 
datatypes without having negation.

best,
Axel


-------- Original Message --------
Subject: Re: status of rdf:text
Date: Thu, 18 Sep 2008 18:22:43 +0100
From: Axel Polleres <axel.polleres@deri.org>
Organization: DERI, National University of Ireland, Galway
To: Sandro Hawke <sandro@w3.org>
CC: Jie Bao <baojie@cs.rpi.edu>, team-rif-chairs@w3.org,
team-owl-chairs@w3.org
References: <26835.1221747240@ubuhebe>

Sandro Hawke wrote:
> How stable is rdf:text?  Can we do reviews and get it published as a WD
> in the next two weeks?  (Any open issues should be marked in the draft.)
> 
>     -- Sandro

Just went through [1] again. The spec is in stable status from my point
of view and ready for WD publication, except the discussion about facets:

   "It is currently not clear whether this document will contain a
    definition of facets on rdf:text."

cf [2], which I sugget to discuss at our (RIF) next F2F.

Maybe the example/explanation on why we understand the set of characters
to be infinite in the preliminaries should be moved to a remark/footnote
in the end of the document, since at this point, it might confuse
readers not concerned about this detail.


For clarification about the facet issue, see below.

with best regards,
Axel

1. http://www.w3.org/2007/OWL/wiki/InternationalizedStringSpec
2. http://www.w3.org/2007/OWL/wiki/Syntax#Datatype_Maps

Explanation facets issue:

OWL2 allows, similar to XML Schema's simple datatype definitions to
define "facets" of primitivce datatypes based on restrictions such as
e.g. minInclusive, max:Exclusive for numeric datatypes or
minLength, maxLength, Length, pattern, for string/text based datatypes.

The OWL people asked whether they can/shall include the list of allowed
  facets in the basic datatype definition. I somewhat oppose that and
think [1] shoukd be kept "as is" and the facets should stay in [2],
since in RIF we can emulate all these facets by built-in predicates,
example below for the interested...



OWL2 facet example:

PropertyAssertion( a:hasAge a:Meg 17^^xsd:integer )

SubClassOf(
     SomeValuesFrom( a:hasAge
        DatatypeRestriction( xsd:integer
           minInclusive 13^^xsd:integer
           maxInclusive 19^^xsd:integer
        )
     )
     a:Teenager
)


I would write this as follows in RIF:


a:Meg[a:hasAge -> 17 ]

Forall ?X ?Y

?X#a:Tennager :-
     And(
          ?X[a:hasAge -> ?Y ]
          External( isInteger(?Y ) )
          External( pred:numeric-less-than-or-equal( 13 ?Y ) )
          External( pred:numeric-less-than-or-equal( ?Y 19 ) )

        )

In RIF, you can formulate arbitrarily mor complex expressions (e.g. Y
being odd, etc. by calling respective built-ins)... Thus, I suggest to
keep facets separate.

Received on Friday, 26 September 2008 16:58:23 UTC