W3C home > Mailing lists > Public > public-rdf-wg@w3.org > May 2011

Re: ISSUE-12: xs:string VS plain literals: proposed resolution

From: Alex Hall <alexhall@revelytix.com>
Date: Mon, 9 May 2011 11:42:05 -0400
Message-ID: <BANLkTi=94RhF0xUAbUd=fmpJPqJTF+JExQ@mail.gmail.com>
To: Steve Harris <steve.harris@garlik.com>
Cc: Andy Seaborne <andy.seaborne@epimorphics.com>, public-rdf-wg@w3.org
Apologies for top-posting, I'm trying to catch up on 3 days' worth of
conversation on the matter.  My thoughts can be summarized as follows:

I think it's a mistake to tell systems to silently convert from one form to
another.  If "a" and "a"^^xsd:string remain syntactically distinct terms,
then systems should allow the user to decide which one to use even if one is
deprecated.  Silently converting one of those forms will really surprise and
annoy any users who do care about the difference.  Since normalization is
optional it also means that the same Turtle file given to two different
conformant systems could result in different graphs.

The decision to normalize to plain literals as opposed to xsd:string
surprised me at first.  It apparently aligns with the decision taken by
RDF2RDB, but seems to be going in the opposite direction from OWL2 and RIF.
 It's disappointing that there's not consistency on the use of typed vs.
untyped literals across the communities.  Since I wasn't at the F2F, I'd be
interested to hear why it was decided to deprecate xsd:string as opposed to
plain literals.

If there truly is a desire across all RDF-based technologies to treat
same-valued plain literals and xsd:string literals as the same term, then I
think we should go ahead and change the abstract syntax in RDF Concepts to
make that so.  Otherwise, as Pat says we essentially do nothing and let
SPARQL decide whether to use {xsd:string}-entailment.  Delaying a decision
on the matter will only allow the confusion to propagate to new technologies
and make it more difficult for future WGs to address the issue.


On Sat, May 7, 2011 at 6:45 PM, Steve Harris <steve.harris@garlik.com>wrote:

> On 2011-05-07, at 14:19, Andy Seaborne wrote:
> > On 07/05/11 05:39, Pat Hayes wrote:
> ...
> >> We can say this all we want, but saying it does not make it true.
> >> Right now, it is false. Those are two triples. If you want this to be
> >> one triple, you need to explain how to rewrite RDF Concepts to make
> >> it come out that way. Good luck.
> >>
> >> Pat
> >>
> >>> Telling the world that "abc"^^xsd:string is a deprecated form of
> >>> "abc" (and systems are encouraged to normalize) is probably the
> >>> best balance between simplification and disruption.
> >>>
> >>>
> >>>
> >>>> Pat
> >
> > To clarify my example:
> >
> > That's not a graph - it's a serialization.  My question is what graph
> does it produce when parsed given the proposed text.
> >
> > If it were:
> >
> > :x :p "foo" .
> > :x :p "foo" .
> >
> > then it produces a graph with one triple.
> >
> > :x :p "foo" .
> > :x :p "foo"^^xsd:string .
> >
> > Is the effect of "and tell systems to silently convert xs:string literals
> to plain literals without language tag" supposed to cause one or two
> triples?
> My recollection / understanding from the F2F was that the above
> serialisation would result in a graph with one triple.
> - Steve
> --
> Steve Harris, CTO, Garlik Limited
> 1-3 Halford Road, Richmond, TW10 6AW, UK
> +44 20 8439 8203  http://www.garlik.com/
> Registered in England and Wales 535 7233 VAT # 849 0517 11
> Registered office: Thames House, Portsmouth Road, Esher, Surrey, KT10 9AD
Received on Monday, 9 May 2011 15:42:33 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 17:04:06 UTC