Re: Let's get the Literals out of the RDF Graph

Hi Sandro,

[...]
> content negotiation thing is...  difficult.  I've tweaked it
> a bit, but since it worked for me the first time, I don't
> know how it's changed.

I don't know how it's changed either, but at least it works for me now :-)
Perhaps it's a bug in IE6.

> > Also, in that file, could you please change rdf:ID to
> > rdf:about? [...]
>
> Interesting.   I just copied DAML for this.   Do you
> want to suggest they change?    I took your advice.

I will suggest that they change... I've already asked for the RDF Schema
namespace to s/ID/about/ [1], and it's been on the issues list for ages
[2].

[...]
> So what I'm saying is that an n3 parser should pretend
> to treat "abc" as shorthand for (string:U0041 string:U0042
> string:U0043) [which is shorthand for some daml:List
> stuff] and then *HORRAY* we have nice, simple, useful,
> interoperable, standard semantics for RDF Literals.

That seems fair enough. Don't forget that there's already a URI scheme for
literals, "data:". Instead of str:U0043, why not just use:-

   data:text/plain;UTF-8,a

and for "abc":-

   data:text/plain;UTF-8,abc

which still gets the notion of literals out of the model?

Note also that CWM has a log:concat property with which one can bring
literals together:-

   this log:forAll :x .
   { :x string:concat ("a" "b" "c") } log:implies { :abc :is :x } .

although that retains literals in the model.

I like the idea that there won't be any literals in the model, but will it
work for datatyped strings?

   :year rdfs:range :Year .
   :SomeEvent :year (str:U0032 str:U0030 str:U0030 str:U0031) .

implies:-

   (str:U0032 str:U0030 str:U0030 str:U0031) a :Year .

implies:-

   :Year rdfs:subClassOf str:String .

Which I guess is reasonable, just as any current datatype is a subClassOf
rdfs:Literal. Are there any further advantages to this idea other than it
gets rid of LV from the model theory?

I also think it exposes a certain amount of nitpicky pedanticism behind the
current rdfs:Literal threads... I've not had more problems in implementing
literals in parsers than I have bNodes :-) Of course, until any behaviour
is mandated by an RDF specification, it's not going to be consistently
implemented (and even when it is, it's touch and go!).

[1] http://lists.w3.org/Archives/Public/www-rdf-comments/2001JulSep/0300
[2]
http://www.w3.org/2000/03/rdf-tracking/#rdfms-difference-between-ID-and-abo
ut

--
Kindest Regards,
Sean B. Palmer
@prefix : <http://webns.net/roughterms/> .
:Sean :hasHomepage <http://purl.org/net/sbp/> .

Received on Friday, 5 October 2001 16:38:52 UTC