Re: Tidy literal nodes incompatable with current MT?

On 2002-01-28 13:57, "ext Jeremy Carroll" <jjc@hplb.hpl.hp.com> wrote:

> 
> 
> Patrick
> 
> you are losing my support this morning :)

Oh dear me, was it something I said?!  ;-)
 
> TDL inevitably requires interpreting one occurrence of a literal differently
> from another occurrence of the same literal. e.g. the literal node label
> "1984" will be interpreted as a string when it is a book title, and as an
> integer when it is an age.

I agree.

> If this literal happens to be stored in the same memory location as the
> other is completely irrelevant. In as much as they are being interpreted
> differently they *are* different literals.

Exactly.

> Thus any implementation detail about whether or not we share storage and
> thus technically about whether the representation of the graph in the
> computer memory is tidy or not is beside-the-point.

Agree.

> What S requires (as you yourself have pointed out) is that there is one
> literal node for each literal label and that has one interpretation. If
> multiple meanings are needed that is up to the application semantics.
> 
> What TDL requires is that different occurrences of literal nodes with the
> same label have different meanings.

Well, different occurrences of literals -- with occurrence being
differentiated by different context, not necessarily whether some
node with that literal label is "shared" in the expression of
different contexts.

What counts is that the context is distinct, and the distinct
context denotes a distinct occurrence of the literal, and that
results in a unique interpretation for that context.

> It is an implementation detail how
> literals are actually store, but the clearest exposition of the TDL position
> is that the graph is untidy. Anything else is mere sophistry.

Well, I guess I don't see a difference. Meaning, that the TDL
interpretation (based on the labels, not the nodes) is the same
whether or not the graph is tidy or untidy -- and that goes for
URI labeled nodes as well. So I don't see TDL *requiring* an untidy
graph, per se.

Of course, as I pointed out in

http://lists.w3.org/Archives/Public/w3c-rdfcore-wg/2002Jan/0327.html

there may be other problems with tidy literals, but I didn't see
any problems insofar as the TDL interpretation was concerned.

Patrick

--
               
Patrick Stickler              Phone: +358 50 483 9453
Senior Research Scientist     Fax:   +358 7180 35409
Nokia Research Center         Email: patrick.stickler@nokia.com

Received on Monday, 28 January 2002 07:25:11 UTC