Re: A basis for convergence and closure?

On 2002-02-05 17:48, "ext Jeremy Carroll" <jjc@hplb.hpl.hp.com> wrote:

> 
>>> 
>>>> 5:    Do we allow S-A idiom?
> 
>> 
>> From the ontology maintenance perspective, I don't
>> think that the problem of multiple vocabularies goes
>> away with the combination of the proposed pair
>> of bNode idioms and the S-A idiom, so all my earlier
>> expressed concerns apply as well.
>> 
> 
> The new MT makes them interchangeable. There would not be a multiple vocabs
> problem with Pat's magic dust.
> 
> In particular we would have
> 
> <rdf:int.map> <rdfs:subPropertyOf> <rdf:value> .
> 
> etc.
> 
> I don't like the URI bloat, but apart from that I see S-A as attractive. I
> believe we could use S-A without the URI bloat, but we're then getting onto
> third-order considerations.

The tradeoff has always been between multiple URIs
or multiple vocabularies. I don't consider either to
be reasonable -- especially since the current modified
TDL/S-P/++ proposal on the table makes either choice
unnecessary.

> I think the S-A idiom may well prove more attractive to the document writer
> than the D idiom. (It is less characters to type).

We have to look at the sum total of usability. I still
think the D idiom is "lighter", all things considered.

There is also the loss of the elegant symmetry between
the global and local variants differing only by
the presence or absence of the rdf:dtype arc(s) and
one could easily make all implicit typing explicit
by simply adding those arcs, if one wanted to e.g.
export some knowledge to an environment which either
did not support global schema knowledge.

Such relationships between the S-A idiom and the
TDL/S-P/++ idioms are far far more complex. Why
burden the users. Common perception is that RDF
is already hard to understand. Please let's keep
it simple.

But, I've stated my position, so I won't belabor the
point further.


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

Received on Tuesday, 5 February 2002 11:04:30 UTC