[SCL] Issues and proposals (details)

Tanel Tammet tammet at staff.ttu.ee
Mon Jan 13 06:50:27 CST 2003


Hi,

John F. Sowa wrote:

> PH> RDF has an XML syntax already defined. It is hideous and irrational,
> > but it is now a standard. There is also an obvious translation of RDF
> > into SCL (positive binary existential conjunctions) . Should we try to
> > produce an XML syntax for SCL which preservers the XML syntax of the
> > RDF-definable fragment? It would not look at all like the DTD that
> > Chris suggested.
>
> I agree that we need an XML representation at the level that Chris
> suggested, but there is no reason why it has to have any syntactic
> resemblance to RDF.   As long as we provide a definition of how the RDF
> subset version can be translated to and from the more rational syntax
> that we propose, that should be sufficient.


I'd support John here. RDF is not designed with ordinary FOL in mind. While
IMHO it is extremely important that we can translate to and from RDF, I see
no need in giving a "default" XML encoding in XML.

Rather, we should give a simpler (as simple as possinle) and close-to-lisp
default encoding in XML, bearing in mind the later need to do "scl 
XML"<->RDF
conversions.

Where I am a bit worried concerning XML is that IMHO we should keep
the information content in lisp and XML syntaxes as similar as possible.
When we design the XML syntax in a way that the XML explicates a lot
more than the lisp syntax (it hardly explicates less) then we have nasty
unbalance. Why this is nasty: I do not really know, but it just feels 
unpleasant :-)

>
> Our claim is that we can provide a common semantics that can support
> interoperability with any notation for any subset of classical FOL that
> conforms to the semantics.  If someone wants to invent a truly ugly
> syntax, as the RDFers have done, they have to live with it.  We can
> support it semantically and show how to translate to and from it,
> but we don't have to adopt it.

IMHO it would be good to have an annex with the algorithm for
converting "scl XML" <-> RDF.

Regards,
             Tanel Tammet






More information about the Scl mailing list