RE: Surface vs. Abstract Syntax, was: RE: What do the ontologists want

>perhaps the greatest benefit of XML is that its surface syntax directly
>represents its abstract syntax, and for someone familiar with XML, this
>means that one can look at a document, even in the absense of a schema, and
>get a pretty good idea of its structure.

Another pragmatic benefit is that we already have pretty good tools for XML.

>A couple of points. First the RDF XML syntax does not make full
>use of XML's
>abilities to represent structured data. Every XML element node
>already _has_
>a list of child nodes (e.g. the DOM NodeList). However the RDF abstract
>syntax does not naturally maintain order so the syntactic hack is
>introduced
>converting <rdf:li> into <rdf:_1> <rdf:_2> etc. This is _painful_ to my XML
>accustomed eye.

Agreed, this is really ugly. In fact RDF really seems to have a problem with
containers generally.

>Second: When I did alot of LISP programming, I recall it often
>took a bit of
>work to create complex datatypes out of LISP. If you recall of Dan Corkill
>and John Lowrance's GRASPER language (UMass c. 1981), one can indeed
>represent nodes, arcs, spaces etc in LISP - on the other hand adding trees
>and maps as native datatypes _tremendously_ increases the speed of such a
>language. XML naturally represents trees and somewhat naturally handles
>maps.

Looking up GRASPER now...ooh, pretty pictures :
http://www.ai.sri.com/~grasper/

also ftp.ai.sri.com/pub/papers/karp-grasper-tr.ps.Z

>> This proposal for thinking about RDF would make
>> the RDF into an implementation language in which to write expressions
>> in some other language (which would have a semantics), rather than
>> just adding a handy datatype to RDF, in the way that DAML adds lists
>> as a shorthand for nested-triple structures of a certain kind (which
>> would be too painful to spell out in detail).
>>
>
>again, the data structures naturally represented in an RDF abstract syntax
>are orthogonal to what semantics are assigned to the expressions. put
>another way, whether we code these using s-expressions or XML is not
>relevent. but _having_ XML already provides an entirely natural way to
>represent lists. what i am suggesting is that this concept could be
>reflected in the RDF abstract syntax -- a trivial way to do this might be s
>:= <p,s,o,i> where i is the index of the statement as reflected by document
>order.

Could you please (have you already?) expand on this last idea - I'm not sure
what you mean.

Received on Friday, 18 May 2001 22:24:07 UTC