Thread Links Date Links
Thread Prev Thread Next Thread Index Date Prev Date Next Date Index

SUO: RE: SUO Comment #3




"West, Matthew MR SSI-GPEA-UK" wrote:
> 
> Dear James,
> 
> The point to be careful of in the comment below is not to interpret the
> requirement as meaning that the SUO should not contain anything that cannot
> be "compiled" into a more restrictive format such as XML. Otherwise there is
> no problem.

Matt makes a good point, but there is another point lurking here.  In
order for the statement "compiled into X" (X /= SUO language) to be
meaningful, we have to define what "compiled" means.  For example, it
is trivial to simply strip out the type-like constants out of the SUO
ontology and embed them in RDF, or XML-schema, or RDF-schema, or 
whatever.  However, you'll lose semantics in the process.  So you have
three choices: 1) characterize just what kinds of loss occur and whether
or not you can live with them; 2) make sure the target language is
as expressive as the SUO standard; or 3) drop the requirement and face
the possibility that nobody will use the SUO. 

Now, (2) isn't going to happen with any of the RDF-y/XML-y standards
likely to gain popular acceptance in the near future, and I assume that
nobody wants (3) because otherwise we'd be wasting our time.  That 
leaves (1) - a monumentally hard job.  Comments?

  ...bill