[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
- From: "Sean B. Palmer" <firstname.lastname@example.org>
- To: "Simon St.Laurent" <email@example.com>,"Bullard, Claude L (Len)" <firstname.lastname@example.org>
- Date: Wed, 30 May 2001 18:04:57 +0100
> RDDL does at least provide an opportunity to gather the
> rest of the parts and make their alignment explicit.
True, but there needs to be some vocabulary of terms for providing
further information on whether the method of constraint is subjective
or objective. For example, XML schema languages such as XSD, TREX,
RELAX and so forth are seen as objective languages, whereas RDF and
XTM are more subjective (although inclusions of XSD and DAML to the
vocabularies means that this doesn't have to be the case... you could
probably use RDF as an XML schema language). UML is a bit of an
oddity - more akin to object oriented programming methods, and hence
procedural rather than necessarily sub/objective.
In XML, what we often have is the namespace, and then a method of
constraint on top of that. There are some complications in the RDF
uses the concatenation style mechanism to create terms for these
"names", and hence methods of constraint can clash, such as in XHTML
where "title" is used as both an element and an attribute. Such a
language would need to specify URIs for inclusion in RDF.
But I digress. We have the names, we have the URIs, we have the
constraints. I agree that mapping between them can sometimes be
useful, but you must consider the layering that goes on - you have the
abstract concepts, inference data and prose on the one hand, and then
you have the syntactic constraints below that. It is usual for one to
follow the other (usually the syntactic schemata following the
abstract stuff), and so the mappings can be one way. You also have to
allow for evolution, which is a consideration.
Sean B. Palmer
@prefix : <http://webns.net/roughterms/> .
:Sean :hasHomepage <http://purl.org/net/sbp/> .