Re: Q to implementers: Resource identifiers - XML Names and/or (concatenated) URIs? (was RE: rdfs.isDefinedBy...)

> 
> (This may end up as a double-post, so if it does, please disregard whichever
> version appears second)
> 
> Setting aside for the moment the whole issue of namespace URIs and their
> potential relation to RDF Schema documents and the network resolution
> thereof I have a question for RDF application implementers out there
> regarding their use of Resource identifiers in the form of XML Names vs.
> concatenated URIs:
> 
> Since it is obvious that the working group(s) responsible for RDF for one
> reason or another (i.e. current charter) are not in a position to properly
> address RDF's current mangled interpretation of Namespaces in XML and other
> identifier-related issues such as:
> 
> - if RDF/XML requires prefixes before each attribute, it is simply not
> Namespaces in XML -compliant XML

I think you state this too strongly, but I have long argued that the current 
requirement of prefix before all attributes falls foul of the *spirit* of 
XMLNS 1.0, and of the almost universal practice of XML namespaces.

I get nowhere with this argument.  Members of the RDF Core WG effectively told 
me: you're too late; we've made our decision; we don't expect to revisit it; 
deal with it.  So I'm dealing with it.


> - the suggested XML Names -> URI concatenation process produces collisions

I don't see that it would make collisions inevitable.  I would consider such 
collisions from namespace concatenation to be the error of the document 
author.  RDF/XML does make this matter awkward, but again I think it would be 
too strong a statement to say that RDF/XML makes naming collisions inevitable.


> - the suggested URI -> XML Names splitting method is so flawed as to not be
> responsibly implementable (i.e. you can't generate valid LocalParts by
> splitting on non-Name characters, but can if splitting on non-NCName
> characters. However, using the latter method produces differently invalid
> results, e.g. if splitting
> urn:NewsML:afp.com:20000811:010607144425.x6pxrl6k:1)

Yes.  We have run into this problem quite a bit at Fourthought because we use 
UUID URNs in many projects.  The only solution we have found is a horrible 
hack: to give resources UUIDs, and also aliases in forms that are compatible 
with RDF/XML, and then introduce equivalence statements between the ID and 
alias.

I think the WG desperately needs to fix this.


> Without getting into a huge discussion about each of the example bullets
> above (unless you have specific comments regarding their effect on your
> implementation) and without starting yet another discussion about how simple
> it would be to fix these issues and how many benefits could be reaped from
> their correction if only an RDF 1.1 or some such could be created that
> natively includes the concept of namespaces, I'd like to hear how
> implementers have dealt with these issues in light of the fact that the
> Working Group(s) can't/won't. Are you fully supporting Namespaces in XML in
> your serialization? Internally? Are you following the WG-recommended
> concatenation and splitting processes? If not, what are you doing instead?

Mostly, we've been scrambling to balance common-sense solutions with whatever 
hacks get the job done.  I've given examples to your specific questions.  
RDF/XML is nice because it is quite flexible.  It is nasty because it involves 
a frightful number of hacks and inconsistencies with general XML practice.

I have come to think that a much more important task than setting a canonical 
RDF/XML interchange format is developing a specification for translating RDF 
models to user-defined XML formats, and vice versa.  After all, I think 
mappings from XML formats in general use is the realistic future of RDF.


-- 
Uche Ogbuji                                    Fourthought, Inc.
http://uche.ogbuji.net    http://4Suite.org    http://fourthought.com
Track chair, XML/Web Services One (San Jose, Boston): 
http://www.xmlconference.com/
DAML Reference - http://www.xml.com/pub/a/2002/05/01/damlref.html
The Languages of the Semantic Web - http://www.newarchitectmag.com/documents/s=
2453/new1020218556549/index.html
XML, The Model Driven Architecture, and RDF @ XML Europe - 
http://www.xmleurope.com/2002/kttrack.asp#themodel

Received on Sunday, 9 June 2002 15:21:47 UTC