OASIS Mailing List ArchivesView the OASIS mailing list archive below
or browse/search using MarkMail.

 


Help: OASIS Mailing Lists Help | MarkMail Help

 


 

   The triples datamodel -- was Re: [xml-dev] Semantic Web permathread, ite

[ Lists Home | Date Index | Thread Index ]


On Jun 3, 2004, at 4:26 PM, Joshua Allen wrote:
>
> Well, the big difference is that RDF is a "triples" data model, while
> XML is hierarchy.  Setting aside the issues of how you *serialize* the
> triples, it's inevitable that "triples" will win in these "semantic"
> scenarios.  OSAF Chandler is based on "triples", as is Longhorn's 
> WinFS.
> Both are essentially "personal semantic web stores".  Triples+URIs is
> how you bootstrap the "personal semantic web store" and make it
> universal.
>

I don't know about that "inevitable" part -- the triples model is more 
or less the binary relational data model, which has been around a long 
time and has not exactly set the world on fire.  C. J. Date commented 
on this recently at http://www.dbdebunk.com/page/page/1147347.htm  
Well, actually some anonymous person "RM" gets in the best lick: 'In 
many cases, however, binary is wrong and does not represent the world 
of discourse in which we do our thinking and understanding.  If I'm 
wrong then I'll start all introductions as, "This is Matt.  Matt has a 
last name.  That last name is Brown.  Matt has a middle name... If a 
data model forces us to record information in a format more difficult 
to use than human language then perhaps it is not a good choice"  '

That may or may not be overcome by all the UIs that will support WinFS 
in Longhorn, or OWL's more human-friendly but reducible-to-triples 
syntax, we shall see.   Likewise we shall see whether the whole 
paradigm of reducing information to, uhh, syllogisms (dipping toe in a 
pool of flame that Cory Doctrow created!) is going to be fruitful 
outside  some limited domains, such as the one I referred to earlier:  
In an enterprise information system, you NEED to do exactly what the 
triples model enables, e.g. relate a customer information "resource" to 
an account status resource in a completely rigorous way (and 
irrespective of whether the "resources" are in an RDBMS, XML text file, 
web service invocation, or whatever).   I'm all for applying the 
semantic web paradigm where it fits well, and information integration 
of divers information sources that are individually well defined seems 
to be one of them.

On the open web, or in a personal information system, I'm not so sure 
-- recall that the previous thread *immediately* turned to analogy 
rather than formal logic, i.e. the question of whether Google is to the 
Semantic Web as the Web is to Compuserve, or maybe as Cleveland is to 
Xanadu.  Intriguingly, the analogy works whether "Xanadu" refers to 
Kubla Khan's fictional pleasuredome or to Ted Nelson's hypothetical 
hypertext web!!!  If that doesn't shake one's faith in triples, I don't 
now what will :-)

p.s.  - Given the prevalence of analogy over logic in real human 
thought and discourse, does anyone know whether formal ontologies can 
effectively represent analogies?  Hmml, Google finds 
http://www.qrg.northwestern.edu/papers/Files/IAAI02KForbus.pdf ...





 

News | XML in Industry | Calendar | XML Registry
Marketplace | Resources | MyXML.org | Sponsors | Privacy Statement

Copyright 2001 XML.org. This site is hosted by OASIS