[XML-DEV Mailing List Archive Home] [By Thread] [By Date] [Recent Entries] [Reply To This Message]

RE: The triples datamodel

• To: "'Hunsberger, Peter'" <Peter.Hunsberger@S...>, martind@n..., XML Developers List <xml-dev@l...>
• Subject: RE: The triples datamodel
• From: "Bullard, Claude L (Len)" <len.bullard@i...>
• Date: Wed, 9 Jun 2004 09:57:22 -0500

And negotiate.  Sorry, Elliotte, but we really do negotiate. 
Hopefully, we have a standard schema.  Yep, we really do use 
those.  For the industry I'm in, that is currently Global 
Justice XML.   Here's an example.

A programmer develops a web service for returning arrest 
information.  Arrest information is related to incident 
information, but at the heart of both is a Name entity. 
To do this, a PersonType is defined because names can name  
businesses and people.  Now comes the fun. 
Another entity is Alias (the kind an arrestee gives when 
they want to disguise their identity).  Is Alias just 
another name or is it a name of type PersonType?  

The naive answer proves to be wrong.  JusticeXML gets 
this definition right, but without it to appeal to, 
we have to negotiate for much longer. Costs go up 
for even a single negotiation so one can envision 
the cost savings as hundreds of producers and 
consumers come online.   The tradeoff for not having 
to buy wall-to-wall systems from a single vendor 
IS commodization pressure by standardization.  

len


From: Hunsberger, Peter [mailto:Peter.Hunsberger@S...]

The rush to standardization of everything XML should tell us that people
value meta-data with broad scope. However, people constantly try to get
ahead of the game and institutionalize their metadata without having any
proof of its value. Personally, I'll write no schema before it's time,
and never consider a schema the final world...
 

• Prev by Date: RE: The triples datamodel
• Next by Date: Re: The triples datamodel -- was Re: SemanticWeb permathread, iteration n+1
• Previous by thread: RE: The triples datamodel
• Next by thread: Fun With Mathematics
• Index(es):
  • Date
  • Thread