Re: XML to tagtriples (and mapping heuristics)

Laurent Szyster — Tue, 20 Sep 2005 14:38:02 -0000

Simplification of RDF/XML without URI constraint for subject, predicate or objects is a good idea. But I have to disaggree on this:

"statement order in a graph is maintained"

Because there is little semantic to be found in a sequence of XML elements and it's a big restriction on its practical usage.

An XML element tree can represent sets of hierarchical relationships, display ordered sequence, but making an assumption about the meaning of a sequence restricts the liberty of the author to change this order of elements. That's exactly why XML is so popular: it is less complex than SGML and provides a more flexible text format than *fixed* position records or flat database interchange formats.

The first example XML string with your name and e-mail address could be rewritten as:

pdawes@users.sf.net

explicitely assign the name "Phil Dawes" as the context of the e-mail address .

Or vice-versa, using a well-supported URL,

Phil Dawes

A metabase (a semantic server, a metada management system) should not re-articulate the information submitted. It is up to the application or its users to produce well articulated statements.

By the way, I tried to send you a private mail to but that did not work out so well. The mailer at sf.net bounced it back.

Kind Regards,

Phil Dawes' Stuff - Latest Comments in XML to tagtriples (and mapping heuristics)

Re: XML to tagtriples (and mapping heuristics)