Re: A certain difficulty

----- Original Message -----
From: David Megginson <david@megginson.com>
|
| You must keep more than that -- at the very minimum, you probably keep
| a fourth property indicating whether the RDF object is literal text,
| non-RDF XML markup, or a reference to another resource.  Otherwise,
| there would be no way to distinguish the string "http://www.foo.com/"
| from the resource "http://www.foo.com.

Yes that's right.

Aside: At one point our simplified syntax allowed XML attributes to
represent reference properties as well as literal properties.  So we
couldn't complete the mapping to RDF without consulting the metadata for
the property.

Lots of XML languages have this same context dependence (i.e. you must
recognise the attribute to determine if the value is a literal or a
reference).  However, RDF syntax avoids it on principal.

This is often the root cause of extra verbage in an RDF serialization
relative to a custom-designed XML language.  And that is what
application designers seem to dislike when I contrast an RDF solution
with a custom XML solution.

Hence my earlier speculation that we will end up needing a whole
technology to map XML to RDF rather than a single, standard,
context-free mapping.

Regards,

Arnold deVos
Langdale Consultants
adv@langdale.com.au

Received on Friday, 25 February 2000 18:40:16 UTC