Some questions

Didier PH Martin martind at netfolder.com
Wed Dec 1 22:40:44 GMT 1999


Hi David,

David said:
The problem is that the RDF-Syntax spec confounds even its bravest
readers by trying to do two things at once:

a) define a model and syntax for exchanging object-oriented
   information in XML; and

b) apply the model and syntax to the problem domain of representing
   knowledge about Web pages.


Didier reply:
I guess that what is causing the confusion right at the beginning is the
triad stuff. Instead, it would have been more useful to present the concept
or the atomic unit as a record or an object without the methods. But,
contrary to RDB records, there is inheritance relationship between the RDF
entities.

So, instead of a model based on the triad "object property value" as an
atom, it would have been a lot easier to say "object as a collection of
properties/values". a schema is like a template, an object is just this
template with slots filled (values added to properties). A template can
inherit from an other template.

But, from the RDF document point of view, what we always see is the objects
and their associated collection of properties/values.

Its funny, one of the ancestor of RDF is the MCF (not from Netscape but from
Apple research/Talva ref - http://www.netfolder.com/SDK/MCF.htm and
http://www.netfolder.com/SDK/MCF11.htm). This ancestor language was designed
as a simple set of units and each unit having a collection of
property/value. It seems that instead of being simplified it just became
more obscure. its sad, it is so easy to use when well explained and
understood.

Obviously the choice of word like "about", "description" lead to think of
data about something instead of the data being _the_ something. This is why
I use a structure like this:

<rdf:description id="MyID">
<location>http://www.netfolder.com</location>
... etc...
</rdf:description>

What are the gains?
a) the object is location independent.
b) its location is just an other property (and in fact it is a property)
c) Then, it is simply an object without any reference, what is giving
references is the properties.
d) I can relate the object to others with properties.
e) its easier to remember and understand.
d) this is the object not an object about an other object.

I discovered that, in some cases I want to express certain object's
relationship like for instance a hierarchy. Let's say that I want to
transfer the content of a directory service from one place to an other, then
in that case:

<rdf:description about="context1/context2/context3">
<location>http://www.netfolder.com</location>
... etc...
</rdf:description>

That way,  all objects are transported as a small independent hierarchy. The
hierarchical relationship is expressed with the string in the about
attribute. And because we are used to express relative position in a
hierarchy with "/" I use it.

I didn't used it for other kind of data structures.

I do not know what went wrong??? Probably OCCAM was in vacations :-))

Cheers
Didier PH Martin
----------------------------------------------
Email: martind at netfolder.com
Conferences:
Web Boston (http://www.mfweb.com)
Markup 99 (http://www.gca.com)
Book to come soon: XML Pro published by Wrox Press
Products http://www.netfolder.com



xml-dev: A list for W3C XML Developers. To post, mailto:xml-dev at ic.ac.uk
Archived as: http://www.lists.ic.ac.uk/hypermail/xml-dev/ and on CD-ROM/ISBN 981-02-3594-1
To unsubscribe, mailto:majordomo at ic.ac.uk the following message;
unsubscribe xml-dev
To subscribe to the digests, mailto:majordomo at ic.ac.uk the following message;
subscribe xml-dev-digest
List coordinator, Henry Rzepa (mailto:rzepa at ic.ac.uk)





More information about the Xml-dev mailing list