RE: SV: A certain difficulty - lack of action!

Hi Jonas,

Jonas said:
Would this be the same as

<rdf:Description about="your resource link here">
   <dc:description>your description</dc:description>
   <dc:author>the author here</dc:author>
                 etc....
</rdf:Description>


Didier replies:
the answer is a simple "no".

Why? because:

a) xlink has a lot of chance to be implemented because linking is
fundamental to the Web. Can we imagine the web without linking? therefore,
linking is fundamental to the XML based web.
b) an xlink interpreter can interpret this as a link. We are linking to
something
c) if you add to the link element the rdf:type you then have your element to
inherit from two AF. 1) a link AF, 2) an RDF schema AF. For example the
following expression is at the same time a link, a content defined by the
RDF schema, an our own element.

<myElement xlink:type="simple"
           xlink:href="your resource link here"
	     rdf:type="DublinCore">
  <dc:description>your description</dc:description>
  <dc:author>the author here</dc:author>
  		etc....
</myElement>

Thus, a link engine can be used to navigate (or simply express a link), an
RDF schema can be used to define the object's class and therefore determine
the element content which can be meta information about the resource, and
finally our own attributes provide what needed by our model. We have then
the following model

   link
     |
     |
   rdf schema
     |
     |
   our element

we can say that the type element indicate that our element inherit from two
different architecture or inherit from two kind of objects, one providing
linkage facilities and which is recognized by a link engine, the second by
an RDF schema aware engine and the third by our own interpreter.

So Jonas, this is a totally different ball game we are talking here,

a) we re-use different technologies or interpreters. The interpreters
provide the behavior, the XML based language provide simply a domain
language in need to be interpreted.
b) We stop seeing the world as a whole or nothing and start practicing
re-use by combining different domain languages and interpreter attached to
these.
c) we resolve the overlap between xlink and rdf.
d) we have a second chance to be understood with a different language.

lesson to remember from SGML:
a) people said it was hard to learn - we hear the same thing with RDF
b) people said that the syntax is to heavy (DTD needed, omitag, etc...) - we
created XML to redo SGML, simplify it and get an other chance to present the
same concept with a new image (PR, selling the idea, etc...)
c) a useful construct was the architectural form - xlink and rdf schema
through the "type" attribute allow to implement a kind of architectural form
and thus provide a lot of diversity when creating a domain language. A
particular domain language can inherit the characteristics of other domain
language and thus we can create more sophisticated solutions by assembling
different interpreter and different level of interpretation.
d) a lot of other lesson we should learn but human are humans and not every
body is learning at the same pace.

So, if we where to do an RDF version 2, it would be useful to have some
synergy between xlink and rdf. Very powerful application can be created by
allowing the same document to be interpreted in different ways.

Finally the about="URI" is a nice attribute but an xlink:href="URI" is maybe
more useful because it is a link. Both construct are used to point to a
resource but the latter can be interpreted by an xlink interpreter not the
former. At least it worth to think about it. So Sergey question was a clever
one.

cheers
Didier PH Martin
----------------------------------------------
Email: martind@netfolder.com
Conferences: Web Chicago(http://www.mfweb.com)
             XML Europe (http://www.gca.org)
Book: XML Professional (http://www.wrox.com)
column: Style Matters (http://www.xml.com)
Products: http://www.netfolder.com
b)

Received on Monday, 28 February 2000 17:28:34 UTC