W3C home > Mailing lists > Public > www-rdf-interest@w3.org > September 2000

Re: Nodes and Arcs 1989-1999: WWW history and RDF (revisited...)

From: Dan Connolly <connolly@w3.org>
Date: Wed, 27 Sep 2000 22:27:44 -0500
Message-ID: <39D2BAB0.34E41CE9@w3.org>
To: www-rdf-interest@w3.org, danbri@w3.org
> From: Dan Brickley (Daniel.Brickley@bristol.ac.uk)
> Date: Wed, Sep 27 2000
> I've my hands full with other RDF docs right now, so figured it was
> about time I circulated this sketch at a high-level background/context
> paper on RDF rather than sit on it still longer.
> 
>         http://www.w3.org/1999/11/11-WWWProposal/thenandnow

This is an interesting data set. I spent some time noodling
on it too. I took a different approach to transcribing
the figure in RDF: I tried to represent exactly what's
in the picture, no more, and no less:

[1]	http://www.w3.org/2000/08/proposal43/proposal.rdf
	revision 1.1 date: 2000/08/08 13:44:41

For reference, the RDF stuff in your paper is:

[2]  http://www.w3.org/1999/11/11-WWWProposal/proposal.rdf

Some of the differences:

[2] <!-- it was created by a Person uniquely identified as... 	-->
[2] <dc:Creator>
[2] 	<util:Person
[2]  web:about="http://www.w3.org/People/all#timbl%40w3.org"/>
[2] </dc:Creator>

but the URI http://www.w3.org/People/all#timbl%40w3.org
is nowhere in the figure. What's in the figure is:

[1] <Document web:ID="thisDocument">
[1] <util:label>This document</util:label>
[...]
[1]   <Person util:label="Tim Berners-Lee">
[1]    <wrote web:resource="#thisDocument"/>
[1]   </Person>

I did make up the local identifier "thisDocument",
because an ID is necessary to serialize the information,
but the label is right there in the figure.
The name "Person" isn't in the figure literally,
but I think the concept of Person is implied
by the org chart diagram idiom.

It might be interesting to annotate the proposal and
relate its nodes to nodes like
http://www.w3.org/People/all#timbl%40w3.org , but
I took an approach of "this is what the figure says"
and kept "this is what I know about things in
the figure" out of it.

Excercise for the reader #1: separate out
the post-hoc knowledge in [2] and merge it,
by machine, with the original knowledge in [1].

Another difference of a similar sort... you
seemed to be avoiding anonymous nodes; I used
them liberally, except where I needed an ID
to make up the difference between a sequential
tree and an arbitrary graph:

[2] <!-- the subject of the document is... 			
[2] -->
[2] <dc:Subject>
[2] <!-- ...a example of a LinkedInformationSystem -->
[2] 	<util:LinkedInformationSystem
[2] web:about="http://fictional.w3.org/entities/MESH">

vs.

[1] <Document web:ID="thisDocument">
[1]     <util:label>This document</util:label>
[1]     <describes>
[1]       <Proposal web:ID="mesh" dc:title="Mesh">


A stylistic difference: the figure doesn't seem to
appeal to any established vocabulary for its
arc labels and such, so I included the schema
info in [1] rather than putting it in
a separate file. While I was at it,
I did some noodling on including visualization
info in schemas:

[1]  <util:Property web:ID="describes"
[1]    util:label="describes">
[1]    <util:domain web:resource="#Document"/>
[1]    <util:range web:resource="#Document"/>
[1]
[1]    <kpic:arcStyle>@@default + get label from
util:label</kpic:arcStyle>
[1]  </util:Property>

I also tried to do some swi-prolog hacking on the data
to generate SVG or something, but I couldn't get
my subProperty and subClass inference rules to
terminate, no matter how hard I tried. I'm not
much of a prolog hacker yet.

I got as far as converting the arcs to prolog ala:

says(replyX, 'http://www.w3.org/1999/02/22-rdf-syntax-ns#type',        
'http:proposal43/proposal.rdf#thisDocument',   
'http:proposal43/proposal.rdf#Document')
.
says(replyX, 'http://www.w3.org/2000/01/rdf-schema#label',     
'http:proposal43/proposal.rdf#thisDocument',    "This document").

and I tried to implement RDF schema:
http://www.w3.org/2000/08/proposal43/rdfs.pro

but my subProperty rules kept looping.


misc bugs in [2]:
dc:Title -> dc:title

-- 
Dan Connolly, W3C http://www.w3.org/People/Connolly/
Received on Wednesday, 27 September 2000 23:29:32 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:51:44 GMT