Re: Index of URI Schemes in Notation3

[most of my comments are about the art of meta term selection, not about URIs
per se.  - Al]

At 11:01 PM 2001-03-13 -0600, Dan Connolly wrote:
>"Sean B. Palmer" wrote:
>> 
>> > 23:46:45 <DanC> yes, schemes.html doesn't get enough of
>> > my attention. 1000 brownie-point bonus for an N3 version
>> > of schemes.html
>> 
>>     
<http://infomesh.net/2001/03/schemes/>http://infomesh.net/2001/03/schemes/
>>      - An Index of WWW Addressing Schemes (in Notation3)
>
>Wow! that was quick...
>
>> That'll be 1000 brownie-points please :-)
>
>Well... 500 so far; your transcription loses a bunch
>of info; e.g. the links to all the background documents!
>The 1000 point offer was for something that I could
>generate schemes.html from using cwm and XSLT.
>
>I'd like you to take my jumble of metadata about schemes and
>documents and to separated it into
>
> * descriptions of documents
>
> * descriptions of schemes
>
> * links between schemes and documents.
>
>e.g.
>
><<http://www.ietf.org/rfc/rfc1959.txt>http://www.ietf.org/rfc/rfc1959.txt>
>  dc:title "An LDAP URL Format";
>  :rfc "1959";
>  :content-length "7243";
>  dc:creator [ foaf:name "T. Howes"], [ foaf:name "M. Smith" ];
>  dc:date "June 1996";
>  :specifies
>    [ :schemeName "ldap" ].
>
>
>where
> 
> :rfc rdfs:subPropertyOf dc:identifer.
> :specifies rdfs:subPropertyOf dc:subject.
>
>and rather than saying "a :URIScheme" each time, just say
>
>  :schemeName rdfs:domain :URIScheme.
>
>oh... btw...
>
>  :schemeName a daml:UnambiguousProperty.
>
>and
>
>  :schemeName rdfs:isDefinedBy
<<http://www.ietf.org/>http://www.ietf.org/rfc/rfc2396.txt>
>
>(hmm... is isDefinedBy appropriate when that spec
>doesn't actually specify RDF properties per se?
>or is it just seeAlso?)

The URI syntax RFC is [a feasible entry point for elaborating] the
authoritative source for what a URI scheme name is.  Or the URL [scheme]
registration RFC.  Whatever index we construct in RDF, it doesn't _define_ the
concept "URI scheme name."  It uses and depends on the externally published
concept.  I would have thought in traditional usage that "see also" would
often
be used for informative references which don't have to be followed to build a
fully formed definition.

This reference is a "q.v." normative reference.

Is there already in RDF usage a generally accepted synonym for "q.v."?

The English phrase I have found comfortable as reflecting the relationship we
need to catch (for schema references etc.) is "as used in."  This is a
normative reference to another utterance which is senior in authority to the
current utterance making the reference, as far as the usage of the subject
term
[or more generally pattern] is concerned.

>
>By the way... your N3 transcription claims that I'm
>its author, which is not so; I think what you meant was
>that I wrote X, you wrote Y, and Y is a transcription
>of X:
>
>=======
><> is :xhtml2n3 of
<<http://www.w3.org/>http://www.w3.org/Addressing/schemes>;
>dc:contributor
>     [ foaf:name "Sean B. Palmer"; 
>       foaf:mbox
<<mailto:sean@mysterylights.com>mailto:sean@mysterylights.com>
] .
>
><<http://www.w3.org/>http://www.w3.org/Addressing/schemes>
>  dc:creator
>     [ foaf:name "Dan Connolly"; 
>       foaf:mbox <<mailto:connolly@w3.org>mailto:connolly@w3.org> ];
>  dc:date "1996/01/02 03:37:23".
>
>:xhtml2n3 rdfs:comment "Converted from XHTML to Notation3";
> rdfs:subPropertyOf dc:relation.

Please see/use the dc:source subPropertyOf dc:relation for what you want to
say, here.

Compare with the metadata for the digital talking book standard, where they
have the same issues about documenting a derived work, see

 <http://<http://www.niso.org/Z3986.html>www.niso.org/Z3986.html>.

>========
>
>Also... foaf:name is a handy concept; surely there's
>a standardized version of it somehwere, no?
>Isn't there a p3p property
>you could use? grumble... P3P isn't mentioned in
> <http://www.w3.org/RDF/#projects>http://www.w3.org/RDF/#projects
>
>aha... here it is... 10 Aug 2000
><http://lists.w3.org/Archives/Public/www-rdf-interest/2000Aug/0043.html>ht
tp://lists.w3.org/Archives/Public/www-rdf-interest/2000Aug/0043.html
>so...
>
>@prefix p3p:
<<http://www.w3.org/2000/07/p3pmodel/p3prdfschema#>http://www.w3.org/2000/07
/p3pmodel/p3prdfschema#>
>
> [ p3p:name [ p3p:nickname "Dan"; p3p:family "Connolly" ];
>   p3p:email <<mailto:connolly@w3.org>mailto:connolly@w3.org>;
>   p3p:uri
<<http://www.w3.org/People/Connolly/>http://www.w3.org/People/Connolly/>.
> ].
>
>ugh... maybe it's better to just define foaf:name in
>terms of p3p:name:
>
>
> { { :who foaf:name :n } log:implies
>   { :who p3p:name [ p3p:formatted :n ] } } log:forAll :who, :n.
>
>
>Hmmm... to call me the author when the thing is a list
>of pointers seems generous; what's the guy who
>builds a bibliography usually called? editor?
>

Sometimes it doesn't pay to put too fine a point on it.  dc:creator covers
both.  Use it. 

It is far more useful to associate your relationship to this document with
dc:creator than to distinguish it from 'author.'

And, for names of people, shadow X.500 schema or, if you find another standard
you consider better for this purpose I would be interested to hear about it. 
Where I say 'shadow,' in particular I mean make it clear that you respect the
seniority of the ISO documents and that the RDF gloss is a derived work.

If we could come up with an _authorized_ recapitulation of ISO/IEC 11179 in
RDF
we would then have a fit standard foundation for formally writing such
transcriptions (e.g. X.500 person in RDF).

See also 

<http://www-diglib.stanford.edu/cgi-bin/get/SIDL-WP-1999-0126>

For an example of using RDF to build schema reconciliation maps.

Al

>
>-- 
>Dan Connolly, W3C <http://www.w3.org/>http://www.w3.org/People/Connolly/
>  

Received on Wednesday, 14 March 2001 10:47:27 UTC