Re: RDF version of IETF RFC metadata (and relationship to FOAF files? )

>I've posted an RDF version of the IETF's listing of RFCs and associated
>metadata ( <http://www.ietf.org/iesg/1rfc_index.txt>
>at <http://www.snee.com/rdf/1rfc_index.rdf>

I did something similar a while ago; I didn't publish the output
because I didn't plan to keep it up to date, but the conversion
is

# $Id: rfcIndexGrok.pl,v 1.8 2001/08/16 22:28:39 connolly Exp $
#
# USAGE:
#
# GET http://www.ietf.org/iesg/1rfc_index.txt \
#   | perl rfcIndexGrok.pl >,rfc-index.rdf
  http://www.w3.org/2001/02pd/rfcIndexGrok.pl

and it seems to still work, ala:

connolly@dirk:~/w3ccvs/WWW/Addressing$ make rfc-index-1630on.rdf
wget ftp://ftp.isi.edu/in-notes/rfc-index.txt
20:07:52 (397.53 KB/s) - `rfc-index.txt' saved [638939]

perl ../2001/02pd/rfcIndexGrok.pl 1630 <rfc-index.txt
>rfc-index-1630on.rdf

> I think it makes great semantic web
>metadata for several reasons, among them being the typed relationships
>between RFCs (e.g. dcterms:isReplacedBy, pr:hasCorrection) and the nice
>connections to the existing semantic web: the existence of several RFC
>authors who have FOAF files. 
>
> Which brings me to my question for the list: what's the best way to model
> the latter? For example, what's the best RDF/XML way to indicate that "N.
> Walsh" in the following entry for an RFC has a FOAF file at
> <http://norman.walsh.name/foaf> http://norman.walsh.name/foaf? I don't want
> to say "the creator of  <http://www.ietf.org/rfc/rfc3120.txt>
> http://www.ietf.org/rfc/rfc3120.txt has a FOAF file at
> <http://norman.walsh.name/foaf> http://norman.walsh.name/foaf"; the resource
> has more than one creator, and I want to make it clear which one has which
> FOAF file.

It's easier to say "a creator..." than "the creator" because
(as far as I know) dc:creator is not a FunctionalProperty; i.e.

  <book1> dc:creator ppl:bob.
  <book1> dc:creator ppl:mary.

doesn't mean that bob = mary.

In your data, I see:


<rdf:Description rdf:about='http://www.ietf.org/rfc/rfc0001.txt'>
  <title>Host Software.</title>
  <creator>S. Crocker</creator>

which suggests the creator is a 10 character string starting with S.
(see http://esw.w3.org/topic/ThingsVersusTheirNames ).


The way I modelled it is:

    :_rfc3501-1
<http://purl.org/dc/elements/1.1/creator> :_MCrispin3;
         <http://purl.org/dc/elements/1.1/date> "2003-03";
         <http://purl.org/dc/elements/1.1/format> "text/plain";
         <http://purl.org/dc/elements/1.1/title> "INTERNET MESSAGE
ACCESS PROTOCOL - VERSION 4rev1";
         rfc:author :_MCrispin3;
         rfc:bytes "227640";
         rfc:date "March 2003";
         rfc:obsoletes :_rfc2060-2;
         rfc:rfc "3501";
         rfc:status "PROPOSED STANDARD" .

    :_MCrispin3     c:fullName "M.Crispin" .

which still sorta cheats re rfc:status and perhaps even the dates.

(My tool spits out RDF/XML, using a very primitive serializer,
so I ran it thru cwm's n3 pretty-printer for this message).



p.s. In your data, I see:

      xmlns:rfc2026='http://www.ietf.org/rfc/rfc2026.txt'
 ...
      <rfc2026:status>UNKNOWN</rfc2026:status>

That says that
  http://www.ietf.org/rfc/rfc2026.txtstatus
is an RDF property. The IETF owns that name, and I don't
think they'd agree. The told their web server to give a 404
for that name, while the obvious way for a URI owner to
acknowledge that something is an RDF property is to arrange
for their web server to say so.

Hmm... oops... http://www.w3.org/2001/02pd/rfc56 is
still 404 too. Better get my own house in order...

Anyway... webarch sorta makes this point. The old RDF Schema
spec used to; I hope the Best Practice WG picks it up again.

-- 
Dan Connolly, W3C http://www.w3.org/People/Connolly/
D3C2 887B 0F92 6005 C541  0875 0F91 96DE 6E52 C29E

Received on Thursday, 6 January 2005 02:29:16 UTC