- From: Sandro Hawke <sandro@w3.org>
- Date: Wed, 05 Oct 2005 22:20:40 -0400
- To: Joshua Tauberer <tauberer@for.net>
- Cc: David Powell <djpowell@djpowell.net>, "Hammond, Tony" <T.Hammond@nature.com>, "'SWIG'" <semantic-web@w3.org>
>
> Giovanni Tummarello wrote:
> > you might want to specify that the labeled graph reppresentation
> > of RDF is an approxiamtion and works just in certain cases since
> > labels ... can be subjects or objects of rdf annotations.
>
> Good point; will do.
>
> David Powell wrote:
> > You could also consider tag: URIs
> > http://www.taguri.org/07/draft-kindberg-tag-uri-07.html
> > They do use the existing domain name registry as an authority.
>
> Great. I knew of tag:, but I had wrongly assumed the authority part of
> the URI must be an email address followed by a YYYY-MM-DD date, which
> seemed a bit verbose. Now that I've read the spec, I see the only
> required thing in a tag: URI over my suggested URI format (a scheme
> equivalent to http: except with a different name) is the year, which is
> great.
>
> Thanks for pointing out tag:. I'll use that.
This is very old ground, but can I ask your reason for using tag: URIs
instead of HTTP URIs?
>From the FAQ I wrote for http://taguri.org/ :
Q: Are tags good RDF identifiers?
A: As with XML Namespaces, there are two schools of thought. Do you
want the web's default mechanism at your disposal for fetching
information about the identified thing? If so, then HTTP URIs are
probably better used.
I'm in the slightly odd position of co-authoring the tag URI spec
[which should hit RFC any day now, finally, I *think*], and yet I
think most applications of tag URIs would be better off using HTTP
URIs.
-- sandro
Received on Thursday, 6 October 2005 02:20:49 UTC