W3C home > Mailing lists > Public > w3c-rdfcore-wg@w3.org > November 2002

RE: freenet URIs and URI ownership

From: Dan Connolly <connolly@w3.org>
Date: 21 Nov 2002 23:47:55 -0600
To: Jeremy Carroll <jjc@hplb.hpl.hp.com>
Cc: Frank Manola <fmanola@mitre.org>, w3c-rdfcore-wg@w3.org
Message-Id: <1037944076.28129.13.camel@dirk>

On Thu, 2002-11-21 at 12:55, Jeremy Carroll wrote:
> > > > As far as I understand freenet URIs have an owner, who often wishes to
> > > > remain anonymous. (Hence the use of freenet).
> > > >
> > > > Thus, putting too much weight on URIs having an authoritative owner
> >
> > Where do we put any weight on that?
> >

Thanks for the pointer and excerpt; this is about wording
details, after all, I believe...

> Sorry:
> (that's a funny fragID)
>  http://www.w3.org/TR/2002/WD-rdf-concepts-20021108/#section-RDFDifference
> also third para of
> http://www.w3.org/TR/2002/WD-rdf-concepts-20021108/#section-Interaction
> Here is the current text:
> [[
> The social conventions surrounding use of RDF include the idea that each URI
> 'belongs to' somebody who has authority and responsibility for defining its
> meaning.

That's pretty good, except we should perhaps s/each URI/most URIs/.

i.e. it is often the case that the chain thru the URI, HTTP, and DNS
specs gets you pretty clearly to an authoritative owner.

But it's not, strictly speaking, the case that

	For all URIs u, there exists socially-responsible party p
	such that owns(p, u).

If I were going to be precise, I'd say that

	For all URIs u, there is some global contract
	(that you can find by taking the scheme part of u
	and looking it up in the IANA registry, and following
	your nose thru references from there) that
	says how the name u takes on a meaning that the
	global community agrees is authoritative.

This more precise defintion applies to news:comp.lang.python;
if you look up news:, you'll find references to the
USENET news protocols and (I think) eventually to
the USENET policies for creating newsgroups by
proposal/discussion/vote etc. So news:comp.lang.python
means what the community agreed, thru that process,
that it means.

But I think just changing "each" to "most" sufficies...
let's see...

> The social conventions are rooted in the URI specification
> [RFC2396] and registration procedures [RFC2717]. A URI scheme registration
> refers to a specification of the detailed syntax and interpretation for that
> scheme, from which the defining authority for a given URI may be deduced.

well, there again, the "... from which" part isn't universally true.

can't think of a wording fix just now, but I think I've made the point

> In
> the case of http: URIs, the defining specification is the HTTP protocol
> specification [RFC2616], which obtains a resource representation from the
> host named in the URI; thus, the owner of the host's DNS domain controls
> (observable aspects of) the URI's meaning.

That part is fine.

> ]]
> and
> [[
> RDF assumes that for any URI some individual or organization has the
> authority to define the meaning of that URI.

bzzt. RDF doesn't really assume that or rely on that. Practical
use of it does rely on *some* grounding of URIs to social context; but
the social context doesn't have to be an owner; it can be
a historical context like a USENET newsgroup vote.

> An RDF predicate is defined by
> the individual or organization with such the authority with respect to the
> its URI, and misuse by others should not be permitted to undermine that
> authority.

That's true (and worth saying) in the case of HTTP and DNS,
but the general case is more subtle.

> ]]
> while the text could do with polishing, I think the idea is, at least in
> part, needed, for our resolution of the rdfms-assertion issue.


> http://www.w3.org/2000/03/rdf-tracking/#rdfms-assertion
> [[
> Resolution: On 23rd August 2002, the RDFCore WG resolved:
> that the text in section 2.3.2 of the Concepts and Abstract Data Model
> document resolves this issue and it be closed.
> ]]
> The text referred to in that resolution is found in:
> http://lists.w3.org/Archives/Public/www-archive/2002Aug/att-0002/02-Overview
> .htm#section-Social
> which reads
> [[
> RDF/XML documents, i.e. encodings of RDF graphs, can be used to make
> representations of claims or assertions about the 'real' world. RDF graphs
> may be asserted to be true, and such an assertion should be understood to
> carry the same social import and responsibilities as an assertion in any
> other format. A combination of social (e.g. legal) and technical machinery
> (protocols, file formats, publication frameworks) provide the contexts that
> fix the intended meanings of the vocabulary of some piece of RDF, and which
> distinguish assertions from other uses (e.g. citations, denals or
> illustrations).
> [[[This needs reviewing...]]]
> For example, a media type, application/rdf+xml [RDF-MIME-TYPE] is being
> registered for indicating the use of RDF/XML that might be published with
> the intent of being such an assertional representation (as distinguished
> from other XML or text that may just happen to look like RDF assertions).
> ]]
> ++++++++++
> The key problem here is that the issue resolution really only talks about a
> single RDF document making an assertion.
> The editors have bravely tried to extend this issue resolution to multiple
> related independently authored RDF documents which when combined have
> slanderous entailment.
> (See particularly:
> http://www.w3.org/TR/2002/WD-rdf-concepts-20021108/#section-InteractionExamp
> le
> )
> To make that work then the idea of a URI owner is needed,

I don't think owners are critical, but rather: social context.

> and this idea
> seems, somewhat problematic.
> However, a single document is hardly a semantic web!
> And without text such as that I am questioning, we could end up with the
> situation where in the clown example none of the original authors are liable
> but someone who sucks up that part of the semantic web and spits it out as a
> single document then has a legal liability!
> Jeremy

Dan Connolly, W3C http://www.w3.org/People/Connolly/
Received on Friday, 22 November 2002 00:48:12 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 20:24:18 UTC