W3C home > Mailing lists > Public > public-lod@w3.org > April 2010

Re: Using predicates which have no ontology?

From: Phil Archer <phil@philarcher.org>
Date: Tue, 06 Apr 2010 16:22:16 +0100
Message-ID: <4BBB51A8.7030801@philarcher.org>
To: Michael Hausenblas <michael.hausenblas@deri.org>
CC: Niklas Lindström <lindstream@gmail.com>, Kingsley Idehen <kidehen@openlinksw.com>, nathan@webr3.org, Danny Ayers <danny.ayers@gmail.com>, Linked Data community <public-lod@w3.org>
Hi all,

Thanks for keeping me in this loop and apologies for radio silence thus far.

On a theoretical level - making the link registry available as data is, 
clearly, a jolly good idea and should happen.

On a practical level I am sorry to say I don't think I can help. In the 
e-mail that Michael sent to bring me in to this discussion he said that 
I was an editor of the Atom registry. Sorry, no, I'm not.

The ATOM Link registry is under the control of the IESG [1]. To get 
'describedby' in there I had to send an e-mail to IANA [2].

But... it's all meant to be temporary. Version 09 of Mark Nottingham's 
HTTP Link header Internet Draft has just been published and, if, as 
we've been hoping for longer than I can remember, it becomes a full RFC 
then the ATOM Link registry will be replaced by a new registry [3].

The current XML version of the registry has a bunch of declarations that 
suggest that IANA is open to making different versions available if they 
can be automated. An XSLT that produced triples would be pretty simple I 
guess (linked GRDDL-style?)

The informal place to raise issues around MNot's draft is the HTTP WG's 
mailing list (see announcement at [4]). Mark may be open to persuasion 
on seeking a data version of the registry. Alternatively one could write 
directly to IANA.

Sorry I can't be of more direct practical help.


[1] http://www.ietf.org/iesg/
[2] http://lists.w3.org/Archives/Public/public-powderwg/2009Feb/0007.html
[3] http://tools.ietf.org/html/draft-nottingham-http-link-header-09
[4] http://lists.w3.org/Archives/Public/ietf-http-wg/2010AprJun/0014.html

Niklas Lindström wrote:
> Kingsley,
> 2010/4/6 Kingsley Idehen <kidehen@openlinksw.com>:
>> Niklas Lindström wrote:
>>>> Niklas,
>>>> Nice!
>>>> I would once again suggest adding local "owl:equivalentProperty"
>>>> assertions
>>>> which enables a reasoner to treat the IANA URIs as synonyms. This is in
>>>> line
>>>> with what I like to call the: owl:shameAs pattern :-)
>>>> Kingsley
>>> Hi Kingsley,
>>> thanks!
>>> Yes, I think that'd be good. But my sketch already describes the IANA
>>> URI:s directly (by, unsolicitedly, using
>>> @xml:base="http://www.iana.org/assignments/relation/"), so *if* that
>>> RDF (or preferably Michael's richer and RDFa-based one) were official,
>>> we wouldn't need that, right? (As those would be self-referential
>>> statements..)
>>> Otherwise, if we were to mint our own ("community official") URI:s for
>>> each of these properties, I'd agree that owl:equivalentProperty should
>>> definitely be there..
>>> .. Well, unless it would be decided in the future that values in
>>> @rel:s at least in Atom are to be viewed as *indirect* references to
>>> relations via a document (akin to e.g. foaf:interest). Of course,
>>> that's not the case in XHTML+RDFa, but for the default names in @rel:s
>>> there the IANA URI:s aren't used (we have the
>>> <http://www.w3.org/1999/xhtml/vocab#>-based ones instead).
>>> So to nail down the definitions of (the nature of) the things the IANA
>>> relation URI:s identify, we'd either have to make it clear that they
>>> *are* relations (i.e. properties) in the RDF sense (and
>>> object-properties in the OWL sense), or that they're not. If it's
>>> undefined, we still can't really make any statements about what they
>>> are, even if we make up our own properties based on how we view them.
>>> (Well maybe, if it was declared that their precise meaning will be
>>> "perpetually undefined".)
>>> So if they (the URI:s) are (direct references to relations), it'd be
>>> wonderful to have IANA publish some kind of RDF discoverable via [1]
>>> to make that clear.
>> Thing is that we need RDF data representation now, and if we put the linked
>> data somewhere (some data space) ASAP we can point to what will someday
>> exist in an IANA data space -- the "shameAs" pattern is a productive
>> mechanism for letting folks like IANA understand why this is so important
>> etc. :-)
> absolutely. But do you think we should describe and use the IANA URI:s
> directly as properties, or that we need to mint new URI:s for them?
> The location of the document(s) containing these descriptions may very
> well be unreachable from iana.org for now (albeit less than ideal),
> but if we need to mint new ones, we cannot really say the iana.org
> ones are properties, right*? Since if they are, we should just use
> them..
>> Got to be fast :-)
> True. And durable. ;)
> Best regards,
> Niklas
> [*] =  Excluding owl:equivalentProperty as well since it's range is
> rdf:Property (via rdfs:subPropertyOf).
>>> [1]: http://www.iana.org/assignments/relation/*
>> --
>> Regards,
>> Kingsley Idehen       President & CEO OpenLink Software     Web:
>> http://www.openlinksw.com
>> Weblog: http://www.openlinksw.com/blog/~kidehen
>> Twitter/Identi.ca: kidehen


Phil Archer
+44 (0)1473 434770

i-sieve Technologies                   |      W3C
Sentiment Analysis Beyond Impressions  |      Open Media Web
http://i-sieve.com                     |      http://www.w3.org
Received on Tuesday, 6 April 2010 15:22:57 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 7 January 2015 15:16:04 UTC