W3C home > Mailing lists > Public > public-xhtml2@w3.org > March 2009

Re: @rel syntax in RDFa (relevant to ISSUE-60 discussion), was: Using XMLNS in link/@rel

From: Dan Brickley <danbri@danbri.org>
Date: Sun, 01 Mar 2009 18:58:15 +0100
Message-ID: <49AACCB7.6030401@danbri.org>
To: Sam Ruby <rubys@intertwingly.net>
CC: Karl Dubost <karl+w3c@la-grange.net>, Henri Sivonen <hsivonen@iki.fi>, Ben Adida <ben@adida.net>, Julian Reschke <julian.reschke@gmx.de>, Mark Nottingham <mnot@mnot.net>, HTMLWG WG <public-html@w3.org>, RDFa mailing list <public-rdf-in-xhtml-tf@w3.org>, public-xhtml2@w3.org, "www-tag@w3.org WG" <www-tag@w3.org>
On 1/3/09 18:29, Sam Ruby wrote:
> Karl Dubost wrote:
>>
>> Le 1 mars 2009 à 09:54, Henri Sivonen a écrit :
>>> Now, to put an actual technical proposal in here:
>>>
>>> I suggest changing RDFa to use full IRIs instead of CURIEs. Then,
>>> suggest making it a conformance requirement for rel in both text/html
>>> and application/xhtml+xml that a rel token MUST NOT contain a colon
>>> or MUST be an absolute IRI and MUST NOT start with the string
>>> "http://www.iana.org/assignments/relation/". Authors SHOULD NOT mint
>>> relation IRIs that differ only in case.
>>
>> After discussing with henri this [morning (EST) on IRC][1], a possible
>> solution for solving the issue without creating too much hurdles for
>> authors and spec would be to use urn.
>>
>> urn:dc:title
>>
>> is a URI and as henri mentioned, "one could register a URI scheme for
>> dc".
>
> If a URI scheme for dc were registered, HTML5 could treat 'xmlns:dc'
> attributes as talismans, and treat as conformance errors cases where
> such attributes have a value that differs from
> 'http://purl.org/dc/elements/1.1/'.

I've toyed with registering a FOAF uri scheme, such that foaf:title is a 
foaf: URI scheme URI for the thing normally known by the URI 
http://xmlns.com/foaf/0.1/title

...however I don't like the dynamic this creates: everyone who creates 
an RDF vocabulary will now be under pressure to go register their 
favoured short name as a URI scheme too, so that someone else couldn't 
take that name later and subvert the intended meaning of instance data 
using the original schema.

If you make a vocab known as "sam" and define it in RDFS/OWL at 
http://intertwingly.net/same-vocab-1# ...

... then next year, someone registers sam: as an abbreviation for a 
different RDFS/OWL vocab defined at 
http://sam.evilorannoying.example.com/some-other-sam# ...

Suddenly the proper interpretation is very much up for grabs.

Does <same:userid> mean what http://intertwingly.net/same-vocab-1#userid 
says, or what 
https://sam.evilorannoying.example.com/some-other-sam#userid says? 
Since machine-readable schemas associated with these URIs can be used 
for data generation and merging, this creates plenty of scope for mischief.

Should I (try to) register a foaf: URI scheme before someone else does?

cheers,

Dan
Received on Sunday, 1 March 2009 17:59:05 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 23 February 2010 18:12:53 GMT