W3C home > Mailing lists > Public > public-rdfa-wg@w3.org > November 2011

Re: Maybe a new pandora's box? '/' in a term (not CURIE!)

From: Ivan Herman <ivan@w3.org>
Date: Thu, 17 Nov 2011 11:52:09 +0100
Cc: Gregg Kellogg <gregg@kellogg-assoc.com>, Shane McCarron <shane@aptest.com>, "public-rdfa-wg@w3.org" <public-rdfa-wg@w3.org>
Message-Id: <1FDBAD6B-ACDC-459F-BCC1-966D1F40D4D0@w3.org>
To: Niklas Lindström <lindstream@gmail.com>

On Nov 17, 2011, at 11:42 , Niklas Lindström wrote:

> Hi,
> 
> On Wed, Nov 16, 2011 at 4:13 PM, Gregg Kellogg <gregg@kellogg-assoc.com> wrote:
>> On Nov 16, 2011, at 5:55 AM, "Ivan Herman" <ivan@w3.org> wrote:
>> 
>>> Unless somebody vehemently opposes this, I would like to record it as an issue. May I?
>> 
>> +1
>> 
>> Clearly required for the schema.org extension mechanism. Re-defining Term should do it.
> 
> I haven't looked into the details, but couldn't relaxing TERM from
> being an NCName get us into trouble? We'd have to redefine TERM very
> carefully, so that CURIEs aren't also considered as terms. Since
> colons are allowed in IRI segments, we'd have to determine whether
> they'd be allowed at all in TERMs, and if so, only after an NCName
> separated by any *other* gen-delim (as defined in the IRI spec)..
> 

I would be very conservative indeed for this. My proposal would be to stick to ncname _plus_ the possibility of a '/' character. For the reasons you say I would _not_ be in favour of allowing the ':' character, for example.

In other words, the only question is whether the '/' would get us into trouble. And I do not see how it would.

Ivan


> Another way to solve it could be to replace the use of
> TERMorCURIEorAbsIRI with TERMorCURIEorIRI. I.e. not relax the lexical
> space of terms, but rather allow for any IRIs including relative ones,
> and define that these are appended to the current vocab. (Note
> *appended*, not to be confused with *resolving* relative IRIs against
> a base IRI.)
> 
> (Not sure what I'd vote for yet, I just wanted to remark on some specifics.)
> 
> Best regards,
> Niklas
> 
> 
>>> Ivan
>>> 
>>> On Nov 16, 2011, at 14:47 , Shane McCarron wrote:
>>> 
>>>> In my opinion, TERM can be extended to permit anything we want.  TERMs are tokens that are matched in a very specific way.  But you are correct that, today, a TERM cannot contain a slash (nor a question mark, nor a hash).
>>>> 
>>>> On 11/16/2011 6:58 AM, Ivan Herman wrote:
>>>>> As I said, I made a mistake, I refer to the term and not the CURIE...
>>>>> 
>>>>> 
>>>>> On Nov 16, 2011, at 13:42 , Toby Inkster wrote:
>>>>> 
>>>>>> On Wed, 16 Nov 2011 09:33:54 +0100
>>>>>> Ivan Herman<ivan@w3.org>  wrote:
>>>>>> 
>>>>>>> <div vocab="http://schema.org/" typeof="A/B">  ...</div>
>>>>>> That should work. There's an example along those lines in section 7.4
>>>>>> of RDFa Core:
>>>>>> 
>>>>>>    <div prefix="db: http://dbpedia.org/">
>>>>>>    <div about="db:resource/Albert_Einstein">
>>>>>>    ...
>>>>>>    </div>
>>>>>>    </div>
>>>>>> 
>>>>> This is an example for a CURIE. What I am looking for is a term usage, ie, together with @vocab
>>>>> 
>>>>> Ivan
>>>>> 
>>>>> 
>>>>>> The suffix part of a CURIE is an irelative-ref as defined by IRI. Not
>>>>>> only can it contain a slash, it can contain a hash, a question mark,
>>>>>> and pretty much anything else you like except unescaped whitespace.
>>>>>> 
>>>>>>    <div prefix="http: http:">
>>>>>>    <div about="http://tobyinkster.co.uk/#i">
>>>>>>    ...
>>>>>>    </div>
>>>>>>    </div>
>>>>>> 
>>>>>> In that example, the CURIE is perfectly legal. The @prefix attribute
>>>>>> that defines it may or may not be legal - not quite sure. The part
>>>>>> after the whitespace is supposed to be an xs:anyURI, and I'm not 100%
>>>>>> sure whether "http:" is sufficient to satisfy that requirement, but it
>>>>>> probably is, because xs:anyURI is defined very broadly.
>>>>>> 
>>>>>> We may actually want to drop the requirement for the part after the
>>>>>> whitespace in @prefix to be an xs:anyURI. (The only important thing
>>>>>> should be that when it's combined with a suffix, it forms a valid IRI.)
>>>>>> We should perhaps allow any string that does not contain whitespace
>>>>>> there.
>>>>>> 
>>>>>> --
>>>>>> Toby A Inkster
>>>>>> <mailto:mail@tobyinkster.co.uk>
>>>>>> <http://tobyinkster.co.uk>
>>>>>> 
>>>>> 
>>>>> ----
>>>>> Ivan Herman, W3C Semantic Web Activity Lead
>>>>> Home: http://www.w3.org/People/Ivan/
>>>>> mobile: +31-641044153
>>>>> FOAF: http://www.ivan-herman.net/foaf.rdf
>>>>> 
>>>>> 
>>>>> 
>>>>> 
>>>>> 
>>>>> 
>>>> 
>>>> --
>>>> Shane McCarron
>>>> Managing Director, Applied Testing and Technology, Inc.
>>>> +1 763 786 8160 x120
>>>> 
>>>> 
>>> 
>>> 
>>> ----
>>> Ivan Herman, W3C Semantic Web Activity Lead
>>> Home: http://www.w3.org/People/Ivan/
>>> mobile: +31-641044153
>>> FOAF: http://www.ivan-herman.net/foaf.rdf
>>> 
>>> 
>>> 
>>> 
>>> 
>>> 
>> 
>> 


----
Ivan Herman, W3C Semantic Web Activity Lead
Home: http://www.w3.org/People/Ivan/
mobile: +31-641044153
FOAF: http://www.ivan-herman.net/foaf.rdf
Received on Thursday, 17 November 2011 10:49:42 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Friday, 27 April 2012 04:55:18 GMT