W3C home > Mailing lists > Public > public-rdf-wg@w3.org > November 2012

Re: Factoring of entailment regimes

From: Peter F. Patel-Schneider <pfpschneider@gmail.com>
Date: Sat, 17 Nov 2012 14:13:53 -0500
Message-ID: <50A7E1F1.7010606@gmail.com>
To: Richard Cyganiak <richard@cyganiak.de>
CC: RDF Working Group WG <public-rdf-wg@w3.org>
I'm not seeing that there is any need to fix up RDF-entailment. Sure, who 
would want to use it, but does that mean that it is broken?   I'm similarly 
not seeing any pressing need to have to handle datatypes without handling RDFS.

I'm certainly not seeing any need to change the treatment of "bad"^^xsd:int.  
I didn't like the treatment previously and I don't like it now, but I don't 
see that it is any worse now than then.l


On 11/17/2012 11:37 AM, Richard Cyganiak wrote:
> On 17 Nov 2012, at 16:06, Peter F. Patel-Schneider wrote:
>> I'm having trouble figuring out why the working group should do anything here, particularly at this late date.  Is it something that needs to be done because the current situation is horribly broken?
> Yes -- in particular:
> 1. There is no entailment regime that makes 1 and +1 equal without requiring all of RDFS
> 2. The decision to make rdf:XMLLiteral an ordinary datatype has gutted RDF-entailment, and its distinctionfrom RDFS-entailment has become rather spurious
>> Or is it something that needs to be done because the WG charter requires it?
> The charter asked us to clean up string literals and deal with the errata. This is tying up the loose ends of that job.
> Best,
> Richard
>> I don't see that either is the case.
>> peter
>> On 11/17/2012 09:57 AM, Richard Cyganiak wrote:
>>> On 17 Nov 2012, at 13:15, Ivan Herman wrote:
>>>> On Nov 16, 2012, at 20:06 , Pat Hayes wrote:
>>>>> We could define a whole range of entailment regimes, in fact, even ones that "overlap" others, by mixing and matching groups of rules and axioms. So maybe we can re-define the 2004 regimes as we go, as well as our newly preferred "backbone". The 2004 regimes were written to exactly match the rdf, rdfs and rdfs+datatypes namespaces, so it would not be outrageously strange to repeat these in the revized specs, if only for historical interest.
>>>> We could also use the renaming trick: we can use another term for RDF and RDFS entailments new style, keep the old ones unchanged, defined but declared 'deprecated' or something.
>>>> I am not sure it is the best approach, just and early-morning-after-breakfast thought...
>>> Perhaps best to proceed in the knowledge that we *could* still define entailment regimes corresponding to the 2004 ones if we want, and leave these question for later:
>>> - whether to really define the 2004 entailment regimes
>>> - whether they would be normative, informative, deprecated or whatever
>>> - what the names for all the entailment regimes (new and old ones) will be
>>> Best,
>>> Richard
Received on Saturday, 17 November 2012 19:14:22 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 22:02:09 UTC