Re: [PROPOSAL: Turtle: EDITORIAL. Concepts: Non-ED] Language Tag Case Conflict (between RDF1.1 and BCP47)

Dear Jeremy,

Thank you very much for your explanation, I agree with your statement on 
the abstract syntax.

Some observations on existing rule/query engines about the language-tag 
case sensitive/insensitive:

ARQ is insensitive on language-tag case,
CWM is sensitive on language-tag case.

EULER was sensitive on language-tag case, and following the discussion, 
Jos has made a new release to make it case insensitive on language-tag.
http://eulersharp.sourceforge.net/DONE 
[Euler-2013-04]
- making language tags case insensitive

Thanks and kind regards,
Hong




From:   Jeremy J Carroll <jjc@syapse.com>
To:     Hong Sun/AXIFX/AGFA@AGFA
Cc:     eric@w3.org, David Wood <david@3roundstones.com>, "Eric 
Prud'hommeaux" <ericw3c@gmail.com>, gavin@carothers.name, Ivan Herman 
<ivan@w3.org>, public-rdf-comments Comments <public-rdf-comments@w3.org>
Date:   04/02/2013 06:48 PM
Subject:        Re: [PROPOSAL: Turtle: EDITORIAL. Concepts: Non-ED] 
Language Tag Case Conflict (between RDF1.1 and BCP47)





On Apr 2, 2013, at 1:46 AM, Hong Sun <hong.sun@agfa.com> wrote:


2. It is better to use "text"@en-gb in RDF 1.1 


As an aside, speaking for myself, the clear philosophy in RDF 1.0 was to 
not answer the equivalent question. How an implementator chooses to 
represent any item of the RDF abstract syntax was simply not something 
that that recommendation would address. The abstract syntax is a model 
that informs the test cases that informs the implementator whether their 
external behavior is correct or not. The specs carefully avoided requiring 
any particular concrete representation of anything, for example, of the 
case of lang tags - thus, the only person who can answer your question is 
you. I think several of the active participants in that WG (the 2004 one) 
were influenced by this text:


===

6. Guidance in the use of these Imperatives

   Imperatives of the type defined in this memo must be used with care
   and sparingly.  In particular, they MUST only be used where it is
   actually required for interoperation or to limit behavior which has
   potential for causing harm (e.g., limiting retransmisssions)  For
   example, they must not be used to try to impose a particular method
   on implementors where the method is not required for
   interoperability.

===


To the extent that the offending text in RDF 1.1 violates this guidance, I 
feel some editorial attention is merited

Jeremy J Carroll
Principal Architect
Syapse, Inc.

Received on Thursday, 4 April 2013 09:57:45 UTC