W3C home > Mailing lists > Public > public-xml-core-wg@w3.org > March 2008

Re: Transition Request: (2nd) PER Request for XML Base Second Edition

From: Chris Lilley <chris@w3.org>
Date: Mon, 17 Mar 2008 14:57:20 +0100
Message-ID: <34841609.20080317145720@w3.org>
To: Martin Duerst <duerst@it.aoyama.ac.jp>
Cc: "Grosso, Paul" <pgrosso@ptc.com>, timbl@w3.org, steve@w3.org, <public-xml-core-wg@w3.org>, <webreq@w3.org>, <chairs@w3.org>, <w3t-comm@w3.org>, <michelsu@microsoft.com>, John Cowan <cowan@ccil.org>

On Monday, March 17, 2008, 5:48:46 AM, Martin wrote:

MD> Hello Chris, others,

MD> At 23:05 08/03/13, John Cowan wrote:
>>Disclaimer:  I don't speak *for* the Core WG, just as a member of it.

>>Chris Lilley scripsit:

>>> What is the timescape for RFC 3987bis? I see its at draft 2
>>> http://www.ietf.org/internet-drafts/draft-duerst-iri-bis-02.txt
>>> published 18 Dec 2007. When is it expected to become an RFC? I assume
>>> (given the wording above) that this is not expected to happen before
>>> 30 June, but how long after that?

>>RFCs are subject to an unpredictable delay between the time they are
>>approved and the time they are published with an RFC number, typically
>>months.  Nothing can be done about this.

MD> That delay has become a bit shorter recently. There are other delays:
MD> - Time needed for discussion and authoring

That was more the area I was asking about. Publication delays are not really the issue here; I was more asking how stable draft-02 was and how close it was to consensus.

MD> - Time needed for review and testing (including IETF last call)
MD> (help with all the above very much appreciated!)

MD> In addition, there are some inponderables with respect to the relationship
MD> between IDNs and IRIs. The IETF is most probably going to form a new IDN
MD> WG. This WG may finish very soon, or not so soon. It affects IRIs at least
MD> with respect to bidi, where there are some obivous fixes needed. If the
MD> current approach is followed, the IRI spec also may need a bit more wording
MD> on what exactly it accepts as a legal IDN and what not.

Hmm. On the one hand its good to get everything tied down; on the other hand, its good to standardize the stuff that works now and add to it incrementally. 

The current situation is a bit unstable. I don't want W3C specs to regress to allowing only URIs. But that does mean we need an IRI spec to refer to. Putting the creation work, and successful completion of another WG into the critical path for IRI-sing W3C specs does not sound like an attractive prospect.

If bidi in DNS is an issue, I would be much happier with a spec now (that can be referred to) which says bidi IDNS is an issue, and a later revision that says it is no longer an issue.

>>> Should specifications that currently use IRI in XML specify LEIRI instead?

>>No.  The "L" in "LEIRI" stands for "Legacy"; the LEIRI definition
>>simply collects under one roof the existing anything-goes definitions
>>of URI-like objects in older specs (XML 1.0 system identifiers, XLinks,
>>XML Schema anyURIs, etc.)

MD> Very much NO indeed.

MD> The following two conditions both have to be met before you should consider
MD> using LEIRIs:

MD> 1) Your spec currently uses a circumscriptive definition of (LE)IRIs such
MD>    as the one used for system identifiers in the current XML spec
MD>    (http://www.w3.org/TR/2006/REC-xml-20060816/#sec-external-ent)
MD>    rather than using a reference to RFC 3987. So there is no need for
MD>    LEIRIs in new specs.

MD> 2) Despite the fact that the chance for (LE)IRIs with such characters
MD>    actually existing out there (except for testing or to prove a point)
MD>    is extremely low, you can absolutely, positively not risk the chance
MD>    to exclude them with an erratum or technical fix to your spec.

MD> Also, please note that even if your spec should allow LEIRIs, the definition
MD> of LEIRIs explicitly says that LEIRIs that are not IRIs should not be used.


MD> Most of the points above are summarized in the opening part of
MD> Section 7 of
MD> http://www.ietf.org/internet-drafts/draft-duerst-iri-bis-02.txt:


MD>    For historic reasons, some formats have allowed variants of IRIs that
MD>    are somewhat less restricted in syntax.  This section provides a
MD>    definition and a name (Legacy Extended IRI) for these variants for
MD>    easy reference.  These variants have to be used with care; they
MD>    require further processing before being fully interchangeable as
MD>    IRIs.  New protocols and formats SHOULD NOT use Legacy Extended IRIs.
MD>    Even where Legacy Extended IRIs are allowed, only IRIs fully
MD>    conforming to the syntax definition in Section 2.2 SHOULD be created,
MD>    generated, and used.


Uh, OK, veryt clear. So IRI is the preferred term for all uses except specialised ones such as a system identifier. 

How come XML Base needs one?

MD> Regards,   Martin.


MD> #-#-#  Martin J. Du"rst, Assoc. Professor, Aoyama Gakuin University
MD> #-#-#  http://www.sw.it.aoyama.ac.jp      
MD> mailto:duerst@it.aoyama.ac.jp     





-- 
 Chris Lilley                    mailto:chris@w3.org
 Interaction Domain Leader
 W3C Graphics Activity Lead
 Co-Chair, W3C Hypertext CG
Received on Monday, 17 March 2008 13:57:58 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 17 March 2008 13:57:59 GMT