xml-dev - RE: Official ISO 639 changes

["G. Ken Holman" <gkholman@CraneSoftwrights.com>]

At 00/04/25 15:17 -0400, François Yergeau wrote:
>Elliotte Rusty Harold wrote:
> > Has anybody noticed that XML 1.0 requires 2-letter and forbids
> > three-letter language codes? From section 2.1.2 of the XML 1.0 spec:
> > ...
> > I think XML needs another erratum here to fix this.
>
>There is one in the works, but no agreement yet.
>
>Some think than XML parsers should not validate the content of xml:lang and
>just pass it to the application.  The spec would be understood to just say
>that the semantics are from RFC 1766 (or its eventual successor with
>3-letter codes).
>
>Others think that it is important for parsers to validate xml:lang.  Adding
>3-letter codes then means a substantive change to the spec which may have to
>wait for XML 1.1 (or whatever the next version is).

The XML 1.0 "Status of this document" section refers to the errata 
document, and for this discussion I'll assume that means errata are normative.

The erratum identified as E38 changes the reference to ISO 639 from 
normative to informative, reinforcing that RFC 1766 is normative and that 
the description in section 2.12 is informative.  This leads me to conclude 
that the productions 34 to 38 (with the intent to describe the lexical 
pattern for values, and not the semantically valid values) are informative 
because the narrative supporting the productions has just become 
informative.  Therefore, I think this says that an XML parser needn't 
validate against the productions 34 to 38.

The statement *not* moved to being informative is "The values of the 
attribute are language identifiers as defined by RFC 1766", so a validating 
XML parser must check what is valid there.  Looking at 
http://www.ietf.org/rfc/rfc1766.txt?number=1766 I initially see reference 
only to 2-letter language tags ... but ... there is a note in RFC 1766 that 
states "ISO 639 defines a registration authority for additions to and 
changes in the list of languages in ISO 639." and then names the authority.

I think this note leaves open the door in RFC 1766 that the ultimate 
authority for what names are valid (lexically and semantically) is outside 
the scope of RFC 1766 but that RFC 1766 reflects the decisions of the 
registration authority.

>Opinions welcome.

It is my opinion that whatever the registration authority for ISO 639 
defines as valid lexical patterns for identifiers is the normative 
requirement for XML:  XML 1.0 normatively references RFC 1766 which 
normatively references the ISO 639 registration authority which states what 
are valid lexical patterns.  I don't see the intent in XML 1.0 ever to 
judge the semantic correctness of the values, just the lexical patterns.

If the registration authority has introduced 3-letter values, then I feel 
it is defensible that those can be considered as being checkable by a 
validating parser without needing any changes to the (as amended) 
informative components of XML 1.0.

I hope this helps the discussion.  I'm not wedded to this decision and will 
consider perspectives presented by others.

.............. Ken

p.s. François, this is my opinion only (since you invited comment) and not 
that of any committee to which I belong ... I happen to be a member of 
Uma's Canadian CAC/ISO/IEC JTC1/SC2 though I haven't been able to attend 
any meetings lately (as I recall I heard you are a member of that committee 
as well); I am also chair of Canadian CAC/ISO/IEC JTC1/SC34, former chair 
of OASIS XML Conformance and Candidate Chair of OASIS XSLT Conformance


--
G. Ken Holman                    mailto:gkholman@CraneSoftwrights.com
Crane Softwrights Ltd.             http://www.CraneSoftwrights.com/x/
Box 266, Kars, Ontario CANADA K0A-2E0   +1(613)489-0999   (Fax:-0995)
Web site: XSL/XML/DSSSL/SGML services, training, libraries, products.
Practical Transformation Using XSLT and XPath      ISBN 1-894049-04-7
Next instructor-led training:    2000-05-02,2000-05-11/12,2000-05-15,
-                                    2000-06-12,2000-06-13,2001-01-27


***************************************************************************
This is xml-dev, the mailing list for XML developers.
To unsubscribe, mailto:majordomo@xml.org&BODY=unsubscribe%20xml-dev
List archives are available at http://xml.org/archives/xml-dev/
***************************************************************************