W3C home > Mailing lists > Public > public-i18n-core@w3.org > January to March 2006

XML Schema datatypes comments

From: François Yergeau <francois@yergeau.com>
Date: Tue, 28 Mar 2006 12:26:44 -0800
To: public-i18n-core@w3.org
Message-id: <44299C04.1030102@yergeau.com>

Gleaned from the Mandelieu minutes.  If there are no objections I will
forward to the SChema WG by their deadline (March 31).




Section 1.3, 3rd para: "Conforming implementations of this specification
may provide...", the 'may' looks like a markup error, shouldn't it be an
RFC 2119 'MAY' ?


Section 1.3: the Note says that implementations SHOULD offer a
user-override of the choice between using XML 1.0 or 1.1 datatypes.  We
think this 'SHOULD' should be a 'MUST', due to the importance of 1.1
support for i18n.


Section 1.3: The spec should provide explicit ways for other specs to
refer to it normatively, specifying either 1.0 datatypes, 1.1 datatypes
or remaining intentionally ambiguous (1.0 or 1.1 being then determined
otherwise, perhaps, as mentionned here, from the XML version of XML
instances at hand).


Section 3.3.7, 1st para: (Editorial) The sentence ending '"15 days
ending 12 July 1995" are not.' is not exactly clear, should be clarified
by saying for example 'are not duration values.'


Sections 3.4.26 (resp. 3.4.27): this says that yearMonthDuration (resp.
dayTimeDuration) is totally ordered, but the 'ordered' fun damental
facet is said to be 'partial'.  Michael Sperberg-McQueen told us why
this was so, but this looks so much like an error (or worse: obfuscation
;-) that we think the spec should explain it away.


Section 3.4.3, 1st para: "language ... as defined by [RFC 3066]or its
successor(s) in the IETF Standards Track."  RFC 3066 is *not* on the
IETF Standards Track, it is BCP 47.  Its successor is now approved but
not yet published (no RFC number yet, known as 3066bis).  We recommend
to change the reference to BCP 47
(http://www.rfc-editor.org/rfc/bcp/bcp47.txt), which will in due time
point to the new RFC.  This would require no other change to that
section, but ensure that the enriched semantics provided by 3066bis, and
the new language subtag registry (which is already live) are taken into
account.


Section 3.3.18: Thanks a LOT for referencing RFC 3987!


-- 
François
Received on Tuesday, 28 March 2006 20:26:57 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Wednesday, 1 October 2008 10:18:50 GMT