XML Schema datatypes: real world usage of gMonth (fwd)

For info; when we get stuck into the detail of our work on XML Schema
datatypes, we'll want to bear in mind that there is some bugfix work going
on with that spec too. See below for msg from Paul Biron regarding a
problem with the specification of gMonth.

Dan


---------- Forwarded message ----------
Date: Fri, 26 Oct 2001 15:00:48 -0700
From: "Biron,Paul V" <Paul.V.Biron@kp.org>
To: 'xmlschema-dev' <xmlschema-dev@w3.org>
Subject: real world usage of gMonth
Resent-Date: Fri, 26 Oct 2001 18:07:25 -0400 (EDT)
Resent-From: xmlschema-dev@w3.org

The WG has discovered that lexical represenation specified in the schema Rec
for the type gMonth does not match the truncted form of ISO 8601 as we had
intended.

Briefly, the lexical forms are as follows:

		Rec (erroneous)		8601 (correct)

gMonth	--MM--			--MM

The WG has tentatively resolved to publish an erratum :

	1. noting the incorrect form in the Rec (i.e., declare it as
invalid)
	2. and specifying the correct lexical forms.

However, before we institute this decision, we are seeking input on this how
would affect schemas and instances in existence already.  If we declare the
"--MM--" form invalid would that adversely affect anyone's schemas or
instances?  Please let us know.

An alternative resolution the WG has discussed is to publish an erratum:

	1. specifying the correct form
	2. deprecate the "--MM--" form (i.e., still legal for v1.0
processors)
	3. declare the "--MM--" form invalid for v1.1 and later processors

thanx for your input

pvb

Received on Friday, 26 October 2001 19:13:44 UTC