W3C home > Mailing lists > Public > www-xml-schema-comments@w3.org > January to March 2006

Finiteness of xsd:gMonthDay etc

From: C. M. Sperberg-McQueen <cmsmcq@acm.org>
Date: 06 Jan 2006 07:13:45 -0700
To: Jeremy Carroll <jjc@hplb.hpl.hp.com>
Cc: W3C XML Schema Comments list <www-xml-schema-comments@w3.org>
Message-Id: <1136556824.3885.44.camel@localhost>

Jeremy,

long, long ago, you wrote the XML Schema comments list
[1] to inquire about the fact that in appendix A, the 
schema document with declarations for the built-in
datatypes gMonthDay, gDay, and gMonth gives their cardinality 
as 'countably infinite', despite the fact that there seem
only to be 366 values for the first and third, and only 
twelve for the second.  

You ask, generously, "is this an oversight?"

With regrets for the length of time it has taken us to
answer you, I can now say pretty definitely yes, it was.
I speculate that the confusion was caused by the reflection
that the number of months in the Gregorian calendar, viewed
as continuous segments of time, is not twelve but countably
infinite.  That is to say, when we wrote that gMonth was
countably infinite, we were confused by the fact that 
each of the twelve values of gMonth itself denotes a
countably infinite set of time periods.  Another capitulation
to the fallacy of composition, I'm afraid.

The WG has not officially approved a fix for the wording,
but the fix is obvious and I'm pretty sure it will be
approved in due course.  If you wish to check our work,
you might consult the proposal in [2] and let us know
whether you approve.  If we don't hear from you in the next
month or so, we'll assume you approve.

Thank you for your patience.

Michael Sperberg-McQueen

[1]
http://lists.w3.org/Archives/Public/www-xml-schema-comments/2002OctDec/0068.html

[2] http://www.w3.org/Bugs/Public/show_bug.cgi?id=2179
Received on Friday, 6 January 2006 14:14:18 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 7 January 2015 14:50:04 UTC