RE: float and double ordered facet value, error in 2E PER

Dave:
Here's how I see this issue:

- IEEE-754 says NaN is unordered WRT all other numbers including itself.
- XML Schema says NaN = NaN
- F&O says that the result of comparing NaN with any other value is
'false'. This is an attempt to mimic the IEEE-754 semantic without going
into partial order.  I recommend Schema do the same.

All the best, Ashok
-----Original Message-----
From: www-xml-schema-comments-request@w3.org
[mailto:www-xml-schema-comments-request@w3.org] On Behalf Of Dave
Peterson
Sent: Monday, May 03, 2004 8:31 AM
To: www-xml-schema-comments@w3.org
Subject: float and double ordered facet value, error in 2E PER


I had understood that the ordered facet value for float and double in
C.1
was to be changed from total to partial.  This correction does not show
in the undiffed version out for comment.

In point of fact, float and double are partially ordered.  This is
because
NaN compares only to itself.  E.g.,  29 <> NaN .
-- 
Dave Peterson
SGMLWorks!

davep@iit.edu