W3C home > Mailing lists > Public > www-xml-schema-comments@w3.org > January to March 2000

RE: Floating point proposal from left field... [long]

From: Arnold, Curt <Curt.Arnold@hyprotech.com>
Date: Mon, 17 Jan 2000 10:22:05 -0700
Message-ID: <00E567D938B9D311ACEC00A0C9B468730C74CB@THOR>
To: "'Edward Jason Riedy'" <ejr@CS.Berkeley.EDU>
Cc: "'www-xml-schema-comments@w3.org'" <www-xml-schema-comments@w3.org>
Thanks for your comments, but I assume that the discussion will go over most
people's heads.  You do give a good demonstration of the complexity of
floating point systems.  I believe that tying XML Schema's data types to a
specific implementation creates an enormous implementation issue for
applications on platforms that do not have that specific type.  From your
comments, it appears that you would agree with that.

With the lexical representation, we have an unlimited range and unlimited
precision.  I believe that it is strongly preferrable for an application to
express a value to the best of its knowledge and let a receiving system
appropriately handle any problems fitting it into its implementation
datatype.  Since the range comparisions for bounds can be done lexically
(and I believe much faster than through conversion and comparision), there
is no need for schema datatypes to delve into this mess.

In the dec17alt, I believe that I suggested an attribute where a datatype
could hint at what implementation datatypes should be used.  This hint would
imply no constraints on the value and would not compell an application to
use that datatype, but it might address the objective of those who are
trying to bind schema datatypes to specific implementations.

I'll email you the HTML files for the Help file.
Received on Monday, 17 January 2000 12:24:40 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Sunday, 6 December 2009 18:12:46 GMT