Precision and CD's

[Rüdiger Loos]

>I wonder if it is just a disagreement over the wording, but
>in fact I view "precision" as an attribute of a representation.

1. There is an
International Standard ISO/IEC 10967-1:1994, Information technology -
Language independent arithmetic, Part 1: Integer and floating point arithmetic.
The standard is final; a draft is available from my ftp site:
  ftp://ftp-CA.informatik.uni-tuebingen.de/documentation/LanguageIndependent-Standards/

  lia1.ps.gz                                           187 Kb    Wed Jul 21 17:58:00
1999

2. The new C-standard-draft ISO/IEC 9899 refers in Annex H to this standard as 'more
general' than IEC 60559:1989, alias ANSI/IEEE 754-1958, which is an implementation
standard.

3. Another standard one can get from ISO, Case Postale 56, CH-1211 Geneve 20,
Switzerland (not cheap):
International Standard ISO/IEC 11404:1996(E), Information technology -
Programming languages, their environments and system software interfaces -
Language independent datatypes.

The conceptual clarity of these documents is impressive and may help the
OM discussions. For example, in LiA1, section 4, 'precision' is defined as
'the number of digits in the fraction of a floating point number (s. 5.2)' and in 5.2
'precision' appears as one of 5 attributes which specify any floating point type by
explicitely stated formal requirements. In 5.2.3 the axioms of 'Approximate operations'
are stated (similar to Gonnet's remark), explicitely refering to 'precision.'
Why not separate the fp specification from the representation and implementation
issues?

Ruediger Loos

-
om at openmath.org  -  general discussion on OpenMath
Post public announcements to om-announce at openmath.org
Automatic list maintenance software at majordomo at openmath.org
Mail om-owner at openmath.org for assistance with any problems