[std-interval] Revised document available

[Gabriel Dos Reis]

Guillaume Melquiond <guillaume.melquiond at ens-lyon.fr> writes:

| Le lundi 18 septembre 2006 à 22:48 +0100, Dr John Pryce a écrit :
| > Dear std-intervalers,
| 
| > B.
| > I do not feel the text makes it clear whether the 
| > platform's arithmetic MUST support oo - which 
| > more or less amounts to using IEEE arithmetic.
| 
| Unless I'm mistaken, infinities are mandated by LIA-1 (Language
| Independent Arithmetic 1)

LIA-1 does NOT mandate infinities.


What LIA-1 says is  (section 5.2.9 Conformity to IEC 559)
  
  # The parameter iec_559 shall be true only when the type F
  # completely conforms to the requirements of IEC 559 (also known as
  # IEEE 754).  Fa may correspond to any of the floating point types
  # defined in IEC 559.

A further note following the above statement

  # NOTE 1 -- The IEC 559 facilities include values for infinities and
  # NaNs, extended comparisons,  program control of rounding, an
  # inexact exception flag, and so oon.  See annex C for more
  # information. 

Last and not least, you have to read section "A.5.2.0.4 Infinity and NaNs"
from Appendix A:

  # [...]  Therefore, LAI-1 makes no provision for infinity or NaNs.
  # The behavior of operations with an infinity ro a NaN as input is
  # not defined by LIA-1.  owever, be sure to read 5.2.9 and clause C.1

-- Gaby