[Cfp-interest 2258] terminology issues

[Vincent Lefevre]

IIRC, this has already been mentioned, but 5.2.4.2.2p3 (and p4)
defines the term "floating-point number", which is not used in a
consistent way.

For instance, 7.12.7.3 in the latest C2x draft N2731 specifies the
fabs function, saying

  The fabs functions compute the absolute value of a floating-point
  number x.

But this should be specified for all values of the floating-point
type, including infinities (which are not regarded as FP numbers
by 5.2.4.2.2p3 and p4).

About the term "range":

  * 5.2.4.2.2p7 could be reconsidered. By "most positive finite
    floating-point number representable in that type", does this
    limit to *_NORM_MAX or should that be DBL_MAX, etc.? i.e.
    should "floating-point" be dropped? What about infinities?
    (I don't think that whether infinities are included or not
    matters in the other parts of the standard, but it would be
    better to make sure that the definition is clean.)

  * The use of "range" in the other parts of the standard should
    be checked. For instance, 5.2.4.2.2p13 ("[...] the values of
    floating constants are evaluated to a format whose range and
    precision may be greater than required by the type.", etc.)
    and similar paragraphs that say "greater range and precision"
    or "extra range and precision" use a different definition:
    here, this actually means that the values are in a super-set.

-- 
Vincent Lefèvre <vincent at vinc17.net> - Web: <https://www.vinc17.net/>
100% accessible validated (X)HTML - Blog: <https://www.vinc17.net/blog/>
Work: CR INRIA - computer arithmetic / AriC project (LIP, ENS-Lyon)