[r6rs-discuss] The definition of flonum is in the wrong place

From: Aubrey Jaffer <agj>
Date: Sun Oct 8 20:56:13 2006

 | Date: Sun, 8 Oct 2006 01:41:40 -0400
 | From: John Cowan <cowan_at_ccil.org>
 |
 | Aubrey Jaffer scripsit:
 |
 | > The definition of flonums should be moved out of the "Rationale" on
 | > page 39:
 | >
 | > The flonums are a subset of the inexact reals, but may be a proper
 | > subset.
 | >
 | > and put on Page 100, in section "16.4. Flonums".
 |
 | +1
 |
 | > The phrase "but may be a proper subset" adds nothing to the
 | > description and should be removed.
 |
 | It is a very useful clarification, especially given that there will
 | probably be many systems in which the flonums will not be a proper
 | subset of the inexact reals.

Then the sentence should read:

  The flonums are a subset of the inexact reals, but may include all
  of the inexact reals.

 | > Near the bottom of page 39, the lines:
 | >
 | > (finite? x) procedure
 | > (infinite? x) procedure
 | > (nan? x) procedure
 | >
 | > should be changed to:
 | >
 | > (finite? z) procedure
 | > (infinite? z) procedure
 | > (nan? z) procedure
 |
 | I consider it useful to apply these procedures to non-numbers.

Should ZERO?, POSITIVE?, NEGATIVE?, ODD?, and EVEN? also accept
non-numbers?

 | > The text after this (page 40) states:
 | >
 | > ..., finite? tests if it is not an infinity and not a NaN,
 | > infinite? tests if it is an infinity, nan? tests if it is a NaN.
 | >
 | > This leaves in question whether (infinite? +nan.0). Changing the text
 | > thus eliminates the unintended interpretation:
 | >
 | > ..., finite? tests if it is not an infinity and not a NaN,
 | > infinite? tests if it is an infinity or NaN, nan? tests if it is a
 | > NaN.
 |
 | This seems to be a very unlikely interpretation. NaNs are not
 | infinities.

In mathematical parlance, "infinite" means "not finite", just as
"incomplete" means "not complete". IEEE-754 talks about "infinities"
which exclude NaNs. How about replacing INFINITE? with:

  (infinity? z) procedure

..., infinity? tests if it is +inf.0 or -inf.0, ...
Received on Sun Oct 08 2006 - 20:56:11 UTC