[
Date Prev][
Date Next][
Thread Prev][
Thread Next][
Date Index][
Thread Index]
comments on draft of 10 Sept
Dan, my hat is off to you. Job well done.
Page 13, 3.2.4 canonical encoding: As I mentioned yesterday, the first
sentence should refer to floating-point datum because it does apply to
NaNs.
Page 14, 3.2.25 normal number: Is finite redundant, non-zero numbers are
a subset of finite numbers according to your hierarchy. BTW, does this
suggest your hierarchy might be included somewhere in the document?
Page 14, 3.2.28 quantum: is quantum defined for zero and infinity, or
does it apply to just non-zero numbers?
Page 18, Table 1: should "an algebraically completed system" be "a
closed system". The issue is what does this mean. I assume it means that
FP operations applied to "data" produce a "datum". In mathematics that
property is called "closure".
http://en.wikipedia.org/wiki/Closure_%28mathematics%29
Page 20, paragraph before Table 3: in bullet a), do we want to say here
that "v" is undefined (for NaNs).
Page 22, last paragraph of c): can we say something stronger here, e.g.,
"A significand is canonical precisely when its declets are canonical."
That seems to be implied by the glossary item 3.2.4.
Page 35, 7.7.2: Lots of references to the undefined "x", I assume it
should be "source".
Page 38, 7.8: Has Jim responded yet to the parenthesized comment?