Re: Fw: Expression evaluation
On May 18 2011, Ian McIntosh wrote:
... 754 has been the de facto standard for a couple of decades,
and I don't know of ANY language that supports it in full, or even
supports most of it at all well. And the semantic problems are a
major part of the reason why not.
If you replace "language" with "language implementation", there are a
number of vendors who support almost all of 754-1985, at least if the right
options are used. We need to continue working on moving from "language
implementation" to "language standard and implementations".
Perhaps. If you were to say "almost all of the mandatory requirements
of 754-1985", I might accept that for at least C99. When I last checked
(which was some time ago), the support wasn't actually usable by most
programmers, for a variety of reasons, but let that pass.
My point was that you just aren't going to get it into the standards,
because it is so semantically incompatible.
The C Floating Point Study Group is well along on a detailed proposal for
adding essentially all of 754-2008 to C. Type names, literal suffixes,
function names and so on have been dealt with. We're almost finished
Signaling NaNs and have half a dozen more topics (eg, alternate exception
handling, extendible types, debugging) to go. I think it will be ready
before 1788 is. 8<) I hope C++ will adopt all or most of it too,
and other languages need similar projects.
Ah. NCEG redux.
Regards,
Nick Maclaren.