Common Lisp standard and IEEE questions

[Barry Margolin <barmar@Think.COM>]

    Date: Mon, 23 Nov 1992 04:17 EST
    From: Guy Footring <Guy@AVON.sltd.dialnet.symbolics.com>

    Am I correct in assuming that whilst conforming CL implementations
    are permitted to conform to the IEEE proposed standard they are not
    required to?  

Correct.  It's more likely that a CL implementation will use whatever
floating point format the computer system it's running on provides.  If
CL required IEEE FP, this would force implementations on non-IEEE
hardware to implement FP in software, which is generally not what the
user expects.  I don't know of any language standard that requires a
particular FP format.

    Are there any other standards that an implementation might choose to implement instead?

Not that I know of, but there are many proprietary FP formats, such as
VAX floating point.

    How much of the CL (draft) standard relating to floating-point is actually 
    in line with (or conforms to) the IEEE proposed standard, in effect forcing implementations
    to conform?

None.  We specify the names of some conditions that correspond to IEEE
traps.  *If* an implementation provides IEEE FP, it should signal those
conditions when the traps occur.

                                                barmar