Re: Floating-point formats and standards
Aubrey Jaffer 21 Jan 2005 05:13 UTC
| Date: Mon, 17 Jan 2005 04:36:07 -0800
| From: "Bradd W. Szonye" <xxxxxx@szonye.com>
|
| Aubrey Jaffer wrote:
| > SRFI-47 is about fixed-size number representations, not about type
| > declaration. Here are the defaulting rules from SRFI-47's replacement
| > (submitted today) ....
|
| These rules look reasonable. I'm guessing that SRFI 58 will also use
| them; is that correct?
Yes.
I am adding this section to the Rationale:
Homogeneous Numeric Arrays
SRFI-63 introduces homogeneous arrays. This SRFI introduces
denotations for homogeneous arrays following the rank and/or
dimensions; and separated from them by a colon (:).
The names of the array-prototype-procedures of SRFI-63 and aliases
for the array-prototype-procedures of SRFI-47 match these type
denotations with A prepended. Having the array-prototype-procedure
names match the array type denotations reduces the memory load for
users.
All implementations must accept all the type denotations, but need
not implement all the homogeneous array types. Those homogeneous
types which an implementation doesn't support it stores in next best
homogeneous array type, defaulting to heterogeneous arrays in the
worst case. The rules for casting unsupported types are given in
SRFI-63.