[r6rs-discuss] [Formal] Rename bitwise-arithmetic-shift to bitwise-shift etc.
On Apr 17, 2007, at 3:56 PM, Ben Harris wrote:
> ---
> This message is a formal comment which was submitted to formal-
> comment at r6rs.org, following the requirements described at: http://
> www.r6rs.org/process.html
> ---
> submitter: Ben Harris <bjh21 at bjh21.me.uk>
> type: Simplification
> priority: Trivial
> component: Arithmetic
> version: 5.92
> summary: Rename bitwise-arithmetic-shift to bitwise-shift etc.
>
> The distinction between "logical" and "arithmetic" shift operations is
> only relevant when working with a fixed word-length and both signed
> and unsigned numbers. In those cases, the choice of shift operation
> determines how bits inserted at the high-order end of the word are
> chosen, with "logical" shifts being useful for unsigned values, and
> "arithmetic" shifts for signed ones. Scheme numbers, as represented
> in (r6rs arithmetic bitwise) are of indefinite length and are always
> signed, so a bitwise "logical" shift operation on them would be
> meaningless on two counts.
>
> In consequence, I think that the following procedures should be
> renamed to remove the word "arithmetic":
>
> Page From To
> 42 bitwise-arithmetic-shift bitwise-shift
> 42 bitwise-arithmetic-shift-left bitwise-shift-left
> 42 bitwise-arithmetic-shift-right bitwise-shift-right
> 38 fxarithmetic-shift fxshift
> 38 fxarithmetic-shift-left fxshift-left
> 38 fxarithmetic-shift-right fxshift-right
I see how your first paragraph provides evidence for your conclusion
with respect to page 42, but page 38 is part of the specification for
the (r6rs arithmetic fx) library, where the numbers are specified to
have definite length, and therefore a hypothetical fxlogical-shift
operation is not meaningless.
Do you have a separate justification for your suggested changes to
page 38?
Received on Fri Apr 20 2007 - 16:03:05 UTC
This archive was generated by hypermail 2.3.0
: Wed Oct 23 2024 - 09:15:01 UTC