[r6rs-discuss] Hashes in the number syntax

From: Ben Goetter <goetter>
Date: Mon Nov 20 11:33:53 2006

From: William D Clinger <will_at_ccs.neu.edu>
> [Jens:]
>> I tried searching for this pattern in the PLT Source base
>> including all PLaneT packages:
>
> The notation is mainly for use by engineers and scientists
> who want to express the number of significant digits more
> precisely, without resorting to scientific notation. The
> PLT source base contains a lot of code, but the MIT Scheme
> source base probably contains more code written by engineers
> and scientists.

I searched an additional ~140Mb of Scheme source. My search did not
include PLaNeT, but did include scmutils and MIT Scheme, both of MIT. I
also searched /Structure and Interpretation of Classical Mechanics/, a
MIT engineering text that uses Scheme. Nowhere did I find hash-number
notation.

The only places where I found hashes following decimal digits were
in data meant to be read by SRFI-38's read-ss, denoting shared
structure, and in unportable assumptions that hash is an acceptable
character within identifiers.

I believe that nobody uses this notation, and that it should be dropped
from R6.
Received on Mon Nov 20 2006 - 11:32:56 UTC