This is a list of small issues I have handled that I do not feel are worth
discussion. Feel free to raise them however -- they are only here to try
to keep the "issues.txt" file from larding up with random noise. If you
disagree with my classification and think an issue here *should* be discussed,
then it should be! So bring it up and we'll move the topic over until it's
resolved.

This document currently lives at
    ftp://ftp.ai.mit.edu/people/shivers/srfi/srfi-1/small-stuff.txt


* UNZIP2 defn typo

From: Doug Currie <e@flavors.com>

1 Typo. Perhaps:
    (unzip2 '(1 one) (2 two) (3 three))
should be:
    (unzip2 '((1 one) (2 two) (3 three)))

-- Fixed.

-------------------------------------------------------------------------------
* UNZIP5 typo

From: John David Stone <stone@math.grin.edu>

There's a typo in the last line of the reference implementation of
UNZIP5 (in http://srfi.schemers.org/srfi-1/srfi-1-reference.scm):

                        (values '() '() '() '()))))
should be
                      (values '() '() '() '() '()))))

-- Fixed.

-------------------------------------------------------------------------------
* ANY typo

From: erik hilsdale <ehilsdal@cs.indiana.edu>
    There's a typo in the last line of the reference implementation of
ANY  (in http://srfi.schemers.org/srfi-1/srfi-1-reference.scm).  Here's
the fixed version:

-- Fixed.

-------------------------------------------------------------------------------
* REDUCE terminology

From: John David Stone <stone@math.grin.edu>

(2) The terminology used in the specification for REDUCEL is messed
up.  In the parlance of mathematicians, a ``right zero'' of a binary
procedure F is a value RZERO such that, for any value X that F accepts as a
first argument, (F X RZERO) is RZERO.  What SRFI-1 calls a right zero is
what is usually called a ``right identity'' -- a value RIDENTITY such that,
for any value X that F accepts as a first argument, (F X RIDENTITY) is X.
There are also left zeroes -- such that (F LZERO X) is LZERO -- and left
identities -- such that (F LIDENTITY X) is X.

        The value that REDUCER returns when its list argument is empty
should be a right identity.  The value that REDUCEL returns when its list
argument is empty should be a left identity of F if the argument order for
F is MIT-Scheme-style, with the fold-state argument on the left, or a right
identity of F if the argument order for F is Standard-ML-style, with the
fold-state argument on the right.

        (3) The specification for REDUCEL also mentions LZERO at one point, 
which I assume is a typo.

        (4) The three dots in the syntax picture at the beginning of the
specification of REDUCEL should be deleted.  (Note that the implementation
doesn't handle multiple list arguments.)

        (5) In the syntax picture at the beginning of the specification of
REDUCER, `lis' should be `list'.

-- "right zero" changed to "right identity"
-- Typos (3), (4) & (5) fixed
-- I will extend the reducer implementation to be nary as spec'd.

-------------------------------------------------------------------------------
* NTH is bad name

From: Brian Harvey <bh@CS.Berkeley.EDU>
Subject: Re: The Scheme Underground list library

Am I the only one who finds confusing the incompatibility between
zero-origin NTH and FIRST, SECOND, etc?

-- Name changed back to old LIST-REF, in keeping with other zero-origin
   <foo>-REF functions.

-------------------------------------------------------------------------------
* CIRCULAR-LIST tastelessness

Many want CIRCULAR-LIST

-- OK, OK. I spinelessly accede to your every tastless whim.

-------------------------------------------------------------------------------