[R6RS] proposed generalizations of equal?
William D Clinger
will at ccs.neu.edu
Wed Jun 28 08:26:57 EDT 2006
For the record, let me clarify something I just wrote:
> > Does this imply an opinion on =, char=?, char-ci=?, string=?, and
> > string-ci=?
> Yes, it implies my opinion. All of those are total orders,
A total order consists of a set and a binary relation.
The following are total orders:
the reals ordered by <=
the characters ordered by char<=?
the equivalence classes of characters modulo char-ci=?
ordered by char-ci<=?
the strings ordered by string<=?
the equivalence classes of strings modulo string-ci=?
ordered by string-ci<=?
In the reals ordered by <=, the equality predicate is =.
Hence my general rule would have = take two or more arguments,
even though = also serves as the equality predicate on complex
numbers, which are not totally ordered.
In the characters and strings ordered by case-insensitive
operations, individual characters and strings serve as
representatives of the equivalence classes. Hence my
general rule would have char-ci=? and string-ci=? take
two or more arguments, even though the characters are
not totally ordered by char-ci<=? and the strings are
not totally ordered by string-ci<=?.
More information about the R6RS