The idea of a SRFI for easy formatting of numbers
are good.

Comments
--------

   - the name OBJECT->STRING is more "Schemy"
     (it's pretty long though)

   - I find the name of the parameter DEPTH in the documentation
     confusing. (It makes me think in two dimensional output)

   - An way to get a hard limit on the width of numbers would be
     convenient.

     > (fmt -1.3333 5 4 'hard-width)
     "-1.33"

   - The
         [[<width>] [<depth>] [<char>] [<radix>] [<plus>] [<exactness>]]
     should probably be
         [<width>] [[<depth>] [[<char>] [[<radix>] [[<plus>]
         [<exactness>]]]]]]

   - I think the exactness parameter is underspecified.
     E.g. what happens if I use 'e with an inexact number?

   - I find the specification of the (FMT <others>) a little too short

BUGS
----

1. There appears to be a problem in the case
    where a negative number is truncated:

      > (fmt -1.55555 3 2)
      "-1.54"

    My TI-83 writes "-1.56".

2.

    I don't understand this behaviour:

     > (fmt -5.0 0 #\space 10 + 'e)
     . fmt: exact number cannot have a decimal point 10 depth

     > (fmt -5 0 #\space 10 + 'e)
     . fmt: exact number cannot have a decimal point 10 depth

3.

     ; DrScheme prints exact rationals as p/q
     > (fmt 1/2 0 #\space 10 +)
     "+1/2.0000000000"

     > (fmt 1/2)
     "1/2"
     > (fmt 1/2 5)
     "  1/2"
     > (fmt 1/2 5 0)

     ; The period should be there when depth is 0
     " 1/2."
     > (fmt 1/2 5 2)
     "1/2.00"

     ; my fault forgot #\space, but strange error
     > (fmt 1/2 5 2 10)
     . fmt: bad argument (10) (null? (10))

      ; An error saying that radix should be 'b 'd 'o or 'x would
      ; be more helpful
     > (fmt 1/2 5 2 #\space 10)
     . fmt: bad argument (10) (null? (10))

--
Jens Axel Søgaard