These two sort programs complement one-another.

MSORT is extremely fast, but is limited to files it
can load into memory, and cannot specify a sort
field.  On most systems it will not be able to sort
a file larger than about 48k.  It will announce any
overflow, and preserve all files, so go ahead and 
try it.

SORT is about 4 to 5 time slower than MSORT, but can
sort on a specified field.  It has NO limits on file
size, apart from disk space available.  (You should
have room for at least two copies of the file to be
sorted).  SORT puts its temporary files on the
default (logged in) drive.

Both SORT and MSORT can write back to the original
file if desired.  MSORT will give a warning.

Both SORT and MSORT can ignore case differences in
letters, by the "-u" switch.  Sorting is based on
the ASCII collation sequence only.  Tabs are just
another character, and may confuse the results if
present.  RETAB program is available to expand or
insert tabs if needed.

MSORT gives help when run with no parameters.  SORT
gives help when run with a "?" parameter.  SORT i/o
parameters require the "<" and ">" redirection 
indicators.  MSORT filenames may be separated with
commas or spaces as desired.

MSORT can input/output across DU (drive/user) areas.
SORT uses only the logged-in user area.  SORT can
input and output from devices (default console),
using the file names RDR, PUN, LST, CON, while MSORT
can only operate on disk files.  MSORT accepts wild
cards in the input filename, and sorts ONLY the first
one found.  MSORTS default output file name is the
input filename, and a different DU may be specified.

Both SORT and MSORT use C.A.R Hoares'  quicksort
algorithm, and thus the sorts are not stable, i.e.
lines that are equal may not appear in the output in
the same order as in the input.

SORT.COM and MSORT.COM are protected by CCITCRC
checksums, and should yield a 0 value under CCITCRC.

	C.B. Falconer.  (86/2/2)