On Thursday, 95/11/30, Bob Quinlan wrote to Murray Lesser about "Rexx
Compilers For Os/2" as follows:

BQ> ML>     AFAIK, there are no REXX compilers for OS/2, or for anything else
BQ> ML> other than a mainframe.  Even that one is not a real compiler, it is
BQ> ML> more like a p-code builder.
BQ> 
BQ> Interesting.  I haven't looked into it in any detail, but I'm
BQ> surprised there aren't any real compilers available.  Most languages
BQ> that were designed to be interpretted still have compilers these
BQ> days.  

Hi Bob,

I see that Murray has given you the scoop on REXX compilers, or their
absence.

The reason nobody has written a real REXX compiler is because of the
INTERPRET statement. One of my current workmates "compiled" some REXX
execs on a mainframe running VM/CMS. Because of the INTERPRET
statement, the "compiled" versions ran slower than the interpreted
versions because VM had to load the REXX interpreter afresh for each
such statement; when being interpreted, the interpreter was already in
memory and need only be re-entered.

Other languages for which compilers are almost impossible are APL,
LISP and SNOBOL. I know of no true compilers for these languages
either.

However, if you want a "compiler" that runs the tokenized code there
is a very simple freeware one written by Bernd Schemmer. It can be
downloaded from CIS in OS2DF1, from Hobbes, and I think Pete Norloff's
BBS might have a copy too.

There is (or soon will be) a commercial compiler (or perhaps
"tokenizer", I'm not sure) available from Markus Pelt-Layman.

  Markus Pelt-Layman            !      Tollfree voice: 1(800) 741-4322
                                !  or outside the USA: 1(303) 442-7700
     PELT INDUSTRIES            !                 FAX: 1(303) 442-3198
        Software Experts        !       Rexx-R-Us BBS: 1(303) 440-1351
                                !         8027 N. 41st Street         
  Email: PELTLAYMAN{at}DELPHI.COM  !       Longmont, CO  80503-8843      

I have not tried it, so I cannot testify how good or bad it is. I
don't know whether it supports the INTERPRET statement or not.

Regards

Dave


 * KWQ/2 1.2i * Karaoke: Japanese for "tone-deaf drunk with a microphone."
--- Maximus/2 3.00