Eddy Thilleman wrote in a message to Jack Stein:

ET>>> Call loadrxx.cmd in your startup.cmd:

JS>> Why not just put the lines in the startup.cmd?

ET>> Then you have to turn the whole startup.cmd file into a REXX
ET>> file (I don't know if that's possible).

JS> Sure.  I do it that way here.

 ET> Then you start your startup.cmd file with a REXX comment?

Yes, one can't call REXX functions from a batch file, right?

JS> If you are calling a rexx script from startup.cmd, then, why 
JS> would you not be able to use the start-up to call the Rexx 
JS> Utilities?

 ET> That's what I do: I call a REXX file from my startup.cmd.

That's OK too, but not necessary if your startup.cmd is a REXX file.

 ET> Thinking about startup.cmd: startup.cmd is just a normal
 ET> batch file and REXX uses also text files with the .cmd
 ET> extension (because of the command processor), so now I see
 ET> how you can think of making startup.cmd a REXX file. :) 

Yep, but you must understand how to use regular batch commands in a REXX
script if you are also mixing in regular batch commands.  This is part of
knowing how to use REXX, so the concepts are important either way.  

JS> Hmmm, I guess you are thinking because startup loads before the
JS> WPShell loads?

 ET> No, I just hadn't thought about making my startup.cmd file
 ET> into a REXX file. 

JS> I don't get any error messages though, so I guess PMShell is running
JS> when startup starts.

 ET> At the time startup.cmd starts, the screen is already a
 ET> blank graphic screen, that's Program Manager (PM).

I remember long ago people used to say REXX didn't work w/o PMShell, but that
is wrong, it works fine w/o PMShell, but the extended utilities don't work w/o 
PMShell.  Not many people run OS/2 w/o the "windows" shell, even hard core
windows haters like me.  Too much stuff depends on PMShell to be bothered
avoiding it completly, so I learned to live with it.

                                              Jack 
--- timEd/2-B11