Stan does a double-take:

> It occurs to be that the right way to approach this is to make the
> SHUTDOWN command a real command, not just a control-A command.

> Then, you could create a SYSSTOP script with a SHUTDOWN
> command in it.

> This reverses the problem statement, but solves the overall goal.

....  and at first glance seems to me to offer more flexibility;  and
would perhaps make it easier to add additional "features" in the
future ??....   blue-sky example of part of a background job that I
would like to be able to run as OPERATOR.SYS someday soon:

!.......
!SETVAR HPAUTOCONT TRUE
!WHILE  <UPS has enough battery left and room temp O.K.> DO
!PAUSE <whatever seems appropriate>
!DOEND
!SETVAR TIMETOSD <# of seconds of battery remaining>
!SETVAR TIMETOSD TIMETOSD - <some safety factor>
!comment  YIKES !!...  losing power or overtemp.... shutdown !!
!WARN <the world>
!STREAM  JBGRACFL.SYSOPER (graceful SD:  stop logging, etc.)
!PAUSE !TIMETOSD; JOB="GRACFL,OPERATOR.SYS"; EXISTS
!SHUTDOWN  (Stan's above)
!EOJ

It is a bit embarrassing that the only way for MPE to respond to a
message from an UPS (HP or any other, AFAIK) is to crash the
O.S.  Since the code to actually *do* a shutdown is already there
to be executed on a =SHUTDOWN from the CONSOLE, seems
like it shouldn't be a huge task to put a regular MPE command
"header" on that same code (restricted to MANAGER and
OPERATOR.SYS and maybe ALLOW list, of course)...  We have
:RECALL & =RECALL;  why not :SHUTDOWN & =SHUTDOWN..

....  add my name to the <plug> list for putting this on the SIGMPE
or SIGSYSMAN enhancement wish list.....  maybe this is a another
good candidate for the next round of Customer Funded Enhancements
(CFE)  ??..

Ken Sletten