As Chris Blaicher and Sam Siegel have suggested, PAUSE/RELEASE is
logically complete.  Anything that can be accomplished with WAIT/POST
can be accomplished, more elegantly and with fewer barred windows and
culs de sac, with PAUSE/RELEASE.

As Chris mentioned, intertask communication can be accomplished in
many ways.  It is useful to think of it as of two kinds.

The first kind is metabolic.  A subtask may wish to communicate (to
its parent) that it has completed its work normally or abnormally
(roughly, successfully or unsuccessfully).  Or again, a parent task
may wish to tell one of its subtasks to wind up its work ASAP.  The
machinery provided addresses most of these requirements adequately,
and for this reason designers should try to think in terms of them
where they can.

In other cases more and different classes of information must be
communicated, and it is for them that more complex, often very much
more complex, machinery is needed.  In particular, queues are useful
chiefly when they sometimes/often contain multiple elements.
Degenerate queues that never contain more than a single element should
be avoided.

John Gilmore, Ashland, MA 01721 - USA

----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN

Reply via email to