Tony said:
>Hmm.  In 1977 (?) I sent a list of 35 enhancement suggestions to
>various people within HP.  Number 7 (verbatim) follows:
>
>Add a parameter to the ":JOB" statement to permit the following:
> :JOB . . .;INCLASS=classname[,inpri within class]
>where "classname" is an installation defined name.  The LIMIT
>command would also have to be changed.  The intent here is to
>permit a single-thread effect without blocking access to batch
>processing.  An example might be:
> LIMIT 2             overall batch limit.
> LIMIT 1,SINGLE      only one job at a time in class "SINGLE".
>a) Job A (a master file update) is introduced in class "SINGLE" and
>   enters execution.
>b) Job B (a report program that cannot run until A finishes) is
>   introduced class "SINGLE" and will wait for A to finish.
>c) Job C (a compile) is introduced and runs, even though B is
>   waiting for A to complete. (No class was specified.)
 
This really brings back memories. Deja Vu all over again!
 
Yes!  Tony's description is in fact very much like my request of years gone
by. I had forgotten what a spooler really looked like after all these years.
 
So they ignored two very similar requests made oh so many years ago.  It
reminds me of the time I spent with the HP folks while an employee (in the
late 70's also) explaining how a data dictionary should really work (after
attending the Rapid / David Dummer workshops as a product champion) and
seeing Dictionary/3000.  Nothing every came of that either.  Gee Wiz, we
give away the best years of our lives for knowt.
Brian Duncombe  [log in to unmask]