On Tue, 6 Jan 1998 02:15:18 -0800, "John D. Alleyn-Day"
<[log in to unmask]>
wrote:

>Why not just call FLOCK unconditionally?  Then you don't have to mess with
>PAUSE at all.  I've been doing this for years for incremental counters
>(usually on an IMAGE dataset) and have never had a problem with it.
>Performance has NEVER been a problem, even on a Classic machine.

Because I don't like "wait_forever_if_you_can't_lock_it" designs.  Too 
easy to get into a deadly embrace.

>I gave up using that bit-pattern arrangement from COBOL some time ago,
>particularly in situations where I might want to vary the delay time.  Now
>when I want a pause I open a temporary message file as a reader at the
>start of the program and for each "pause" I post a timed FREAD.  Delay
time
>is in seconds expressed as an integer, so it's easy to change it or
DISPLAY
>it for debugging.

What I would like to be able to do is either use REAL's from COBOL or call 
an intermediate intrinsic (like CPAUSE) that accepts a 32-bit integer and 
converts it to REAL (to call PAUSE).  You have proposed an interesting
idea, 
it's just that I don't like multi-level mis-direction.  If the purpose of 
a call is to wait a specified number of seconds, then call the intrinsic 
designed for that purpose and not another intrinsic which just happens to 
provide what you want as a byproduct.

Just my opinion.  I'm sure your scheme works well for you.  Thanks for the 
alternatives.

Jim Phillips                            Manager of Information Systems
E-Mail: [log in to unmask]      Therm-O-Link, Inc.
Phone: (330) 527-2124                   P. O. Box 285
  Fax: (330) 527-2123                   Garrettsville, Ohio  44231