Excerpted from message dated 05-23-95, Paul Sidorsky to Murray Lesser:

PS>One related question - how accurate are the delay/wait times passed
  >to the OS/2 API functions, such as DosSleep() and the
  >DosWait...Sem() functions?  If they're accurate to the millisecond,
  >then there logically should be some way to time to at least the
  >millisecond level without a device driver.

Hello Paul--

    The "times passed to" are very accurate ;
"times returned from"
are not quite so.  I quote from the "Remarks" for DosSleep (and a couple
of related API descriptions) in the Warp "Control Program Programming
Reference":

      Time intervals for DosSleep, DosAsyncTimer, and DosStartTimer are
      specified in milliseconds; however, it is important to recognize
      that the actual time interval will be affected by two factors...

      Because clock ticks are less precise than millisecond values, any
      time specified in milliseconds will be rounded up to the next
      clock tick.

      ...[B]ecause the system is a priority-based multitasking operating
      system, there is no guarantee that a thread will resume
      immediately after the timer interval expires.

    As far as [apparent] timing to the millisecond under DOS is
concerned, there is the same clock-tick problem for reported timings in
DOS (when using the DOS API [INT 21H] functions) as in OS/2; but the
clock ticks in DOS are a little further apart!  Roughly 55ms instead of
31.25ms.  Don't believe everything you see on a display screen!

PS>Thanks again for the info.

    You're welcome.

                --Murray
___
 * MR/2 2.23 #120 * Watching for speed bumps on the Information Highwy
                                                                           
                                  
---