Path: utzoo!attcan!uunet!samsung!crackers!m2c!umvlsi!umaecs!amh!mbabramowicz
From: mbabramowicz@amherst.bitnet
Newsgroups: comp.sys.mac.programmer
Subject: Re: Timer
Message-ID: <10719.271f09c1@amherst.bitnet>
Date: 19 Oct 90 14:12:17 GMT
Lines: 44

I think that it is possible to get better-than tick accuracy without installing
any new hardware or relying on certain model Macintoshes.

(Note: I haven't read the articles referenced previously, so perhaps someone
has already written about this.)

The for loop that someone suggested earlier is actually useful, with
modifications:

	Time = Ticks();
	for (count = 0; (count<= 2000) && (!Button()); count++);
	Time = Ticks() - Time;
	oneTick = 2000/Time;

The variable oneTick is now set equal to the number of times you go through
the for loop in a Tick.

Then, if you want to track user reaction time, you use the same for loop as
before. (To preserve the exactness of your count, be sure to keep in the
count<=2000 line since that computation takes some time.  You can make it
count<= something big if you want the user to be able to take longer than
2000.)

When the user hits the mouse button, the loop exits.
count / oneTick * 60 should now give the time the loop took, in seconds.
Dividing by 1000 gives milliseconds.

Note of course that for accuracy oneTick should be a float or a double.

If you use a number higher than 2000 you will obtain greater accuracy.

This will work on any Macintosh, unless they make a Macintosh so fast that it
can do 2000 iterations in less than a tick. (Maybe some now can do this.  I
don't really know.  If so, just use 100,000 or a 1,000,000.)

One little annoying thing about this method is what happens if the user presses
the mouse button while you are doing your test? I haven't worked that out yet,
but I doubt that a workaround would be complex.

I haven't actually tried this out; this is all speculation; all disclaimers
apply.

Michael Abramowicz
Amherst College