In perl, we can’t use the sleep function to do a sub-second sleep, because it takes an integer argument as the number of seconds. Normally, the cleanest way to do this is via the Time::HiRes module. For example if I wanted to sleep for a 100th of a second I would write:

use Time::HiRes qw(usleep);
usleep(10000);

Pretty straight forward, but in the land of golfing efficiency, this is a few too many keystrokes. You may be tempted to shell out to use the unix sleep command, which will take a float of seconds, eg.

`sleep .01`;

But we’ve lost a bit of “purity.” In my travels, I’ve come across this little gem which will achieve such a sleep:

select$a,$a,$a,.01;

Here, $a is an undefined scalar. What’s happening is that select will block for .01 seconds waiting to see if any of the first three parameters is a file handle with data ready to be read. Since all three are undef, we block for all .01 seconds. Voila, a shorter way to express a sub-second sleep.