Update prime_counting example to demonstrate limiting the message rate

[mdickinson]

The bug observed in #442 appears to be at least partly due to a buggy release of PySide2, which was unable to handle too many requests efficiently.

However, even with a non-buggy PySide2, it would be easy to overwhelm the main thread with too many progress messages from a background thread. It would be worth updating the prime_counting example to demonstrate how to implement a simple rate-limiting feature on the progress messages (only send a message if no message was sent within the last 0.1 seconds, say).