Multithreading Waitable Object

This is some terminology in some documentation I cannot be responsible for.

I just took a look at some object and think I can see the difference. To my taste, it is really minor. There is one simplification which makes Critical Section different. It can be considered as a lightweight version of Mutex. It has a bit simplified interface and can only be used for synchronization between threads of the same process, not between threads of different processes.

First of all, the 'Synchronization Objects' primitives bring the thread in a special wait state. OS switches out the thread entering the Critical Section or waiting for other Synchronization Objects and doesn't schedule it back to execution until it is waked up by a certain interrupt condition. In this way the thread consumes strictly zero CPU time until it is awakened.

In case of Critical Section, the thread is blocked when some other thread had already entered the same Critical Section instance and is awakened when the other thread leaves this Critical Section instance.

For the very similar Mutex and many other Synchronization Objects there is another special wake-up condition: a timeout. This is the difference. The Synchronization Objects which have an (optional or additional) blocking method with a timeout parameter are called Waitable.

This is the answer. Your only problem is that this term is not really self-descriptive but somewhat misleading. I personally try to avoid it.

—SA