What are the differences between a process and thread?

Answer 1

Every process has at least one thread, known as the main thread. However, threads can spawn other threads of execution. All threads within a process share the same data segment (constants, static variables and global variables) and the same virtual address space. However each has its own independent stack. All active threads in a process get an equal share of CPU time, thus they operate concurrently. If the system has multiple cores, separate threads can also operate simultaneously.

In multi-tasking environments, concurrent threads are vital. For instance, rather than the main thread performing tasks it can initialise its environment and then enter an "idle mode", a tight loop that simply monitors a message queue and dispatches those messages accordingly. If a message indicates that a lengthy task must be performed, the main thread would be unable to monitor the message queue for the duration of that task and the application becomes unresponsive to user-input while that task is going on. But if the task is delegated to a "worker thread", the main thread can continue to accept user-input. Meanwhile the worker thread can complete its task and signal the main thread via the message queue.

The only problem is that if the main thread is terminated (perhaps the user decided to end the program in the middle of a task), it must signal all worker threads to abort before the main thread can end. To achieve this, the main thread simply posts messages for each of its threads onto its own queue. When one of these messages reaches the front of the queue it will simply ignore it. Meanwhile the worker threads will periodically examine the queue to see if there are any messages intended for it and process them accordingly, removing them from the queue and signalling the message queue when they've completed the abortion. Once all workers have signalled that they have aborted, the main thread can safely terminate.