The last scheduling paradigm that developers can use to schedule tasks is the MicroPython thread mechanism. In a microcontroller-based system, a thread is essentially a synonym for a task. There are some minor differences, but they are beyond the scope of this book. Developers can create threads that will contain task code. Each task could then use several different mechanisms to execute their task code, such as the following:

• Waiting on a queue
• Waiting on time using a delay
• Periodically monitoring for a polled event

The thread mechanism has been implemented directly from Python 3.x and provides developers with an easy method for creating separate tasks in their application. It is important to recognize that the Python thread mechanism is NOT deterministic. This means that it will not be useful for developing software that has a hard real-time requirement. The MicroPython thread mechanism is also currently experimental! Threads are not supported in all MicroPython ports and for the ones that are, a developer usually needs to enable threads and recompile the kernel in order to have access to the capability on offer.

Starting with MicroPython version 1.8.2, there is support for an experimental threads module that developers can use to create separate threads. Using the threads module is not recommended for developers who are just getting started with MicroPython for several reasons. First, by default, threading is not enabled in the MicroPython kernel. Developers need to enable threading and then recompile and deploy the kernel. Second, since the threading module is experimental, it has not been ported to every MicroPython port yet. 

If threads aren't officially supported and not recommended, why are we even talking about them? Well, if we want to understand the different scheduling mechanisms available to us with MicroPython, we need to include the mechanisms that are even experimental. So, let's dive in and talk about threading with MicroPython (even though you may not be able to run a threading application until you have learned how to recompile the kernel, which you will do in Chapter 5, Customizing the MicroPython Kernel Start Up Code). 

When a developer creates a thread, they are creating a semi-independent program. If you think back to what a typical program looks like, it starts with an initialization section and then enters into an infinite loop. Every thread has this structure! There is a section to initialize the thread and its variables, followed by an independent loop. The loop itself can be periodic by using time.sleep_ms() or it can block an event, such as an interrupt.