What are hardware abstraction layers and how do they work in microcontroller architectures?

In my experience, abstraction becomes necessary when you want to incorporate laymen into a system requiring expertise. It also helps expert users transition to a higher-level of interaction. As an engineer you reach into yourself and find analogies that right true in a layman's worldview and you use them to map information in a special logical format that the laymen can rely on to have a handle on system's functions. For example the keyboards we use having letters printed on them and matching graphic accelerators so that when a letter is punched the electrical signal represented shows as a matching letter on screen so users can use keyboards to type meaningful text using hardware whose inner workings remains a mystery them them.

For folks coming from a C++ world, the idea of HAL is comparable to the idea of Abstract Classes in C++. One can imagine every microcontroller board has its own child class realizing the top level abstract class. In this way, the user can continue to use the same API exposed by the top abstract class and not be concerned about the implementation details of different child class for different boards. Implementation of child class needs the knowledge of registers/configuration/gpio pin access to that hardware, etc. These child classes is nothing but the driver code. All in all, HAL is equivalent to abstract classes in the C++ world.

HAL stands for "Hardware Abstraction Layer." It is a software layer that serves as an interface between hardware and higher-level software, providing a standardized way for software to interact with hardware components without needing to know specific details about the underlying hardware. The main purpose of a HAL is to abstract the low-level hardware details and provide a uniform interface for device access and control. This abstraction simplifies the development of software applications by allowing Developer write code that is independent of the underlying hardware platform. It also facilitates portability, as software written to interface with the HAL can potentially run on different hardware platforms with minimal or no modifications.

Third sentence in this paragraph has an error... instead of "...by separating the hardware-dependent and hardware-dependent parts" perhaps I think it should read "...by separating the hardware-dependent and software-dependent parts"..? Happy to help.

Microcontrollers often have limited resources like memory and processing power. Implementing a HAL may consume a significant portion of these resources, leaving less room for inclusion of application-specific code. Efficient code optimization and customization is essential to mitigate this challenge.

Arduino boards come with connectors with writings printed alongside - the printings themselves are abstractions aimed at being understood by higher-level developers rather than being understood by teams that manufacture the boards themselves.