Would you give an example of one of the cases you have seen of abstraction used for the sake of obfuscation? I'm not sure what you mean by that.


Not really. For example, I may not know how to calculate square roots using only the simple arithmetic operations provided by my CPU. But I do know how to call the sqrt() function provided by the standard C library. This is a perfect example where an abstraction (the sqrt() function) allows me to do something (calculating the square root) that I don't understand. Far from "guaranteeing failure", it allows me to get my work done without having to reinvent the wheel at every step. I would be much more likely to fail without the abstraction.

I'm starting to guess that you're not really challenging the idea of abstraction itself, but maybe something that you don't like about the Attila's idea of an 8051 API in particular. Is that a good guess? If so, it might be helpful for you to explain your specific objection to what the OP is trying to do.

I think it would be very difficult to make an 8051 API that was flexible enough and universal enough (i.e., "applicable to enough 8051 variants") to be useful, but I don't think the concept itself is flawed. I mean, suppose somebody actually did it, and it was really possible to write something like the following and just know that it would work:

 
#define VARIANT SILABS_C8051F120
#include "8051api.h"

void main() {
    InitializeUART0(BAUD_9600);
    StringToUART0("Hello, Attila");
}
 

Wouldn't that be a liberating change from the endless farting around with SFRs that we do now?

-- Russ