Well, incidentally yes, but it is not the point.
I don't know how it is decided that a feature/function should be part of the core or a lib and what belongs to the core or to a lib. From my arduinoobness, I see the project as (more than, but including) a set of functions that come as a layer over the avr compiler, a set that is willing to grow thanks to a community of users, basically.
And when a function is recurrently used by users on different projects, it seems sensible to add it to the project (it can be a lib, perhaps I shouldn't have said the "core"). I don't know the process for this integration. A suggestion in this forum could be a first step.
This integration would help people by making available a function
- that would be portable (some implementations you find on the web don't set the ADC registry correctly for all ATMegas)
- that would free the developers from the need for looking in the ATMega spec for the registry descriptions
- that would just be correct (because adding it to the project would mean finding a consensus on how it should be done, therefore settling the debate)
What you refer to as "a correct implementation".
I apologise if I'm misunderstanding the goals of the project or the processes of decision. My point was not just to have a final answer to the initial debate. I would have asked in the proper forum section. I'm happy with 1024, I think it is pretty clear in the spec, and anyway, 1204 or 1023, I don't need that kind of precision. I just figured such a function could be a candidate to integration. I mentioned the debate to underline one of the advantages of integration : unification.
I read a lot of arduino.cc already. Perhaps you can point me to a page that would instruct me about the life of the project, decision process, patch submission, etc.