User friendly means "easy to use". We can all agree on that, I think. And it's what we mean when we say it. So far, so good.
Except, not. When people say "easy to use", what they REALLY mean is "easy to learn", or "easy to use without learning". And it's important, also, to realise the difference between the latter two. In my Why Windows Causes Stupidity rant, I elaborated on how Windows is an archetypical example of this latter kind of thinking, and on some of the detrimental effects of it. Now, I want to examine it from the user friendly side of the equation, and maybe figure out where to go from there.
If we accept the non-pejorative definition of "stupid" given in the Why Windows Causes Stupidity rant, then we can coin a phrase. And so now we have "user friendly" (easy to use), and "stupid-user friendly" (easy to use without learning). And we have "easy to learn" floating around without a good name for it.
Let's create a vocabulary here. On the one hand, we have the usability concept. We can call this "user friendly", we can call this "easy to use", or we can just call it "usability". This is the sort of concept that people like Jakob Nielsen go on about, though what he really means is something halfway between real "usability" and "learnability". It's an important concept; the easier something is to use, the better a tool it is.
Then, we have the learnability concept. This is "ease of learning". And "learnability" is a good term for it. It's also very important; if you can't learn something, you can't use it.
And then there's the red-headed stepchild of the group; "ease of using without learning". My Why Windows Causes Stupidity rant is pretty much about that, and I think it illustrates why it's not a positive goal. It's self-defeating, and it's a terrible way to design. It forces the designer to try and anticipate every possible outcome and second-guess the user. And it destroys flexibility.
Now, armed with these definitions, we can move forward and start applying them in the context of operating systems. We'll look at the strengths and weaknesses of various approaches, and see if we can't determine ways of improving things.
So, let's get on with it, and start examining the details.