An extreme (in several respects) and not-necessarily-fair example:
I recently saw a bright 10 year old child learn two editors. He
first learned the Apple Logo program editor, a super pared-down Emacs
clone. He picked it up quickly and was having so much fun using it
that he would spend as much time just sitting and "playing" at
changing things on the screen as actually creating/editing/running
Logo procedures.
Several months later, he learned (a subset of) the Apple (UCSD)
Pascal editor in order to be able to write reports for school, etc.
(The Logo editor can't be used to create a text file.) The Pascal
editor is heavily moded, with separate modes for insertion, deletion,
overtype, cursor movement, etc. Even though the mode was constantly
displayed on a "mode line," he had no end of trouble. Of course he
had the usual problem of forgetting which mode he was in, but in
addition, he seemed to have difficulty forming a conceptual model of
the editor that would correctly tell him which mode he should be in
at any given time. Particularly annoying was the way a single key
could perform completely unrelated editing functions in the different
modes. Learning each mode was like learning a separate editor.
Although at this point he has used the Pascal editor more than the
Logo editor, he still has much more trouble with simple editing tasks
in the Pascal editor.
-jeff
[Editor's note: this might, of course, be a primacy effect. Have
any of you developmental psychologists out there ever studied
children learning to use computers, either editors or command
languages? /jqj]