There's no dependency between learning functional programming, including Haskell, and category theory. I say this as someone who is reading a CT book (Conceptual Mathematics) after being first exposed to the topic by learning Haskell.
I'm currently trying to learn functional programming through Haskell. Glad to have found another resource, plus this cheat sheet linked from the page looks handy: http://cheatsheet.codeslower.com/
"I actually started doing Udacity's CS class when they teach you how to design a search engine, however half way through I realized that this can be done twice as faster with tools that already exist."
I may be misinterpreting your claim, but they're not trying to teach you how to quickly build a search engine. They're introducing fundamental concepts of computer science and showing their applications via a search engine.
"What would help, in my mind, is for kids to be able to learn each subject at their own, separate pace--one "micro-skill" at a time, advancing to the next only when all the micro-skills relied upon as a base for the next micro-skill have been mastered (100%ed). [Think of it sort of like an unlockable "tech tree" of education.]"
This is exactly what the Khan Academy is doing. I believe there's at least one school (not sure if public or private) where they're experimenting with that model. See:
IIRC, the students in the class take most of the lessons via computer/tablet and move forward once they've answered a certain number of quiz problems correctly. The teacher can monitor the progress of each student, and if one is having difficulty than he/she can get some individual help from the teacher or another student.
I took Udacity's cs101 class as a (lazy) way to get introduced to python. While I felt like it was a very well done introduction to computer science, it wasn't a very effective way to learn a language as it made a deliberate effort to decouple python and cs (and rightfully so). I would suspect that this is also true of the other MOOCs listed here.