I don't think illustrating your point by naming an absolute mathematical genius is very optimal.
I mentioned Ramanujan because he's famous, and he obviously didn't get it from taking university classes. I used to do math competitions in high school, and I'd meet the best students in the state or country. You know what? All of us learned the same way Ramanujan did: we read number theory books for fun, continually asked questions that were interesting to us, and then worked on them until we answered them. Not a single person in the group learned by attending lectures or even through supervision by experts. It's just not how it happens. That's the point I was trying to make concisely before.
Math, like CS and say, any of the arts, live on their own. I think then what we have is perhaps an improper classification: Math and Computer Science are arts! :-0
I think the major difference is that in each of the fields you quoted, they only place to actually learn the skill is in the university classroom. The knowledge being passed on is hidden and selective. The same is not true with computer science. (Interesting though, that with things such as MIT Open Course Ware we may see this change in the future)
More and more, a college degree has gone from being a symbol of educational merit to being a caveat against absolute stupidity. It's lost all its original meaning.
I often see university being quoted as "showing that you can finish something/work hard to achieve a goal". Often, this line is spouted by those whose only achievement in life is that degree. It's simply not an accurate measure of anything, but it's the only thing in that a person in that position can relate to.
As somebody that spent my early 20's slogging through mud and being responsible for the lives of 8 men while surrounded by explosives and live ordinance, I personally find the argument that I didn't drink like a fish and attend a few classes hungover and therefore don't have a commitment to finishing what I started insulting.