Not that anyone does.
When I was in high school it didn't seem like there even was a viable option other than Computer Science, which I know now is largely unrelated to most software development. Of course, you could argue I didn't do my due diligence, but it was the guidance counselor's job to guide me and I trusted them implicitly. In retrospect, that in and of itself taught me a valuable lesson about critical thinking.
Everyone I've talked to has said they haven't used a single bit of calculus in a professional capacity since graduating. I may be succumbing to confirmation bias, but I've never heard to the contrary.
Regardless, the issue isn't to pick your path exactly before you go. The point is to be prepared to dedicate some marginal effort to it either at the beginning or before.
You really think it's on the schools to tell students what they should learn or what degree they should get? Certainly people have different amounts of support going into it, but we can't just keep pushing the age of personal responsibility further and further out. To absolve underprivileged young adults of these key steps in life is not only harmful to their personal development, but I'd argue personally insulting to their potential as independent adults.
No, not me. I agree with you that we've pushed the age of adulthood too far out.
But when the mistake you learn from comes with crippling debt, it is on us to guide kids to make low-consequence mistakes before the big ones. That's why I suggest we encourage more real-world exposure with internships, co-op programs, shitty jobs first. We need to foster the environment that allows kids to actually critically think about their future instead of shoving them towards it, and we need to practice the meritocracy we preach by looking past degrees to hire people.
Others require you to pretty much declare from day one, and it might well be another $50,000 if you get it wrong.
I'm not truly blaming the student. It's an honest mistake to make. However, I see this same mistake (CS == programming) propagated online by people who should know better. Whenever someone links to "The Perils of Java Schools" by Spolsky, in relation to a debate about CS, he/she is doing this. Java/C#/C++/${LANGUAGE} doesn't matter in this context because a CS education isn't about learning a particular programming language. And this article, while not written by one of these people who should know better, doesn't help.
"CS != Programming" is also a mistake. An over-generalization, at best. A good CS program will involve a lot of programming and be a great foundation for becoming a programmer.
Indeed, a good CS program is a great foundation for becoming a programmer. But CS is not about programming, at least not in the sense that gets discussed when Spolsky laments the "perils of Java Schools". Also, a CS researcher may not even be particularly interested in actual programming languages, preferring to focus on theoretical matters. Some of them are barely distinguishable from applied mathematicians.
People who want to learn programming may get impatient when they are taught about graph theory, lambda calculus, big-O, Turing machines, algebra, calculus, statistics, computability, etc. Not saying they necessarily will, and in fact all of this may surprise them in a good way, but many times -- which is what you often read in this kind of articles -- they become frustrated: "what is this nonsense? Why don't they teach $LANGUAGE? Everyone is using $LANGUAGE! All these CS types live in their ivory tower and don't even understand what is being used to build $MOBILE_PLATFORM apps! None of this helps me get my startup running!". The conclusion is obvious: academia "doesn't get" the real world.
In reality, however, it's more of a case of mismatched expectations. CS is -- if you look at it with squinty eyes -- a kind of applied mathematics; one in which you get to play with both idealized and real-world computers. Ok, that's another oversimplification, but it's no coincidence the founding fathers (and mothers :P) of CS were mathematicians, logicians, physicists, and the like. They would scoff at the notion that "you don't need Calculus".
It's perfectly fine to find the research/theoretical aspect of computers uninteresting. It's perfectly fine to want to learn Ruby/C++/whatever and just build your site. But it's not fine to go into CS, and find it too theoretical because they don't just teach you Python (or whatever language is currently being used in the real world), or force you to sit through algebra. In that case, you simply picked the wrong career path, and it's not academia's fault.
Did I know as a 17-year-old applying to college what the hell the difference was? No. Did I study CS? Yes. Was it the right decision? Probably not (though I did stay in it). And I had about every advantage in the world of being exposed to programming from when I was 5 or 6.
Maybe the craziest part of the US college system, and technical fields in particular, is that most people don't know what's available, and don't know what they're applying for, all the while they go somewhere between the cost of a car and a house into debt.