>>30
I'm
>>25 and not
>>29, and your opinion is wrong. There are no reasons for new programmers to want to write free-form code. Force them to use good style when they're new, and they'll use it forever. This is why university-level programming courses will fail you for using things like bad variable/function naming styles. They need to hammer good styles into you to make you want to write maintainable code.
>>34
>Learn Scheme. It will make you a better programmer for the rest of your days. Honestly. Read this book:
http://www.htdp.org/
No, don't learn Scheme, at least not yet. It's too difficult/frustrating for a new programmer, and you won't find it any fun because there just isn't as much online help for writing cool things with it; for example there are a billion tutorials on writing games in C++, and virtually none for Scheme. Python is easy enough that you'll be able to write fun and interactive programs.
Also, don't bother learning assembler. Read through a basic description of how it works (see what the opcodes look like, see a basic description of the internals of a processor, etc), and that's enough to give you an understanding of what a compiler breaks your code down to. You will never, *ever* use assembler, unless programming will be your career; it's only really useful if you're doing speed-critical applications (certain math functions in games, high-level math research/computation, etc).
>>37
Fortran is not surprising, because it's still widely used (in fact it's probably the most common language) in the pure sciences. I did my honours project in physics doing particle accelerator simulations using Fortran. There is *so much* mathematical/scientific legacy code in Fortran that it's just easier to use it instead of something newer, because half the work is already done for you.
As for Ada and COBOL, I don't know what to say. Haskell is just not catching on.