Brian Harvey on Scheme vs. Python
Scheme vs. Python is an excellent read, but I would be more impressed if he took points 3 and 4 out back and shot them in the head. "[Y]ou should be able to learn a programming language (after the first time) over a weekend" is perfectly true (for some reasonable value of "weekend"), for a Computer Science program. "The best language for a course is not necessarily the best language for writing real-world code" is an excellent point as well. "You can learn to program in any language [but] the big ideas in [SICP]...express themselves best in Scheme" is (arguably, I admit) also valid.
But points 3 and 4 are, first of all, the same "Everyone should use Lisp" argument that hasn't actually convinced anyone in the 50-year history of Lisp. They weaken the actual point I think he's trying to make, or at least that I wish he was trying to make.
What, exactly does "the lifespan of a programming language is closer to the lifespan of a dog than to that of a person" mean? What is the lifespan of a programming language? I don't know of any programming languages that have achieved some kind of critical mass that can be meaningfully said to have "died".
Furthermore, fundamentally, what is "Lisp"? LISP 1.5, a la McCarthy? Is anyone still using that? Common Lisp? Good luck with recursion there; recursion without tail call elimination is not an especially good idea. Scheme != Lisp, any more than C, C++, Java, et al., are the same as ALGOL. There are ideas that unite all of these languages, but the fundamental one, that programs are data and that data is a program, is not one of "Lisp’s ideas [demanded by 'real users'] in the programming languages they do use."
Finally, what the heck is up with "[u]sers of strongly typed languages demanded, and got, Lisp's heterogeneous lists." I'm not seeing it.