> This is a nice, comfy thought but it's just not true.
Actually I think most people found it profoundly uncomfortable.
> Any suitably competent and experienced programmer could pick up a new programming language in a week or two of focused study, but they'll basically be starting from scratch if they're learning HTML + CSS for the first time.
This has nothing to do with your point. Most programmers struggle to pick up Lisp, Haskell and ML for the first time, but those are unambiguously programming. TeX is usually a bit tricky too.
> Meanwhile, a web developer who only knows HTML and CSS is entirely unprepared to do any serious work in any real programming language, and would be starting from scratch if he decided to pick up C.
Much as it pains me to do this, I cite Node.js as a counter example. People can and do write systems software using their understanding of Javascript as a basis. They may not do it well, but I know for a fact I could write certain types of servers acceptably well with it (I would prefer other tools, though).
Note that just because Ryah and co do not know how to write systems software, the basic principle is not invalidated. With some serious hacking to the node.js vm, it could be a viable competitor to Erlang.
This argument is incredibly weak, and basically self-refuting. But we have to repeat it every now and then.
> but it does mean that the process of telling a computer what to do and the process of telling a computer "here's some text, and here's some guidelines about how you should display it" don't inform each other.
You keep re-asserting this but haven't given any quantitive metrics for how it is different. You just keep saying it like I'm supposed to agree.