For those of you that don't know, LaTeX is the way to write scientific papers. It's set up to produce nice equations, put figures and tables more or less where you want them, collect references in the proper order (with the help of BibTeX), and generally make you look like you know what you're talking about. Publishers can even set up their own style sheets and macros (see the American Physical Society's REVTeX) so that your article looks right for the journal, and it's a minimal amount of effort to translate your work to the format of another journal.
That said, in scientific articles you often need weird characters, e.g. √, &Delta, ∇, etc. Since LaTeX's source is ASCII, you need to know that √ = \sqrt or \textsurd, Δ = \Delta, and ∇ = \nabla. Not to mention hundreds of other characters.
You'll eventually remember that α = \alpha, but will you ever learn that ⊃ = \supset? I won't. You could spend all your time looking up these symbols in tables (as I just did for the HTML), or you can go to the Detexify2 – LaTeX symbol classifier. Just draw in the symbol you want with a mouse, and you get a set of suggested LaTeX commands.
LaTeX and HTML handle formatting non-ASCII characters in similar ways. It would sure be helpful if someone would adapt this to work for HTML characters as well.