Code generally uses only ASCII!

[pne]

I always get a little bit sad when I see documentation that includes code snippets that have curly quotes, such as print “This is a test.\n”. I doubt that that code would even compile if you copy-pasted it into a file and compiled/ran it with the language in question. What's even worse is German documentation that takes "smart" quotes one step further and gives you print „This is a test.\n“—which not only has a really ugly closing quote in the Courier New so often used for code examples (the lines should point NE/SW rather than SW/SE for a German closing single quote IMO), but is even less similar to the correct print "This is a test.\n".

A related sadness comes when two consecutive hyphens turn into an en or em dash… reading about program invocations such as hello —foo=bar —baz=qux seems to me also a mark of bad typography (so as not to say "incompetence").

And finally, reading bits of code that use ligatures is something along those lines as well: my $file, for example. Eep.