Does any of the new programming languages everyone seems to force upon poor old LLVM support function names with spaces? No weird unicode characters, regular spaces.
That used to be a thing waaaay back in the ALGOL age, but apparently we've got some kind of StockholmSyndrome with CamelCase now.
Follow
I think the main problem lies in the way the parsers/compilers work. They tokenize before they have any clue about syntax, declarations and so on.
Way back in time, compilers weren't cold, off-the-shelf and academic, but handcrafted art pieces.
@niclas On the other hand, we also know a lot more about parsing. And skipping white space is done anyways for other parts, just not within identifiers.
I would guess it's a combination of being used to how variables/functions are "supposed" to be named (upper/lower ASCII plus underscore maaaaybe a minus), grep gripers (ambiguity when using braindead tools) and the inability to "strop" keywords. Can't sell upper-case keywords.