My time to chime in!
No spaces - This was never a good idea to begin with!
Case-sensitive - There are a few reasons why this should be done. I guess the most important is that filenames in non-Windows OSes are largely case-sensitive, so if you make the language that way then they should 'in-theory' natually keep an eye out for their filenames as to not create massive deployment issues which can't be caught at compile time in all cases(pun!).
Plus, when you say case-insensitive, you're only talking about the English language in the ANSI range, these days you should really be supporting UTF-8(and supporting it is just a case of not not-supporting it), so unless you write the never perfect behemoth that is
supporting proper case-insensitivity, then it'll always be partially case-sensitive.
I personally always use upper camel case(LikeThis) for types and functions, but lower camel case(likeThis) for variables. And with the exception of underscores(which haven't even been mentioned here)this approach is virtually standard across all C/C++/Java/whatever code. The only main point of contention you're left with is initialisms and acronyms, I used to do all capitals (RGB, HSV, etc), but I sometimes come across code where I need to have two in a row, this quickly made me change my mind and so ColourHsvRgbWhatever makes far more sense than the alternatives I've seen, so I think that standard should be used.
Brackets around all functions! - This just makes the language far more consistent, in DBPro it made no sense that commands that returned nothing didn't require brackets, but user functions that also returned nothing did. Also, not having brackets could maybe be used to pass functions around in variables? This would allow for callbacks.
I have a lot of ideas for things that go outside of the scope of this topic so maybe I'll make a thread for it!