Remove `?` from allowed characters in ID, Symbol, Keyword and NamedArgument about fluent

[Pike]

sgtm

[mathjazz]

I'm ok, too.

[stasm]

I think we should follow what we did in #50 and forbid ? at the beginning of identiifers, and allow it elsewhere. I feel strongly about the benefit of allowing identifiers like really-close-file? as opposed to really-close-file. Same with function names: VOWEL?() reads better to me than IS_VOWEL().

[zbraniecki]

Just for the record: I feel the opposite to @stasm - that ? as a sigil is unnecessary, unfamiliar (not present in CSS, JS, PHP etc.) and on top of things removes a sigil from the scope of potential semantic sigils in the future.
Due to the nature of the sigil, it's likely only going to be present at the end of the variable/id/arg which increases the confusion factor since it can be both.

I also believe that for dont-make-me-think reasons, IS_VOVEL is better than VOVEL?() and really-close-file = Really close file? is better than really-close-file? = Really close file?.

[stasm]

In #51 we made it illegal for an Identifier to start with -, _ or ?. This answers most of the concerns outlined in the OP. I would still like ? to be a valid character at other positions in identifiers. Most languages disallow it because they use ? for ternary ifs or wildcards. We don't have these limitations and can thus follow in the footsteps of Clojure or Ruby.