Re: Minor last-minute issues David A. Wheeler 19 Sep 2012 00:03 UTC

Per Bothner:
> No. Kawa maps:
> [foo bar] to ($bracket-list$ foo bar)

I think we should *not* require a particular definition for unprefixed [...].  Many Schemes use that as a synonym for (...), and I want to minimize syntactic changes (such differences would be a source of bugs).  A particular *implementation* might support [...] as having a special meaning, but that should be independent.

> x[foo bar] to ($bracket-apply$ x foo bar)

Ah, the proposal is to use $bracket-apply$ instead of bracketaccess.

I think it's critical that the bracket access symbol be available in other Lisps too - at least Scheme and Common Lisp.  My intent is for this notation to work *beyond* Scheme.  But $bracket-access$ is legal in Scheme (at least R5RS) and Common Lisp ("$" is a constituent character), so I think it meets that criteria.

I'm not fond of $bracket-apply$ - it's a little ugly.  But perhaps its ugliness is a virtual; people are unlikely to use it as an identifier.  And being compatible with a previous convention - especially if people actually use it in real code - has its pluses.

Is there any code that depends on $bracket-apply$?  How much?  And are there any other thoughts on this, good or bad?

--- David A. Wheeler