[racket] racket 5.4 syntax tweak wishlist

On Mon, Sep 10, 2012 at 9:52 AM, Neil Van Dyke <neil at neilvandyke.org> wrote:

> Two syntax tweaks I'd like to see in Racket 5.4:
>
> * Colon keywords -- Make symbols beginning with the colon character read
> as keywords.  Possibly also make keywords print as starting with colon
> rather than pound-colon.  I know some people don't mind "#:", but I don't
> like typing it or looking at it.  I think "#:" also looks gratuitously
> different/cumbersome to people coming from Lisps, and unnecessarily cryptic
> to people just coming from anywhere else.
>

+1, but I suspect that some people really prefer the #: because it makes
the fact that it's only syntax apparent?


> * Multiple-value LHS in "let" forms -- In the "let"* forms (i.e., "let",
> "let*", "letrec"), in addition to permitting the LHS of a binding to be an
> identifier, also permit LHS to specify multiple idents for multiple-value
> binding.  Two motivations: (1) "let-values" and friends are painfully long
> for something as basic as a binding form, and increase rightward drift; and
> (2) when editing code to change an existing use of a single-values "let"*
> form to add a multiple-values to it, having a multiple-values LHS doesn't
> require adding a bunch of extra parens to the existing single-values.
>  Syntax-wise, I would be equally happy with any of the three alternatives
> for multiple-value LHS: "(values ID ...+)", or drop the keyword to have
> just "(ID ...+)". or even drop the parens so LHS syntax is simply "ID ...+".
>

I'd something like that too, although if you can write (let ([(a b c)
(values 1 2 3)]) ....), then you may infer that you can write (define (a b
c) (values 1 2 3)), which is wrong.
Yet, you can't write (let ([(f x) (+ x 1)]) (f 3)), so let is already
different from define.

Laurent
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.racket-lang.org/users/archive/attachments/20120910/c8babc37/attachment.html>