The Declaration Approach (was: Some thoughts on the 'reveal' operator)

[Mark S. Miller]

At 11:01 AM 10/21/99 , Chip Morningstar wrote:
>Any non-keyword symbol is going to be pretty arbitrary (which is why I like
>"reveal") but I *really* dislike "=>". For one thing, the arrow points in the
>wrong direction -- the value comes out of the function, but the arrow points
>into it (when this token was originally suggested, the arrow was pointing at
>the name of the variable that was revealed, not at the function itself).

We have opposite directionality intuitions.  For me, "=>" points in exactly the right direction, since an invocation of the function/method declared on the left yields the value that the block on the right evaluates to.  Ie, 

     def square(x) => {x*x}

seems perfectly intuitive to me, while

     def square(x) <-- {x*x}

seems completely wrong.  But since you strongly have the opposite reaction, the search goes on.

>If we're being completely
>arbitrary anyway, then you might use "@" (which I don't recall having another
>meaning in E).

It has an incompatible use in quasi-literal patterns.  See http://www.erights.org/elang/grammar/quasi-overview.html

>  The best choice in terms of meaning seems to be "^", but we
>already dismissed that for ergonomic reasons. 

Perhaps a variant?

>"&" maybe?  

No.  "&" is only associated with two meaning-clouds in my head: and-like and address-of.  I can't imagine stretching either one to fit.


         Cheers,
         --MarkM