[racket-dev] [racket] Plea for neologism

From: Matthias Felleisen (matthias at ccs.neu.edu)
Date: Wed Nov 24 16:56:48 EST 2010


   A[e_1 & e_2] = A[e_1] meta& A[e_2] 

compositional? How about 

   A[e_1 & e_2]env = A[e_1]env meta& A[e_2]env
   A[e_1 & e_2]env k = A[e_1]env (\x_1. A[e_2]env (\x_2. k (x_1 meta& x_2)))
   A[e_1 & e_2]env k s = A[e_1]env (\x_1s_1. A[e_2]env (\x_2 s_2. k (x_1 meta& x_2) s_?) s_1

On Nov 24, 2010, at 1:20 PM, Don Blaheta wrote:

> How about "noncompositional"?  This word and its opposite have fairly
> technical linguistic meanings.  A "compositional" phrase (= expression)
> is one whose meaning can be (correctly) inferred only by knowing the
> meanings of their parts and the semantic rule associated with the syntax
> form of the expression.  So an expression like "a red apple" means
> precisely what you would expect if you knew the meaning of "a", "red",
> and "apple", and knew how to combine a determinative, an adjective, and
> a noun into a noun phrase.
> It's not a perfect match.  Single-word anaphor (like "it") wouldn't
> normally be called noncompositional because there's nothing to compose
> there---the word just has a complex meaning.  And in natural language
> there's no analogue at all to the more complex things that macros can
> do.  But to the extent that a hygienic macro system tries to make it
> difficult (or impossible) to write macros that capture values, and a lot
> of people informally use the term "unhygienic macro" to refer to "macros
> that hygienic macro systems try to prevent", the core thing that's being
> prevented is essentially noncompositionality.
> To turn it around, if I hand you a compositional expression, I also hand
> you the syntax rule and the values of all evaluable sub-expressions,
> then you can 100% reliably hand back the value of the overall
> expression, and this seems to be the core desideratum when people start
> talking about hygienic macro system.
> -- 
> -=-Don Blaheta-=-dblaheta at monm.edu-=-=-<http://www.monmsci.net/~dblaheta/>-=-
> "The "melting pot" theory works in some areas in the larger cities. The
> "salad bowl" theory works rather well for other towns and cities. But I
> propose a third theory that covers vast areas of the US. The "child's plate"
> theory. In this theory all of the foods are separated into their own groups
> and if the ketchup touches the green beans all hell breaks loose."
> 							--Brian Pyle
> _________________________________________________
>  For list-related administrative tasks:
>  http://lists.racket-lang.org/listinfo/users

Posted on the dev mailing list.