Paul writes
| It has occurred to me that unlifted tuples achieved via a special
| "newtype" decl are not the same as those achieved with strictness
| annotations. This is because with "newtype" it seems that people want
| a situation where (_|_,_|_) = _|_. But with strictness annotations on
| b
Regarding Lifted vs. Unlifted Tuples
I pretty strongly agree with the general oberservation that "tuples
for free" is a Good Idea, whether it be motivated from ADT's (as Simon
first did), from function results (as Joe did), or as function
arguments/currying (
Regarding Lifted vs. Unlifted Function Spaces
-
Gee, maybe lifted functions aren't such a good idea after all! (:-)
But before reaching that conclusion, there are a few points that need
clarification. In particular, Simon mentions two problems that
In pub/haskell/glasgow/working on ftp.dcs.glasgow.ac.uk:
-rw-rw-r-- 7588053 Nov 5 20:27 ghc-0.18-bin-sun4.tar.gz
-rw-rw-r-- 141709 Nov 6 11:17 ghc-0.18-hi-files-ghc-0.16.tar.gz
-rw-rw-r-- 101853 Nov 6 11:16 ghc-0.18-hi-files-hbc.tar.gz
-rw-rw-r-- 3028811 Nov 5 19:40 ghc-0.18-sr
