Keith Wansbrough did implement his thesis work in a fork of GHC. But (a) it
was jolly complicated and pervasive, and (b) the performance improvements were
not great. It didn't pay its way. So we dropped it. See his thesis, available
here
Dear Gentle Reader,
Many, many beautiful gems in the Haskell Weekly News archives are worth a
second look.
To give you a taste, I reproduce below excerpts from the quotes section of
the Jan 31, 2007 issue -- yes, that's 9 years ago -- under the editorship
of Don Stewart.
Enjoy.
Best, Kim-Ee