Just going back to this, the channel issue may be solved by the
strict-concurrency package (strict Chans and MVars), and the general
problem of distributing arrays seems to be solved more thoroughly
by the data parallel array library (map, fold, scanl, filter, zip et
al), not just map and reduce?
Hello Haskell-Cafe,
i've written small program which demonstrates how map/reduce may be
implemented in Haskell. it counts amount of words in file, splitting
it into 64kb blocks processed by two threads. their results are
combined by another two threads. how it may be made better? in
particular, is