[Haskell-cafe] Type Directed Name Resolution

[Stephen Tetley]

Is it just me or does this bit in the proposal:

  m .lookup key
    .snd
    .reverse

Which translates to this:

 reverse . snd . (\m -> lookup m key) $ m

make no sense and refuse to type check - i.e lookup is producing a
Maybe not a pair for second?

I can see some benefit with TDNR for record selection as per Neil
Brown's message, but I have serious doubts about the syntax and find
most of the examples in proposal determinedly non-compelling.