[Haskell-cafe] Type-Directed Name Resolution for Records?

[John Smith]

Two significant points which have emerged from the TDNR thread are:

-The wiki page combines two orthogonal proposals: type-directed name resolution, which requires no special syntax, and 
the x.f syntax, which does not require TDNR.
-Implementing both proposals may be a desired fix for the record system, but their use in the rest of the language is 
more controversial.

I would therefore like to propose that TDNR (both parts thereof) be implemented for record accessors. In the fullness of 
time, the community may come to some consensus regarding the implementation of either or both of these features in the 
rest of the language.