[erlang-questions] per function/process locals

[ok]

On 1 May 2007, at 3:35 am, James Hague basically suggested
  - the long-known technique of mapping an imperative style (without
    _shared_ mutable subvariables) to a functional style by setting up
    a one-to-one correspondence between mutable variables and function
    arguments

  - using some kind of preprocessor or language extension to make it
    easier to get this kind of code right

Similar extensions to Prolog have often been suggested.

Speaking only for myself, I find that it is a rare algorithm where
I need more than a handful of "variables" in a loop, and I would
regard the presence of a sufficient number to be painful as indicative
of bad design on my part rather than inadequacy on Erlang's (or  
Prolog's)
part.  However, see Mercury, where they've provided imperative-looking
record update.

I'd rather have the let-for-then construct.