🚀 go-pugleaf

In <BvHIwz.9HM@comp.vuw.ac.nz> brian@comp.vuw.ac.nz (Brian Boutel) writes:
>In article <farrell.718005712@coral>, farrell@coral.cs.jcu.edu.au (John Farrell) writes:
|>However I don't like the ML/Haskell convention of using
|>"let x = y in z" rather than "z where x = y", I like to read programs top down
|> rather than bottom up.

>Consider 
>   let x = y in if a then b else c
>and
>   if a then b else let x = y in c

>Now consider
>    if a then b else c where x = y
>To which of the let forms does this correspond?

>It's ambiguous, and an arbitrary decision has to be made as to which parse to
>use. This is dangerous, as a wrong assumption can result in a syntactically
>correct program with entirely the wrong meaning.

  To me it seems that "let" is a command to the reader, and by the time I have
informed myself that x = y, I have forgotten why I was reading the code in the
first place. I think "let" sounds more like an imperative command than a
functional expression, e.g. "LET" is BASIC syntax.
  Anyway, while I'm at it, I don't like the "a | b = c" syntax either, but I
suspect that's because I haven't learnt to read "|" as being "if" and "=" as
being "then". I much prefer some sort of arrow notation, but I can't think of
a good one right now.
  As people who don't know me might have guessed, I'm a Miranda weenie and I
love Miranda syntax (BTW, if David Turner is reading this, I am moving to Gofer
because it is free and is available on the PC). Now continuing my love of
ambiguity, I would like to say that I don't care if a language is ambiguous,
just so long as I know what it means :-). I.e. I would much rather write "where"
clauses with brackets to resolve ambiguity than write "let" expressions which
confuse me straight away.
  Which reminds me! Why the difference between "let" and "letrec"? Surely we
have got past the stage of explicity flagging recursive functions? I know that
(well, I believe that) an expression can have different meanings depending on
whether a token is a "let" or a "letrec", but surely it's dangerous to have a
name refer to two different values depending on three little letters?

John