The dreaded M-R

Simon Marlow simonmar at microsoft.com
Fri Jan 27 04:51:37 EST 2006


On 26 January 2006 16:07, John Hughes wrote:

> Simon Marlow wrote:
>
>> I wonder if there's an alternative solution along these lines:
>> 
>>  - We use ParialTypeSignatures to make bindings monomorphic:
>> 
>> 
>>
http://haskell.galois.com/cgi-bin/haskell-prime/trac.cgi/wiki/PartialTyp
>> eSigs 
>> 
>>    eg.
>> 
>>      x :: _
>>      x = (+1)
>> 
>>  - we make it a static error for a variable bound by a simple pattern
>>    binding ("x = e") to be overloaded, unless a type signature is
>>    given. The error message would explain the problem, and how to
>>    fix it. Alternatively, we make it a strong warning.
>> 
>> It seems to me that the partial type signatures extension provides a
>> lot of bang for the buck - it gives us a way out of the MR in
>> addition to partial type signatures. 
>> 
> I don't like this. Once students start dropping type signatures (which
> they do
> pretty soon for local variables in where-clauses), they would
> sometimes-- unpredictably as far as they're concerned--get an error
> message telling them they must put one back in again, but it's enough
> to write x :: _. Can 
> you imagine
> explaining to an average student in the first year why they MUST put
> in a type signature, but it doesn't need to include a type???

Understood, but what about when the student writes '=' instead of ':='
by mistake - most of the time it'll work!  But occasionally they fall
foul of the reason we had the MR in the first place.  Presumably the
compiler should emit a noisy warning, but continue anyway?  Won't that
be confusing?

How about an even simpler solution:

  *All* pattern and variable bindings are monomorphic unless a type
  signature is given.

I wonder how much code this would break?  It can't be too bad, because
John is suggesting using := for variable bindings everywhere as a
starting point.  Also, SML users live with something very similar.  How
often do we really need polymorphism in a variable or pattern binding?
I'm guessing probably not that often, because by far the most common use
of polymorphism is in functions.  You lose if you write

  mymap = map

or
  
  mynil = []

and then try to use those at more than one type, but I'm guessing that's
quite rare, and you can still give a type signature.  I'd be very
interested to tweak this in GHC and see how much code still passes the
type checker.

Cheers,
	Simon


More information about the Haskell-prime mailing list