The Future of Haskell discussion at the Haskell Workshop

Robert Ennals Robert.Ennals@cl.cam.ac.uk
Wed, 10 Sep 2003 11:39:39 +0100


> ketil@ii.uib.no (Ketil Z. Malde) writes:
> 
> > Robert Ennals <Robert.Ennals@cl.cam.ac.uk> writes:
> 
> BTW, isn't this more or less exactly what Simon suggested (at the very
> top of this thread)?

Not really, no.

I assume you mean the system suggested by Peter Thieman, outlined in the 
initial email by Henrik Nilsson.


My system has the following differences:


Record updaters become normal functions. (and recold selectors remain 
functions)

Normal type classes are used to implement them. 
	(no magic "has" constraints)

Type classes are not magically inferred. They are manually declared just like
	any other type class would be.



As I interpret it, the system proposed at the top of the thread treats record 
fields as something special, and I am very keen that this should not happen. I 
think that it is important that one should be able to replace a record field 
with accessor functions.


-Rob