[mercury-users] discriminated union question

[Ralph Becket]

> "Ralph Becket" <rbeck at microsoft.com> wrote:
> 	I've been meaning to change the definition for pair to 
> use {}/2 tuples
> 	and replace -/2 over pairs with a function with the 
> appropriate set of
> 	modes.  pair/2 does crop up in several places in the 
> library, but I'd
> 	encourage people to move to the new tuple types where possible.
> 	
> Erlang uses {Key,Value} for pairs rather than Prolog's Key-Value.
> It's probably the single UGLIEST thing in Erlang.
>
Well this one's open to debate.  Last time I suggested carrying out
the change the only people who expressed an opinion supported the 
idea.

> An important advantage of the (-)/2 pair constructor is that it is
> *different* from generic tuples.  It would be a great pity to 
> lose that.

It's not clear to me what the advantage is.  Perhaps it would be
sensible
to use -/2 to denote key-value pairs, but in that case I think the pair 
type should be renamed as such.

The change I was suggesting would still support -/2 for pairing and
pattern matching, the only difference being that pairs would actually
be {}/2s and we could take back fst/1 and snd/1 for tuples.

-- Ralph
--------------------------------------------------------------------------
mercury-users mailing list
post:  mercury-users at cs.mu.oz.au
administrative address: owner-mercury-users at cs.mu.oz.au
unsubscribe: Address: mercury-users-request at cs.mu.oz.au Message: unsubscribe
subscribe:   Address: mercury-users-request at cs.mu.oz.au Message: subscribe
--------------------------------------------------------------------------