Hello all,
In response to my post on coercions in Haskell several people told
me: Don't do it; Haskell's type system is complicated enough as it
is, and coercions give a lot of trouble without many advantages.
Still, I wonder if the properties of type synonyms could be
described in terms of coercions. One could consider a type T
introduced by a type synonym to be distinct from its defining type
U. To compensate for this, if an expression of type U occurs
where one of type T is expected, or vice versa, then silently a
coercion is inserted. As far as I can see, the resulting
behaviour is identical to the current Haskell "type synonym and
its definition are completely interchangeble" rule.
If I understand `newtype' declarations correctly, viewed in this
light they are similar to `type' declarations, except that the
programmer is forced to name the coercion from the old to the new
type (as a constructor), and write it down explicitly.
So, three questions:
- Is the above view of type synonyms really equivalent to the
Haskell report's definition?
- If so, wouldn't it be a better alternative which allows the
introduction of more general coercion functions later?
- Can it be used to replace or simplify the notion of `newtype'?
(Why were `newtype' declarations introduced in the first place?)
Groetjes,
<><
Marnix
--
Marnix Klooster
[EMAIL PROTECTED] // [EMAIL PROTECTED]
