On Wed 26 Jan, Peter Hancock wrote:
> It is clear that many programmers have enormous difficulty
> in grasping monadic IO, and carefree pollution of the
> available vocabulary is a serious matter.
Could you explain who you believe is guilty of carefree pollution of
the available vocabulary, and why?
> Jerzy Karczmarczuk's instincts seem to me to be right.
> Evaluation (calculation, ... etc) belongs to the "fetch"
> phase of the IO machine, and execution to, well, the
> "execution" phase.
Which is precisely why describing an expression of type IO a
(evaluated or otherwise) as an "unexecuted" action is appropriate IMHO.
I suppose if want to be really pedantic I could say we don't know for
sure that the action hasn't been executed at some time in past. The
important point (at least as far as Haskell is concerned) is that
representation does not imply execution (I think).
Regards
--
Adrian Hey
