Hi, I had a monadic revelation at about 3 am. The answer to the question "what is an IO value, really?" is "who cares?" I just posted a blog entry discussing how CT found it's way from Moggi into Haskell at http://syntax.wikidot.com/blog (hence the title; Moggi as functor). It addresses the question of what such things are and why Moggi's insight is so brilliant. Feedback welcome, but please remember this is coming from a non-mathematician who likes to write. If you find anything there that outrages your inner Russell, please correct me, but be gentle. Thanks, gregg
Correction: the correct response is: "Nothing."
On Sat, Feb 7, 2009 at 11:11 AM, Gregg Reynolds
Hi,
I had a monadic revelation at about 3 am. The answer to the question "what is an IO value, really?" is "who cares?" I just posted a blog entry discussing how CT found it's way from Moggi into Haskell at http://syntax.wikidot.com/blog (hence the title; Moggi as functor). It addresses the question of what such things are and why Moggi's insight is so brilliant. Feedback welcome, but please remember this is coming from a non-mathematician who likes to write. If you find anything there that outrages your inner Russell, please correct me, but be gentle.
Thanks,
gregg
On Saturday 07 February 2009 12:11:29 pm Gregg Reynolds wrote:
I had a monadic revelation at about 3 am. The answer to the question "what is an IO value, really?" is "who cares?" I just posted a blog entry discussing how CT found it's way from Moggi into Haskell at http://syntax.wikidot.com/blog (hence the title; Moggi as functor). It addresses the question of what such things are and why Moggi's insight is so brilliant. Feedback welcome, but please remember this is coming from a non-mathematician who likes to write. If you find anything there that outrages your inner Russell, please correct me, but be gentle.
As far as I know, Moggi didn't really have anything directly to do with Haskell. He pioneered the idea of monads being useful in denotational semantics. But it was Wadler that recognized that they'd be useful for actually writing functional programs (see his "The Essence of Functional Programming"). So one might say that it was his doing that brought monads to Haskell proper. -- Dan
Hi Dan,
On Sat, Feb 7, 2009 at 12:00 PM, Dan Doel
On Saturday 07 February 2009 12:11:29 pm Gregg Reynolds wrote:
As far as I know, Moggi didn't really have anything directly to do with Haskell. He pioneered the idea of monads being useful in denotational semantics. But it was Wadler that recognized that they'd be useful for actually writing functional programs (see his "The Essence of Functional Programming"). So one might say that it was his doing that brought monads to Haskell proper.
From what I've read Wadler was clearly the guy who thought of using monads in Haskell, but he explicitly credits Moggi for coming up with the general idea. Moggi just as clearly knew he was on to something powerful and useful (e.g. something " that could lead to the introduction of higher order modules in programming languages like ADA or ML"). What I would be interested in knowing is whether it was Wadler or Moggi who first realized monads (and CT) could be encoded directly in a target language, not just in a semantic metalanguage. Plus there were other people working in the same area; I just don't know the detailed history. Might be a good subject for a blog post for somebody who does.
FYI I made a few corrections to my original post. -gregg
participants (2)
-
Dan Doel -
Gregg Reynolds