Agreed. Great. I still contend that it would be cool to get this to be a real thing at something like the Haskell workshop, I think hearing the different perspectives would be an interesting insight into the many different ways to explain monads. But I suppose the way to start would be to put up a webpage for collecting them..
kris On Sun, Sep 16, 2012 at 3:55 PM, Conal Elliott <co...@conal.net> wrote: > Hi Tillmann. Wow. Lovely and spot on! And I almost never hear monad > explanations without wincing. Thanks for sharing. -- Conal > > On Sun, Sep 16, 2012 at 7:48 AM, Tillmann Rendel > <ren...@informatik.uni-marburg.de> wrote: >> >> Hi, >> >> >> Kristopher Micinski wrote: >>> >>> Everyone in the Haskell cafe probably has a secret dream to give the >>> best "five minute monad talk." >> >> >> (1) Most programming languages support side effects. There are different >> kinds of side effects such as accessing mutable variables, reading files, >> running in parallel, raising exceptions, nondeterministically returning more >> than one answer, and many more. Most languages have some of these effects >> built into their semantics, and do not support the others at all. >> >> (2) Haskell is pure, so it doesn't support any side effects. Instead, when >> Haskell programmers want to perform a side effect, they explicitly construct >> a description of the side effecting computation as a value. For every group >> of related side effects, there is a Haskell type that describes computations >> that can have that group of side effects. >> >> (3) Some of these types are built in, such as IO for accessing the world >> outside the processor and ST for accessing local mutable variables. Other >> such types are defined in Haskell libraries, such as for computations that >> can fail and for computations that can return multiple answers. Application >> programmers often define their own types for the side effects they need to >> describe, tailoring the language to their needs. >> >> (4) All computation types have a common interface for operations that are >> independent of the exact side effects performed. Some functions work with >> arbitrary computations, just using this interface. For example, we can >> compose a computation with itself in order to run it twice. Such generic >> operations are highly reusable. >> >> (5) The common interface for constructing computations is called "Monad". >> It is inspired by the mathematical theory that some computer scientists use >> when they describe what exactly the semantics of a programming language with >> side effects is. So most other languages support some monad natively without >> the programmer ever noticing, whereas Haskell programmers can choose (and >> even implement) exactly the monads they want. This makes Haskell a very good >> language for side effecting computation. >> >> Tillmann >> >> >> _______________________________________________ >> Haskell-Cafe mailing list >> Haskell-Cafe@haskell.org >> http://www.haskell.org/mailman/listinfo/haskell-cafe > > _______________________________________________ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe
