In the author's own words: I should stress that this is very hand-waving imprecise definition...
I don't want to be too hard on this guy, but this is a bad article. Sorry. This article uses incorrect terms to define things. It also doesn't explain or illustrate why monads are useful. You cannot read this article and learn anything about monads.
It's not useful on a theoretics level, and it's also not useful on a practical Haskell level. I don't think the author understands what he is writing about. (Don't get me wrong, by the way – it's totally fine to write about something you don't understand in order to learn more about it. But maybe don't publish it prominently on your website and peddle it out to newbies as learning material.)
The overabundance of bad monad articles is making it harder for Haskell newbies to learn. There are many good, free resources for learning about Haskell, such as Real World Haskell. This is not one of them.
Think of a monad as a spacesuite full of nuclear waste in the ocean next to a container of apples. now, you can't put oranges in the space suite or the nucelar waste falls in the ocean, BUT the apples are carried around anyway, and you just take what you need. - Dons, talking about bad monad explanations
edit: also, if you're really interested in seeing what a Haskell monad looks like literally translated into JavaScript, I wrote a comment here on HN a year ago with some code: http://news.ycombinator.com/item?id=1275860 I could leave better comments if I wrote it again today, but the code is correct and works.
Hmm! That's funny, I thought I was "getting" monads better after reading this article (as someone who doesn't already have a firm grasp of monads, of course).
You see, I already sensed that monads:
- were a way that you could sort of simulate side effects using pure functions
- had to do with the type signatures of functions
What I gleaned from the article is that monads:
- have to do with making functions composable
- work by way of the abstractions bind and unit in order to make type signatures match
Did the article harm my (lack of) understanding of monads?
Monads basically let you add an extra calculation to your functions. Why would you want to do that? Mainly so that you don't have to do it manually. Think of it as tacking an extra calculation onto the side of each function that operates in the monad.
For a "state" monad, that calculation is maintaining/passing the state. For a "logging" monad, that calculation is maintaining the log. For an "error handling monad", that calculation is checking for and propagating errors.
I know lisp better than Haskell, and so I'm more familiar with macros than with monads. From what you describe, it seems to me that the use cases for monads and macros have some overlap.
Take the OP's example for the Writer monad, in which several functions needed to return a debug string "<function name> was called" in addition to their main return value. If I had these two functions+:
(def sine (x)
[(Math.sin x) "sine was called"])
(def cube (x)
[(* x x x) "cube was called"])
I might write a macro like the following (rather than a monad) to abstract away their common pattern:
And then each function's definition could be written more concisely:
(defWriter sine (x)
(Math.sin x))
(defWriter cube (x)
(* x x x))
---
+ Forgive me for indulging in the syntactic sugar from my own lisp->javascript project in these examples, but since we're dealing in JavaScript I couldn't resist. https://github.com/evanrmurphy/lava-script
Actually, bind and unit seem to fit very well with this interesting article, which explains the mathematics behind monads (in fact it uses unit, which you say has nothing to do with monads): http://bartoszmilewski.wordpress.com/2011/01/09/monads-for-t...
If you read the article, it's clearly not intended to be a "resource for learning about Haskell". It's a gentle introduction to monads for people who don't want to read Haskell code.
I don't want to be too hard on this guy, but this is a bad article. Sorry. This article uses incorrect terms to define things. It also doesn't explain or illustrate why monads are useful. You cannot read this article and learn anything about monads.
It's not useful on a theoretics level, and it's also not useful on a practical Haskell level. I don't think the author understands what he is writing about. (Don't get me wrong, by the way – it's totally fine to write about something you don't understand in order to learn more about it. But maybe don't publish it prominently on your website and peddle it out to newbies as learning material.)
The overabundance of bad monad articles is making it harder for Haskell newbies to learn. There are many good, free resources for learning about Haskell, such as Real World Haskell. This is not one of them.
Think of a monad as a spacesuite full of nuclear waste in the ocean next to a container of apples. now, you can't put oranges in the space suite or the nucelar waste falls in the ocean, BUT the apples are carried around anyway, and you just take what you need. - Dons, talking about bad monad explanations
edit: also, if you're really interested in seeing what a Haskell monad looks like literally translated into JavaScript, I wrote a comment here on HN a year ago with some code: http://news.ycombinator.com/item?id=1275860 I could leave better comments if I wrote it again today, but the code is correct and works.