[Haskell-cafe] Question about ST arrays

[Ramin Honary]

(re-sending with "reply to all")

If I understand you correctly, you want to create updates to your hash
function throughout execution of your program. If this is the case, based
on my own experience, I believe the ST monad was not designed for this
purpose and so I think it is probably impossible. You must use the IO monad
for a hash table that receives regular updates during program execution.

ST is designed for doing rapid construction of a large immutable data by
evaluating many stateful updates to structure during the time before it is
made immutable. When monad evaluation terminates, the data structure
becomes immutable. Furthermore, you may not use the immutable data
structure until the ST evaluation terminates, producing the pure value.

For something like a game, you need to take input then update a hash in an
event loop. The IO monad is a good model of this sort of behavior, but the
ST monad is not.

Another possibility is that you can use a pure binary tree structure such
as Data.Map, if you would prefer to maintain purity, and pass this Map
structure around in a State monad transformer. If your State monad
transformer is of type (StateT (Map k v) IO a), you can use "liftIO" to
take inputs, and the "modify" function to update the Map state after each
input.


On Sat, Dec 30, 2017 at 1:05 AM, Jean-Marc Alliot <jm at alliot.org> wrote:

> Well I am going to try to explain why I want to use STArrays and the runST
> function (and I am absolutely ready to be proven wrong, and directed to a
> better way to do it).
>
> I have, as a simple way to learn the language, implemented an Awele
> program in Haskell (I have been programming games for years). It was easy,
> elegant (around 70 lines, 150 with the graphical interface) and completely
> functional, but I soon stumbled upon one problem. While implementing a
> vanilla alpha-beta is easy in functional style, implementing hash-tables (I
> mean hash-tables of the kind we use in game programming, not exactly
> regular hash-tables) is not that easy, and hash tables are mandatory to
> have a fast alpha-beta.
>
> Currently, my idea is to create a module which will hide the
> implementation, and have only two functions in its interface:
> 1) A test_and_retrieve which take as only parameter a position and return
> a Maybe object which contains Nothing if the position has never been
> evaluated or (Just v) if there is already an evaluation for it (I know that
> I need more information than just a value, but for the sake of simplicity
> let's stick with just a value)
> 2) A store function which will take a position and its evaluation and
> store it in the table.
>
> Now if I use IO Arrays, I will have to live in the IO Monad if I
> understand correctly how the IO Monad works, while using ST Arrays seems to
> give me the possibility to hide all the non functional code in my hash
> module, and keep my main code almost purely functional.
>
>
> Le 29/12/2017 à 16:36, Ramin Honary a écrit :
>
> You should use "runSTArray" or "runSTUArray" instead of "runST" to convert
> your STArray to an immutable array:  https://hackage.
> haskell.org/package/array-0.5.2.0/docs/Data-Array-ST.html#v:runSTUArray
>
> Or another option is to use "stToIO"  to convert the "STArray" to an
> "IOArray." https://hackage.haskell.org/package/base-4.10.
> 1.0/docs/Control-Monad-ST-Lazy.html#v:stToIO  but if you want to build an
> IOArray, it is better to just start with an IOArray rather than converting
> an STArray to an IOArray.
>
> The design of ST arrays is to allow you to construct them quickly and then
> make them immutable once you are done constructing it.
>
> An immutable array must have the whole array copied after every single
> update, but ST arrays allow you to make many updates without copying, then
> when you have completed constructing the ST array, you must freeze it to an
> immutable array using the "runSTUArray" function. Freezing happens without
> copying the array, after that it is immutable and may not be made into an
> STArray again unless you unfreeze it using "thaw", which creates a copy of
> it:  https://hackage.haskell.org/package/array-0.5.2.0/
> docs/Data-Array-MArray.html#v:thaw
>
> Once you have constructed your immutable array, you can access it
> arbitrarily using the immutable operator (!).
>
> If you want to make multiple updates at multiple times, you must use an
> IOArray or IOUArray. The ST monad is designed for you to construct pure,
> referentially transparent, immutable values in an isolated and strictly
> evaluated "environment" that lets you perform strict updates during
> construction. Once evaluation of the "ST" monad is complete, the returned
> value becomes pure, immutable, and referentially transparent. The for-all'd
> "s" parameter of the "runST" function ensures you do not mix separate
> "environments," and this is the reason you got your type error. Using
> "runSTArray" or "runSTUArray" does not have this restriction.
>
> I am not sure of the reason for this design decision, but I know it has
> something to do with the compiler guaranteeing that pure immutable
> referentially transparent types are constructed without effecting each
> other, preventing race conditions. There is discussion of this on the
> Haskell wiki:  https://wiki.haskell.org/Monad/ST#An_explanation_
> in_Haskell-Cafe
>
>
> On Fri, Dec 29, 2017 at 11:36 PM, Jean-Marc Alliot <jm at alliot.org> wrote:
>
>> Hi,
>>
>> This is my first post to this list so I apologize in advance if I don't
>> use it properly, or if my question is too simple or inapropriate.
>>
>> I come from the Caml world and I am quite new to Haskell (but not to
>> functional programming). I am currently trying to get the hang of Haskell
>> arrays. I have gone through regular arrays, IO Arrays and I am now working
>> with ST Arrays.
>>
>> This is the problem I am currently stuck with. I write the following code:
>>
>> arr = newArray (-1, 1) 0 :: ST s (STArray s Int Int)
>> get :: Int -> Int
>> get i = runST (arr >>= (\b -> readArray b i))
>>
>> Here everything is perfectly OK.
>>
>> Now I want a more general version that could deal with any array like
>> arr. So I write:
>>
>> get2 :: ST s (STArray s Int Int) -> Int -> Int
>> get2 tab i = runST (tab >>= (\b -> readArray b i))
>>
>> And the compiler is clearly very upset by my code:
>>
>> Couldn't match type ‘s’ with ‘s1’
>>       ‘s’ is a rigid type variable bound by
>>         the type signature for:
>>           get2 :: forall s. ST s (STArray s Int Int) -> Int -> Int
>>         at testst.hs:17:9
>>       ‘s1’ is a rigid type variable bound by
>>         a type expected by the context:
>>           forall s1. ST s1 Int
>>         at testst.hs:18:14
>>       Expected type: ST s1 Int
>>         Actual type: ST s Int
>> I am pretty sure that the compiler is right and I am wrong, but I don't
>> get why... Anyone could help?
>>
>> Thanks
>>
>>
>> _______________________________________________
>> Haskell-Cafe mailing list
>> To (un)subscribe, modify options or view archives go to:
>> http://mail.haskell.org/cgi-bin/mailman/listinfo/haskell-cafe
>> Only members subscribed via the mailman list are allowed to post.
>
>
>
>
> _______________________________________________
> Haskell-Cafe mailing list
> To (un)subscribe, modify options or view archives go to:
> http://mail.haskell.org/cgi-bin/mailman/listinfo/haskell-cafe
> Only members subscribed via the mailman list are allowed to post.
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.haskell.org/pipermail/haskell-cafe/attachments/20171230/267b4f0b/attachment.html>