- Haskell-Cafe - Haskell.org

no dynamic binding
by Andrew Harris 21 Sep '04

21 Sep '04

Hi - I have another question. I am still working on a soccer server and thought it would be neat to create command objects that had a "toString" method. Then, I was going to keep a list of these command objects and at the right time stringify them and send them to the server. So I created a class with a toString method: class ServerCommandClass a where toString :: a -> String And then a few instances: -- dash command data DashCommand = DashCommand { dashpower :: Double } instance ServerCommandClass DashCommand where toString c = "(dash " ++ show (dashpower c) ++ ")\n" -- move command data MoveCommand = MoveCommand { x :: Double, y :: Double } instance ServerCommandClass MoveCommand where toString c = "(move " ++ show (x c) ++ " " ++ show (y c) ++ ")\n" The problem is, I am not quite sure how to describe a *list* of command objects where the list could have both DashCommands and MoveCommands in it. Ideally the list could contain both, and then for each item in the list I could call the toString method. I was reading Simon Thompson's Haskell: The Craft of Functional Programming and I read that Haskell 98 does not support dynamic binding, which (it seems) is what I'm trying to do. Does anyone have a suggestion on an alternative approach? thanks, -andrew

4 3

Re: [Haskell-cafe] Strings - why [Char] is not nice
by MR K P SCHUPKE 20 Sep '04

20 Sep '04

Would it not bet better is String were a typeclass instead of a data type. Then the string operations could be redefined by implementation, so you may have strings as lists or strings as arrays? Keean.

2 1

Re: [Haskell-cafe] no dynamic binding
by Abraham Egnor 20 Sep '04

20 Sep '04

You can use exisential types to do what you'd like. From memory, so there are probably errors: newtype ServerCommand = forall a. ServerCommandClass a => ServerCommand a instance ServerCommandClass ServerCommand where toString (ServerCommand c) = toString c commands :: [ServerCommand] commands = [ServerCommand $ DashCommand ..., ServerCommand $ MoveCommand ...] Also, for a less type-safe approach, see Data.Dynamic. On Sun, 19 Sep 2004 13:48:53 -0400, Andrew Harris <andrew.unit(a)gmail.com> wrote: > Hi - > > I have another question. I am still working on a soccer server and > thought it would be neat to create command objects that had a > "toString" method. Then, I was going to keep a list of these command > objects and at the right time stringify them and send them to the > server. So I created a class with a toString method: > > class ServerCommandClass a where > toString :: a -> String > > And then a few instances: > > -- dash command > data DashCommand = > DashCommand { dashpower :: Double } > > instance ServerCommandClass DashCommand where > toString c = "(dash " ++ show (dashpower c) ++ ")\n" > > -- move command > data MoveCommand = > MoveCommand { x :: Double, > y :: Double } > > instance ServerCommandClass MoveCommand where > toString c = "(move " ++ show (x c) ++ " " ++ show (y c) ++ ")\n" > > The problem is, I am not quite sure how to describe a *list* of > command objects where the list could have both DashCommands and > MoveCommands in it. Ideally the list could contain both, and then for > each item in the list I could call the toString method. > > I was reading Simon Thompson's Haskell: The Craft of Functional > Programming and I read that Haskell 98 does not support dynamic > binding, which (it seems) is what I'm trying to do. Does anyone have > a suggestion on an alternative approach? > > thanks, > -andrew > _______________________________________________ > Haskell-Cafe mailing list > Haskell-Cafe(a)haskell.org > http://www.haskell.org/mailman/listinfo/haskell-cafe >

2 1

monad question
by Andrew Harris 20 Sep '04

20 Sep '04

Hi everyone, Well, I'm trying to build a soccer server client and am using Haskell. I thought I was muddling through well enough... but I have some questions. I am using the Robot monad that I copied from Hudak's book HSOE from page 272 (assuming I would understand it later). Here's the monad code I copied: newtype Robot a = Robot (RobotState -> IO (RobotState, a)) instance Monad Robot where return a = Robot (\s -> return (s, a)) Robot sf0 >>= f = Robot (\s0 -> do (s1, a1) <- sf0 s0 let Robot sf1 = f a1 (s2, a2) <- sf1 s1 return (s2, a2)) updateState :: (RobotState -> RobotState) -> Robot () updateState u = Robot (\s -> return (u s, ())) queryState :: (RobotState -> a) -> Robot a queryState q = Robot (\s -> return (s, q s)) I defined myself a comfortable "chunk of state" (comfortable for today at least): data RobotState = RobotState { player_type :: RobotType, team_name :: String, my_current_position :: Position, visible_flags :: [ObjInfo], visible_ball :: [ObjInfo], visible_friends :: [ObjInfo], visible_foes :: [ObjInfo], audible_player_messages :: [String], my_current_mode :: RobotMode, debug_level :: [RobotDebugLevel], my_uniform_number :: Integer, game_mode :: GameMode, game_clock :: Integer } deriving Show In addition, I added some support functions similar to the ones he wrote in Section 19.3.1: assign_flags :: [ObjInfo] -> Robot () assign_flags flaglist = updateState (\s -> s {visible_flags = flaglist}) assign_friends :: [ObjInfo] -> Robot () assign_friends friendlist = updateState (\s -> s {visible_friends = friendlist}) assign_foes :: [ObjInfo] -> Robot () assign_foes foelist = updateState (\s -> s {visible_foes = foelist}) assign_ball :: [ObjInfo] -> Robot () assign_ball balllist = updateState (\s -> s {visible_ball= balllist}) Now, I read several online monad tutorials and kind of understand the notion of a thread of function calls. And I know that the do notation is syntactic sugar for the bind operation. But the code I've written is really nasty. For example, here is part of my code that handles a see record and returns a new RobotState: handleSeeRecord :: [SeeObjInfo_type] -> RobotState -> IO (RobotState, ()) handleSeeRecord seeobjlist p = do flaglist <- return (morphToList flagFinder seeobjlist) balllist <- return (morphToList ballFinder seeobjlist) friendlist <- return (morphToList friendFinder seeobjlist) foelist <- return (morphToList foeFinder seeobjlist) Robot e <- return (assign_flags (flagSpread flaglist)) Robot f <- return (assign_ball balllist) Robot g <- return (assign_friends friendlist) Robot h <- return (assign_foes foelist) (r', ()) <- e p (r'', ()) <- f r' (r''', ()) <- g r'' h r''' This actually works, but I think it can be improved with the bind operation >>= somehow. I tried a few alternatives but couldn't find one that worked. I ended up with this because it seems I had to "strip" the type off of "return (assign_flags (flagSpread flaglist))" and get the "e" that was a function. Might someone have a suggestion on how to improve this? thanks -andrew

2 1

Writing binary files?
by Ron de Bruijn 18 Sep '04

18 Sep '04

Hi, I would like to write and read binary files in Haskell. I saw the System.IO module, but you need a (Ptr a) value for using that, and I don't need positions. I only want to read a complete binary file and write another binary file. In 2001 somebody else came up with the same subject, but then there wasn't a real solution. Now, 3 years later, I can imagine there's *something*. What's that *something*? Regards, Ron __________________________________ Do you Yahoo!? New and Improved Yahoo! Mail - Send 10MB messages! http://promotions.yahoo.com/new_mail

18 52

Interoperability with other languages and haskell in industry
by Vincenzo aka Nick Name 18 Sep '04

18 Sep '04

Again, I will try to take benefit of the thread on the "senior" list to ask a question to everybody who uses haskell in industry (so you people at Galois Connection can't avoid to answer, I know you are there :D ): are your solutions entierely written in haskell, or are there parts written in other languages? If so, how do you let all parts interoperate? Do you use some form of RPC, or CORBA, do you just use a database to store common data, do you use custom protocols (e.g. command line arguments) or what? Do you have experience with wrong ways to achieve this goal? I ask this because "it might be" that in the next years our sleeping company will produce some software, and I can easily convince other people to use "new" languages, if I can ensure them that in case it proves difficult for any reason, we can finish with a certain module and implement the rest of the system using "more conventional" technologies. Thanks Vincenzo

2 2

Proofs for program testing
by Henning Thielemann 18 Sep '04

18 Sep '04

I saw that many introductions to Haskell contain proofs of properties of implemented functions. This is probably due to the fact that pure functions can be handled more easily than imperative programs with hidden states. I wondered whether one can use proofs of Haskell functions for testing. I found QuickCheck http://www.cs.chalmers.se/~rjmh/QuickCheck/ which, as far as I understand, relies on random inputs and I found http://homepages.inf.ed.ac.uk/wadler/realworld/era.html which sounds like some GUI driven program. Is there something that can be used for automatical testing, e.g. for darcs' tests to check patch integrity?

1 0

RE: [Haskell-cafe] Re: Writing binary files?
by MR K P SCHUPKE 17 Sep '04

17 Sep '04

>You wouldn't want to have to accumulate the >entire body as a single byte string Ever heard of lazyness? Haskell does it quite well... Accumulating the entire body doesn't really do this because haskell is lazy. You don't need a more complex interface in Haskell! Keean.

2 1

RE: [Haskell-cafe] Re: Writing binary files?
by MR K P SCHUPKE 17 Sep '04

17 Sep '04

>In the general case, it needs to be a bit more complex than that, Thats why the functions handled lists not individual characters, I was assuming that each [Word8] -> [Char] represented a valid and complete encoding block... IE at the start of each call it assumes no escapes. All this means is than when reading in chunks you paste those chunks together before conversion, and you can only break outside of escapes. This in my opinion is better behaviour anyway... I don't want some hidden escape state mangling output, just because some earler code generated invalid output. Keean.

2 1

ghc-6-2-1_1.msi MD5 sum
by Sergey Zaharchenko 17 Sep '04

17 Sep '04

Could anyone give me the MD5 sum of http://www.haskell.org/ghc/dist/6.2.1/ghc-6-2-1_1.msi I've been writing in Haskell for FreeBSD, and now I wanted to compile one under windows. The .msi file seems to be broken, and I just wanted to check... -- DoubleF "The more data I punch in this card, the lighter it becomes, and the lower the mailing cost." -- Stan Kelly-Bootle, "The Devil's DP Dictionary"

1 0