- Haskell-Cafe - Haskell.org

Typeclasses vs simple functions?
by Olex P 15 Sep '09

15 Sep '09

Hey guys, It's a dumb question but I'd like to know a right answer... Let's say we have some geometry data that can be Sphere, Cylinder, Circle and so on. We can implement it as new data type plus a bunch of functions that work on this data: data Geometry = Sphere Position Radius | Cylinder Position Radius Height | Circle Position Radius deriving (Show) perimeter (Sphere _ r) = 0.0 perimeter (Cylinder _ r h) = 0.0 perimeter (Circle _ r) = 2.0 * pi * r Perimeter doesn't make sense for Sphere or Cylinder. So we could define a type class for objects that have perimeter and make an instance of it only for Circle (data Circle = Circle Position Radius). Make sense. But these three functions above have desired behaviour. If user has a list of objects like [Sphere, Circle, Circle, Cylinder] he would like to calculate perimeters of each object using map perimerer list (in this case we also have to modify Geometry data type). So we could make instances of "perimeter" type class for all objects and return zero in case if perimeter doesn't make sense. Same as previous version but with typeclasses and with additional constructors (constructors for each type of object + constructors in Geometry data). Looks a bit overcomplicated. Any reasons to use type classes in this case? Maybe there is something I'm missing? Cheers, -O

3 6

Re: [Haskell-cafe] Typeclasses vs simple functions?
by Olex P 15 Sep '09

15 Sep '09

Sure! I completely forgot about Maybe. The only one question is is it good from the point of view of ordinary user who doesn't know about such things like functional programming, monads etc. Imagine average user who is looking into a spreadsheet and sees values 0.1, 1.4, Nothing... From other side it seems to be logical. Why not. Thanks for the idea :) On Tue, Sep 15, 2009 at 12:05 PM, Lyndon Maydwell <maydwell(a)gmail.com>wrote: > I think the problem is that you want to compose a list with no > indication of weather one member can have a perimeter or not. I'm not > sure if this is a good solution or not, but I immediately think to > make all Geometry objects instances of a class that return a Maybe > value for the perimeter: > > e.g. > > --- > > import Data.Maybe > > data Geometry = Sphere Position Radius | Circle Position Radius deriving > (Show) > > type Position = (Double, Double) > type Radius = Double > type Height = Double > > class Encircled x where > perimeter :: x -> Maybe Double > > instance Encircled Geometry where > perimeter (Sphere _ r) = Nothing > perimeter (Circle _ r) = Just $ 2.0 * pi * r > > list = [Sphere (1,1) 1, Circle (2,2) 2] > > main = (print . catMaybes . map perimeter) list > > --- [12.566370614359172] > > On Tue, Sep 15, 2009 at 6:29 PM, Olex P <hoknamahn(a)gmail.com> wrote: > > Hey guys, > > > > It's a dumb question but I'd like to know a right answer... > > Let's say we have some geometry data that can be Sphere, Cylinder, Circle > > and so on. We can implement it as new data type plus a bunch of functions > > that work on this data: > > > > data Geometry = Sphere Position Radius > > | Cylinder Position Radius Height > > | Circle Position Radius > > deriving (Show) > > > > perimeter (Sphere _ r) = 0.0 > > perimeter (Cylinder _ r h) = 0.0 > > perimeter (Circle _ r) = 2.0 * pi * r > > > > Perimeter doesn't make sense for Sphere or Cylinder. So we could define a > > type class for objects that have perimeter and make an instance of it > only > > for Circle (data Circle = Circle Position Radius). Make sense. But these > > three functions above have desired behaviour. If user has a list of > objects > > like [Sphere, Circle, Circle, Cylinder] he would like to calculate > > perimeters of each object using map perimerer list (in this case we also > > have to modify Geometry data type). > > So we could make instances of "perimeter" type class for all objects and > > return zero in case if perimeter doesn't make sense. > > Same as previous version but with typeclasses and with additional > > constructors (constructors for each type of object + constructors in > > Geometry data). Looks a bit overcomplicated. > > Any reasons to use type classes in this case? Maybe there is something > I'm > > missing? > > > > Cheers, > > -O > > > > _______________________________________________ > > Haskell-Cafe mailing list > > Haskell-Cafe(a)haskell.org > > http://www.haskell.org/mailman/listinfo/haskell-cafe > > > > >

2 2

Typeclasses vs simple functions?
by Lyndon Maydwell 15 Sep '09

15 Sep '09

---------- Forwarded message ---------- From: Lyndon Maydwell <maydwell(a)gmail.com> Date: Tue, Sep 15, 2009 at 10:03 PM Subject: Re: [Haskell-cafe] Typeclasses vs simple functions? To: Olex P <hoknamahn(a)gmail.com> I think it depends on what is going to be using the functions, data. As far as I can tell, type classes seem to be used in code designed for reuse. So if it is just for a small part of a project that is hidden behind an interface or something, then it's probably fine to just use data/functions without type classes. Keep in mind that these are the musings of someone who only recently began using Haskell. If someone else has a better explanation, then I'd be interested as well :) On Tue, Sep 15, 2009 at 9:57 PM, Olex P <hoknamahn(a)gmail.com> wrote: > Well... How this: > > instance Encircled Geometry where > perimeter (Sphere _ r) = Nothing > perimeter (Circle _ r) = Just $ 2.0 * pi * r > > differs from this: > > perimeter :: Geometry -> Maybe Double > perimeter (Sphere _ r) = Nothing > perimeter (Circle _ r) = Just $ 2.0 * pi * r > > and from this: > > perimeter :: Geometry -> Double > perimeter (Sphere _ r) = 0.0 > perimeter (Circle _ r) = 2.0 * pi * r > > The latter is even simpler because there is no need in extraction of Double > value from Maybe. > So the question is still there: do I need a type class? > > On Tue, Sep 15, 2009 at 12:21 PM, Olex P <hoknamahn(a)gmail.com> wrote: >> >> Sure! I completely forgot about Maybe. The only one question is is it good >> from the point of view of ordinary user who doesn't know about such things >> like functional programming, monads etc. Imagine average user who is looking >> into a spreadsheet and sees values 0.1, 1.4, Nothing... From other side it >> seems to be logical. Why not. >> Thanks for the idea :) >> >> On Tue, Sep 15, 2009 at 12:05 PM, Lyndon Maydwell <maydwell(a)gmail.com> >> wrote: >>> >>> I think the problem is that you want to compose a list with no >>> indication of weather one member can have a perimeter or not. I'm not >>> sure if this is a good solution or not, but I immediately think to >>> make all Geometry objects instances of a class that return a Maybe >>> value for the perimeter: >>> >>> e.g. >>> >>> --- >>> >>> import Data.Maybe >>> >>> data Geometry = Sphere Position Radius | Circle Position Radius deriving >>> (Show) >>> >>> type Position = (Double, Double) >>> type Radius = Double >>> type Height = Double >>> >>> class Encircled x where >>> perimeter :: x -> Maybe Double >>> >>> instance Encircled Geometry where >>> perimeter (Sphere _ r) = Nothing >>> perimeter (Circle _ r) = Just $ 2.0 * pi * r >>> >>> list = [Sphere (1,1) 1, Circle (2,2) 2] >>> >>> main = (print . catMaybes . map perimeter) list >>> >>> --- [12.566370614359172] >>> >>> On Tue, Sep 15, 2009 at 6:29 PM, Olex P <hoknamahn(a)gmail.com> wrote: >>> > Hey guys, >>> > >>> > It's a dumb question but I'd like to know a right answer... >>> > Let's say we have some geometry data that can be Sphere, Cylinder, >>> > Circle >>> > and so on. We can implement it as new data type plus a bunch of >>> > functions >>> > that work on this data: >>> > >>> > data Geometry = Sphere Position Radius >>> > | Cylinder Position Radius Height >>> > | Circle Position Radius >>> > deriving (Show) >>> > >>> > perimeter (Sphere _ r) = 0.0 >>> > perimeter (Cylinder _ r h) = 0.0 >>> > perimeter (Circle _ r) = 2.0 * pi * r >>> > >>> > Perimeter doesn't make sense for Sphere or Cylinder. So we could define >>> > a >>> > type class for objects that have perimeter and make an instance of it >>> > only >>> > for Circle (data Circle = Circle Position Radius). Make sense. But >>> > these >>> > three functions above have desired behaviour. If user has a list of >>> > objects >>> > like [Sphere, Circle, Circle, Cylinder] he would like to calculate >>> > perimeters of each object using map perimerer list (in this case we >>> > also >>> > have to modify Geometry data type). >>> > So we could make instances of "perimeter" type class for all objects >>> > and >>> > return zero in case if perimeter doesn't make sense. >>> > Same as previous version but with typeclasses and with additional >>> > constructors (constructors for each type of object + constructors in >>> > Geometry data). Looks a bit overcomplicated. >>> > Any reasons to use type classes in this case? Maybe there is something >>> > I'm >>> > missing? >>> > >>> > Cheers, >>> > -O >>> > >>> > _______________________________________________ >>> > Haskell-Cafe mailing list >>> > Haskell-Cafe(a)haskell.org >>> > http://www.haskell.org/mailman/listinfo/haskell-cafe >>> > >>> > >> > > > _______________________________________________ > Haskell-Cafe mailing list > Haskell-Cafe(a)haskell.org > http://www.haskell.org/mailman/listinfo/haskell-cafe > >

1 0

ANNOUNCE: LambdaINet-0.1.0, Graphical Interaction Net Evaluator for Optimal Evaluation
by Paul L 15 Sep '09

15 Sep '09

It's available on Hackage DB at http://hackage.haskell.org/package/LambdaINet Thanks to Kim-Ee Yeoh for pushing me into releasing this piece of code I wrote two years ago. I'll just quote the README from the source tarball below. LambdaINet ========== LambdaINet implements an interaction net based optimal evaluator following Lambdascope [1], with an interactive graphical interface allowing user to view and directly manipulate interaction net. [1] Vincent van Oostrom, Kees-Jan van de Looij, Marijn Zwitserlood, Lambdascope, Workshop on Algebra and Logic on Programming Systems (ALPS), Kyoto, April 10th 2004 USAGE ===== After "cabal install", just type "LambdaINet" to start the application. Once it starts, press H for help, and ESC to quit. To understand all the operations in detail, you'll have to read the above mentioned paper by Oostrom et. al. Currently there is no way to load input programs except modifying the source, Try src/Main.lhs if you want to change the start-up program, or any of the 1..9 preset programs. At this moment, the object language supports lambda expression with recursion, tuples, and primitives such as numbers, strings and functions. SIDE NOTE ========= The bulk of code was put together in two weeks when I was working on the leak problem for FRP in 2007. So it was really a rushed job with no guarantee of correctness, although I tried to stay faithful to the original paper as much as I could. I only did some moderate clean-ups before releasing this application to public, and the code itself was sparingly documented when it was originally written. Some parts are probably still buggy, like the translation from net to term; other parts could use more improvements, like the node positioning and line layout algorithm. But I decide to release it anyway -- maybe some people some where will find it useful. DEVELOPMENT =========== Please forward bug reports or feedbacks to me (Paul Liu at paul at thev dot net), but don't hold your hope high on timely bug fixes. Help is also needed to develop LambdaINet further, for example, it really needs a way to read lambda expressions from a separate file or standard input, which should be a simple feature to add, but alas! I don't have the time in the nearest future to do this kind of things myself. -- Regards, Paul Liu Yale Haskell Group http://www.haskell.org/yale

2 3

HDBC Oracle bindings(?/!)
by Thiago Arrais 15 Sep '09

15 Sep '09

Are there Oracle bindings for HDBC anywhere? The HDBC project page talks about Oracle backends but there isn't an Oracle-specific driver listed (I assume one can reach an Oracle backend through the ODBC interface?). I started writing Oracle bindings for HDBC based on the Takusen code. It was done mostly for playing than for anything else and certainly isn't production ready, but it is available in a darcs repo at http://thiagoarrais.com/repos/hdbc-oracle in case anyone is interested. This code isn't ready for prime time yet and I'm looking for some that *is*. Is there such a thing already? Cheers, Thiago Arrais -- Enjoy Chaos - http://blog.thiagoarrais.com Thoughts, ideas and useless ranting about software development and technology

4 6

Cabal install on Windows 7
by Peter Verswyvelen 15 Sep '09

15 Sep '09

I tried the "cabal install" command on Windows 7, and I had to run it with administrative privileges, otherwise I got access denied (it failed to create the Haskell folder in C:\Program Files) Not sure if this is also the case on Vista. Is this the intended behavior?

10 27

Can't install chp (confused by cabal yet again)
by Colin Paul Adams 14 Sep '09

14 Sep '09

I tried a cabal install chp: It complained that base was hidden. So I unpacked the archive, and tried installing using runhaskell Setup configure/build/install. Now I get (from install): Setup: You need to re-run the 'configure' command. The version of Cabal being used has changed (was Cabal-1.6.0.3, now Cabal-1.6.0.2). So I repeated the process and get the same message again. ghc version is 6.10.3 Also: cabal --version cabal-install version 0.6.2 using version 1.6.0.2 of the Cabal library Where does this 1-6.0.3 come from (Ghc HEAD perhaps?)? What can I do about it? -- Colin Adams Preston Lancashire

3 2

trouble compiling Crypto-4.2.0 / trouble with cabal
by mf-hcafe-15c311f0c＠etc-network.de 14 Sep '09

14 Sep '09

Hi, Cabal is still fighting me all the time. Its latest move is to be oblivious of some of the installed packages: $ cabal unpack crypto Unpacking Crypto-4.2.0... $ cd Crypto-4.2.0/ $ runghc ./Setup.hs configure --prefix=/tmp2/ Configuring Crypto-4.2.0... Setup.hs: At least the following dependencies are missing: HUnit -any, QuickCheck -any $ ghc-pkg list | grep -i -e '$cabal\|hunit$' Cabal-1.6.0.3, array-0.2.0.0, base-3.0.3.1, base-4.1.0.0, HTTP-4000.0.8, HUnit-1.2.2.0, ObjectName-1.0.0.0, OpenGL-2.3.0.0, My ghc version is 6.10.4.20090719. (Perhaps I should go get myself a more stable version?) How does Cabal find its packages other than by ghc-pkg? I read the documentation, and did a quick search, but came up empty-handed. thanks, matthias

3 3

ANNOUNCE: hssqlppp, sql parser and type checker, pre-alpha
by Jake Wheat 14 Sep '09

14 Sep '09

Hello all, I've started on a SQL parser and type checker, which I'm currently planning on evolving into a lint-type program for PL/pgSQL, called HsSqlPpp. It currently parses a subset of PostGreSQL SQL and PL/pgSQL, can type check some select, insert, update, delete and create statements, and is pretty rough all round at the moment. It uses Parsec and UUAGC, comes with a small HUnit test suite, and has a Cabal file. The code is on Launchpad here: https://launchpad.net/hssqlppp You can get a snapshot here: http://launchpad.net/hssqlppp/prealpha/secondtypechecking/+download/hssqlpp… More information in the readme: http://bazaar.launchpad.net/~jakewheat/hssqlppp/trunk/annotate/head%3A/READ… and brief the usage guide gives you an idea of what you can do with it right now: http://bazaar.launchpad.net/~jakewheat/hssqlppp/trunk/annotate/head%3A/usage I'm pretty new to Haskell, parsing, attribute grammars and type checking - any comments, advice, criticism welcome. Thanks, Jake Wheat

3 3

Coercion from a Word32/64 to a Float/Double
by minh thu 14 Sep '09

14 Sep '09

Hi, I'd like to know if this should work: -- GHC coercion getFloat :: BHeader -> Get Float getFloat h = case endianness h of LittleEndian -> fmap (coerce . fromIntegral) getWord32le BigEndian -> fmap (coerce . fromIntegral) getWord32be where coerce (I32# x) = F# (unsafeCoerce# x) -- GHC coercion getDouble :: BHeader -> Get Double getDouble h = case endianness h of LittleEndian -> fmap (coerce . fromIntegral) getWord64le BigEndian -> fmap (coerce . fromIntegral) getWord64be where coerce (I64# x) = D# (unsafeCoerce# x) Loading this into ghci compiles fine (but I haven't try to parse data), but compiling with ghc (6.10.1 or 6.10.4) reports: /tmp/ghc1967_0/ghc1967_0.s: Assembler messages: /tmp/ghc1967_0/ghc1967_0.s:287:0: Error: bad register name `%fake0' /tmp/ghc1967_0/ghc1967_0.s:349:0: Error: bad register name `%fake0' Thanks for any help, Thu

3 4