February 2005 - Haskell-Cafe

Getting an attribute of an object
by Dmitri Pissarenko 13 Feb '05

13 Feb '05

Hello! I have a list of instances of the ClassifiedImage class, which is defined as follows. data ClassifiedImage = ClassifiedImage {imageFileName :: String, subjectID :: String} deriving Show Attribute imageFileName contains a file name of a certain image. I want to "transform" the list [ClassifiedImage] into a list [(ClassifiedImage, Image)], where Image is content of the file with name imageFileName. That is, I want to have a routine, which iterates through the list of ClassifiedImages, reads each file (with filename contained in the attribute ClassifiedImage.imageFileName) and stores its content in a variable. I have already a function, which reads the content of a file. My idea is to use map for this task: readClassifiedImages :: [ClassifiedImage] -> [(ClassifiedImage, Image)] readClassifiedImages classifiedImages = do return map readerFunc classifiedImages readerFunc denotes some function, which takes the attribute imageFileName of a ClassifiedImage instance. I suppose that this readerFunc looks like shown below. readerFunc :: ClassifiedImage -> (ClassifiedImage, Image) readerFunc classifiedImage = (classifiedImage, fileContent) where fileName = classifiedImageFileName classifiedImage fileContent = readImage fileName classifiedImageFileName :: ClassifiedImage -> String In order for this function to work, I need to define classifiedImageFileName, which "returns" the imageFileName attribute of an instance of ClassifiedImage class. How can I define this function? Can I use map for readImage function, which is in the IO domain? If not, what tutorial can help me? Thanks in advance Dmitri Pissarenko -- Dmitri Pissarenko Software Engineer http://dapissarenko.com

5 13

Re: Point-free style (Was: Things to avoid)
by oleg＠pobox.com 11 Feb '05

11 Feb '05

Jan-Willem Maessen wrote: ] Is it really clear or obvious what ] ] > map . (+) ] ] means? Perhaps the following two examples might be more convincing: > u=uncurry > e=((partition $ u(==)).) . zip > f x=(x\\).(x\\) It is obviously clear what 'e' and 'f x' do. The second example, which even contains the type signature to increase comprehension: > prod:: (MCompose a b (c -> d) e, MCompose f g (b,g) d) => > (h -> a) -> (c -> f) -> h -> e > > prod = (. ((. (,)) . mcomp)) . mcomp Here `prod' is indeed the categorical product. The second example is taken from http://pobox.com/~oleg/ftp/Haskell/categorical-maxn.lhs which has the following comment about that code fragment: The constraints in the prod's type are intricately related. The final expression for prod bears some similarity with Unlambda code. Perhaps because both Unlambda and the category theory eschew "elements" in favor of combinations of arrows. Probably there are other similarities.

6 5

Re: [Haskell-cafe] Re: Point-free style
by Matthew Roberts 10 Feb '05

10 Feb '05

just a few quick observations comming from an imperative background, I found point free code very difficult to understand when learning Haskell. It is not that it is hard to understand the concepts of point-free style, it is that it is hard to know when something is point-free. It is "another option" and I think the best way to make a language "readable" is to stick to a few simple rules and clean semantics. Since, you can't write all your code point-free, I say write it all pointed - consistency is the thing! also, the term "point free" confused me for quite a while. Point-free code has lots of points (periods) and so I kept thinking it was the other way around. matt On 11/02/2005, at 7:19 AM, Henning Thielemann wrote: > > On 10 Feb 2005, Peter Simons wrote: > >> Now compare that to the following function, which does the >> some thing but without point-free notation: >> >> incrEach' :: Integer -> [Integer] -> [Integer] >> incrEach' i is = is >>= \i' -> return (i'+i) > > point-free again > > is >>= return . (i+) > > :-] > > _______________________________________________ > Haskell-Cafe mailing list > Haskell-Cafe(a)haskell.org > http://www.haskell.org/mailman/listinfo/haskell-cafe

2 1

Re: [Haskell] [newbye] 'Just a'
by Graham Klyne 10 Feb '05

10 Feb '05

[I've switched my response to the Haskell-cafe list] What do you intend if both (get a) and (get b) return Nothing? You might also be interested in the library function 'maybe': http://www.haskell.org/onlinereport/standard-prelude.html#$vmaybe or maybe (sic) Maybe.fromMaybe in: http://www.haskell.org/onlinereport/maybe.html Using one of these, I think your function could reasonably be expressed as a simple one-liner. #g -- At 19:33 04/02/05 +0100, Yuri D'Elia wrote: >Hi all. I'm experimenting with haskell and its type system. >I've done a function which scans a list, and returns "Just a" >value if the element is found, or Nothing. > >get :: a -> Maybe a > >and implemented > >getAorB :: Maybe a -> Maybe a -> a >getAorB a b = ... > >a function which returns "Just (get a)" or "Just (get b)" if (get a) is >Nothing (excluding failures in both). By now, I've implemented it in >terms of pattern matching: > >getAorB a b = f (get a) (get b) > where f (Just a) Nothing = a > f Nothing (Just a) = a > >but I'd like to know if there are other possible ways to do it, possibly >without enforcing an evaluation order like pattern matching does. > >_______________________________________________ >Haskell mailing list >Haskell(a)haskell.org >http://www.haskell.org/mailman/listinfo/haskell ------------ Graham Klyne For email: http://www.ninebynine.org/#Contact

3 3

Contravariance
by Johan Glimming 09 Feb '05

09 Feb '05

Hi I have a problem with contravariance in the bananas in space setting. If you take catamorphism from a function space such as A=A->A to a type such as BxC, then the negative occurrence of X in the functor GXY=X->Y swaps around BxC into B+C. This means that maybe in this particular case Meijer-Hutton's original Haskell code for cata/ana does not work, since the type for cata would force BxC and B+C to be the same type. I might be wrong about this, and I wonder anybody can comment on this. Med vänliga hälsningar / Yours Sincerely, Johan Glimming

2 1

Name My Functor Classes
by Ashley Yakeley 09 Feb '05

09 Feb '05

These classes seem like they might correspond to concepts in category theory: class Functor f where fmap :: (a -> b) -> f a -> f b -- must satisfy: -- fmap id = id -- fmap (p . q) = (fmap p) . (fmap q) class (Functor f) => C1 f where f1 :: f (a -> b) -> f a -> f b class (C1 f) => C2 f where return :: a -> f a -- must satisfy: -- fmap = f1 . return class (C2 f) => Monad f where (>>=) :: f a -> (a -> f b) -> f b -- must satisfy: -- fmap ab fa = fa >>= (return . ab) -- f1 fab fa = fab >>= (\ab -> fmap ab fa) -- (return a) >>= p = p a -- (fa >>= p) >>= q = fa >>= (\a -> p a >>= q) What are good names for C1, C2 and f1? -- Ashley Yakeley, Seattle WA

2 1

Re: Top 20 ``things'' to know in Haskell
by John Hughes 09 Feb '05

09 Feb '05

> >Message: 9 >Date: Mon, 7 Feb 2005 10:31:30 -0500 >From: "Jacques Carette" <carette(a)mcmaster.ca> >Subject: [Haskell-cafe] Top 20 ``things'' to know in Haskell >To: <haskell-cafe(a)haskell.org> >Message-ID: <000701c50d2a$189f5590$1b447182(a)cas.mcmaster.ca> >Content-Type: text/plain; charset="us-ascii" > >The recent post of Graham Klyne (below) reminds me that I have meant to ask: >is there a ``top 20'' things a serious programmer should know when writing >code in Haskell? Of course there is a lot of programming language theory >that would be great to know, but I mean really down-to-earth things like the >2 items below (module Maybe, the 'maybe' function). > >The Haskell libraries are quite large, and it is unrealistic to try to get >familiar with all of them right away. But getting a ``small'' list would be >very useful - I think of this as step 2 after one learns to get comfortable >with a language. I had done this (for Maple) for training new hires at >Maplesoft, and I definitely noticed that they became more idiomatic >programmers faster this way. > >Jacques > >PS: of course, this could already exist on haskell.org and/or the Wiki, but >not in an 'obvious' enough place as I missed it... > > Beginners might find this useful: http://www.cs.chalmers.se/Cs/Grundutb/Kurser/d1pt/d1pta/ListDoc/ It's guide to finding the right list function which leads you to it by asking you what you are trying to achieve. john >-----Original Message----- >From: haskell-cafe-bounces(a)haskell.org >[mailto:haskell-cafe-bounces@haskell.org] On Behalf Of Graham Klyne >Sent: February 7, 2005 10:09 AM >To: Yuri D'Elia; haskell-cafe(a)haskell.org >Subject: [Haskell-cafe] Re: [Haskell] [newbye] 'Just a' > >You might also be interested in the library function 'maybe': > http://www.haskell.org/onlinereport/standard-prelude.html#$vmaybe > >or maybe (sic) Maybe.fromMaybe in: > http://www.haskell.org/onlinereport/maybe.html > >#g >-- > > > > > > >

1 0

Re: [Haskell-cafe] How to write (FxF' . delta) AxB --> F A
by Johan Glimming 08 Feb '05

08 Feb '05

> proj :: (Functor f) => (f:*:f') (a,b) -> f a > proj (Pair x _) = fmap fst x Yes I think that was what I needed. ;-) It seems I also need fsplit :: (Functor f, Functor f') => (f a, f' b) -> (f :*: f') (a,b) for same data (f:*:f') o = Pair (f o) (f' o) Which takes (FA, F'B) into FxF' (A,B). I am confused by the fact of having two instances of F and somehow I need to map on both of them simultaneously. Med vänliga hälsningar / Yours Sincerely, Johan Glimming

1 0

RE: [Haskell-cafe] how to name third party libraries?
by Simon Peyton-Jones 08 Feb '05

08 Feb '05

Send mail to libraries(a)haskell.org, saying what's in your libraries, and suggesting names. The process usually converges pretty rapidly. Simon | -----Original Message----- | From: haskell-cafe-bounces(a)haskell.org [mailto:haskell-cafe-bounces@haskell.org] On Behalf Of S. | Alexander Jacobson | Sent: 07 February 2005 22:55 | To: haskell-cafe(a)haskell.org | Subject: [Haskell-cafe] how to name third party libraries? | | Is there documentation somewhere about how to name | third party libraries? | | For example, if I have HTTP and SMTP libraries, | what should I call them? | | -alex- | | | | ______________________________________________________________ | S. Alexander Jacobson tel:917-770-6565 http://alexjacobson.com | _______________________________________________ | Haskell-Cafe mailing list | Haskell-Cafe(a)haskell.org | http://www.haskell.org/mailman/listinfo/haskell-cafe

1 0

help with some code. it doesn't whant to compile.
by pablo daniel rey 08 Feb '05

08 Feb '05

hello i'm new to haskell so i'm sorry if this is a stupid question, but i'm having problems with some basic code. the code : data Maybe Dir = Just Dir | Nothing data Dir = Left | Right | Up | Down data Piece = Vertical | Horizontal | CodeA | CodeB flow = [(Horizontal, Left, Left), (Horizontal, Right, Right), (Vertical, Down, Down), (Vertical, Up, Up), ................ etc ] fst :: (a,b,c) -> a fst (x,y,z) = x scnd :: (a,b,c) -> b scnd (x,y,z) = y third :: (a,b,c) -> c third (x,y,z) = z element :: [(Piece, Dir, Dir)] -> Maybe Dir element [] = Nothing element xs = Just (third (head xs)) chgDir :: Piece -> Dir -> Maybe Dir chgDir p d = element (filter (\x -> p == (fst x)) (filter (\x -> d == (scnd x)) flow)) the error i get : Instances of (Eq Dir, Eq Piece) required for definition of chgDir i don't know what's happening. help!!! thanks in advance

2 1