January 2013 - Beginners

can't install hackage package
by Emmanuel Touzery 22 Jan '13

22 Jan '13

Hello, I've been using this hackage package "rex" without problems on two computers (linux, windows) but now I fail setting it up on a third computer which confuses me, and the error message is not really enlightening for me: $ cabal update Downloading the latest package list from hackage.haskell.org $ cabal --version cabal-install version 1.16.0.2 using version 1.16.0.3 of the Cabal library $ cabal install rex Resolving dependencies... cabal: Could not resolve dependencies: trying: rex-0.3.1 Any idea? Thank you! Emmanuel

4 11

Antiderivative (indefinite integral)?
by Martin Drautzburg 22 Jan '13

22 Jan '13

Hello all, not strictly a Haskell question, but anyways ... is it possible to compute the antiderivative of a function f::Int->Int ? I understand that you can compute the definite integral by simply summing up the values of f within a given interval. My first guess would be: no this is not possible. The antiderivative F of a function f::Int->Int needs to have the property that F(b) - F(a) must be the sum of f within [a,b]. To do this I must know all values withib [a,b]. But at the time I compute the antiderivative I do not know this interval yet. What is striking me is that in calculus I can often symbolically compute the antiderivative and I get a simple function, and I can get the value of F for a given x and I get a simple number. Why is that so? -- Martin

5 7

Re: [Haskell-beginners] Antiderivative (indefinite integral)?
by Martin Drautzburg 20 Jan '13

20 Jan '13

On Saturday, 19. January 2013 22:38:09 Paul Higham wrote: > Looking at the antiderivative the same way we can write > > s = [Int] -> [Int] > s [] = [] > s (f:fs) = scanl (+) f fs But your function has a well definied starting point. This won't work anymore if the function is defined for all negative Ints, right? -- Martin

1 0

Re: [Haskell-beginners] Antiderivative (indefinite integral)?
by Martin Drautzburg 20 Jan '13

20 Jan '13

On Saturday, 19. January 2013 20:41:32 Denis Kasak wrote: However, if > the same function was given to you as an infinite list of tuples of (x, > f(x)) you would need to do an infinite number of steps just to compute the > antiderivative, much less prove anything about it. An if the function is given as rows in a table (discrete, finite) then I can compute the derivative by just looking at two successive values, however for the antiderivative I have to look at all values between the lowest possible x and the running x. If the function is discrete but has no lower bound for x, then I cannot compute an antiderivative at all, at least not one which will be correct for any x. I wrote an antiderivative function for discrete values and ended up passing it a lower bound for x. IIUC then there is no way to avoid this. Is this about right? It is still somewhat strange. For a discrete function f(i) I can compute the definite integral F(b) - F(a) but I cannot compute F(a) or F(b) themselves. Right? -- Martin

1 0

helper tools for reading haskell source
by Nathan Hüsken 20 Jan '13

20 Jan '13

Hey, Often, when I read haskell code, I would like to know how and where a datatype or function is defined. Normaly I read the code of package, and I then use grep. But when it is not defined in the same package, grep can not help me. Example: I want to know where the datatype Frame and the function windowGetChildren in the wxHaskell odule are defined. I know there is hoogle, but for these it was unable to help me (or I was unable to use it correctly?). So I was wondering, what other tools or methods do you use to "analyse" haskell source? Regards, Nathan

4 3

Things which predictedly change over time
by Martin Drautzburg 19 Jan '13

19 Jan '13

Hello all, ... and I thought this was easy: find a way to represent a "schedule", i.e. a thing which assumes different values at different days. It turned out I don't know what a schedule is. What is the formalism behind "This train leaves at 11:00 every day except sundays (where it doesn't run at all), except on Easter Sunday where it leaves at 11:10"? I do not even need to know the "most compact" representation, just some representation which is reasonably compact and which can be translated into human language. So far I believe a Schedule is a function which takes a day and retuns a Value. If I had this function, I could easily list values for all days within given interval of days. But the inverse should also be possible. Given a List of (Day, Value) pairs, I should be able to construct a Schedule. The reason why I am interested in this is the following: "On every Monday I take a trip from A to C. To do so, I need to take two trains and change trains in B". So I have to consider the two train schedules to decide whether or not my itinerary will work. So I kind of have a Constraint which is scheduled itself (every Monday) and which depends on two other Schedules. The Constraint can be satisfied for some days and violated for others. I want to find out when it is violated (within a finit interval of days) without having to enumerate all days and check each day individually. And most importantly I want to specify the Constraint without referring to individual days, just like I did in the quoted sentence above. You may also say, I want to lift the fine world of relational algebra to a world where things change over time, i.e. assume different values at different days. This seems to be such a common planning problem! You face it every time you deal with things which are not constant over time, and it becomes a nightmare when you deal with relationships between them. Still I could not find anything useful on the net. How can the world turn at all without having this solved? -- Martin

2 2

How to get qualifiers into type
by Rustom Mody 18 Jan '13

18 Jan '13

I am playing around with some small graph theory problems. [Yeah I know there are good libraries -- as I said just playing around...] And I need to do things like type Vertex = Ix a => a (so that a vertex can be used to index an adjacency-list-array) That is to say that whenever 'Vertex' appears in a type signature, the Ix should 'float' out to the qualifier list After a lot of nudging from the error messages and rewriting as type Vertex = forall a. Ix a => a and giving options LANGUAGE RankNTypes, ImpredicativeTypes, LiberalTypeSynonyms I still get all kinds of errors. I'll report them if required. However first of all would like to know: Is this the way to go? Is this possible?

3 3

Type system for constructor preconditions
by Bryan Vicknair 17 Jan '13

17 Jan '13

I'm interested in forcing data passed to a constructor to meet certain preconditions. I've seen the wiki entry on smart constructors, and would have stopped thinking about the problem if hiding the data constructors didn't also make it impossible to do pattern matching. Ideally, the type system could do these checks, but we don't have dependent types so I kept looking. The pattern below uses the type system to force a client of the Circle data type to pass in valid (positive) radius. This isn't quite what is on the smart constructors wiki page. Does pattern have a name? Where can I read more about it? What are the negatives? The positives are that when a client goes to create a Circle in their code, they'll notice that they need to pass a ValidRadius, not just any number, so then they'll *have* to find a function that gives them a ValidRadius, and thus will be forced to be semi-informed about what sorts of preconditions exist on the Circle constructor; They'll have to deal with the case that validRadius gives them Nothing. If we had simply put a comment by the Circle constructor to use a validateRadius function to make sure their number is valid, they could have ignored it. -- In Circle.hs module Circle ( Circle (Circle), Radius, validRadius, ValidRadius ) where type Radius = Int data ValidRadius = ValidRadius Radius deriving (Eq, Show) data Circle = Circle ValidRadius deriving (Eq, Show) -- A radius must be positive. validRadius :: Radius -> Maybe ValidRadius validRadius x | x > 0 = Just $ ValidRadius x | otherwise = Nothing -- In Client.hs import Circle r = 3 :: Radius vr :: ValidRadius vr = case validRadius r of Nothing -> error "Tell UI about bad input" Just x -> x

2 1

How to represent a (running) network?
by Martin Drautzburg 17 Jan '13

17 Jan '13

Hello all, I am trying to implement a discrete event siumulation for a logistic network. To do this, I need a representation of the network as such and a way to express what each Process does when it gets alerted by an event. A Process has inputs and outputs which are again connected to outputs and inputs of other Processes. Establishing these connections constructs the network as such. But the Processes also need to do something. They need to take something from one of their inputs, so something with it and place something on one of their outputs. So I kind of have a dual usage for inputs and outputs: they play a role in the topology of the network and in the "function" which defines a process. How can I keep the two in sync? Is there a way prevent a Process from trying to use an output which does not exist? What would be a good way to represent a Network anyways? I believe the classic approach is a list of nodes and a list vertices. In the simulation I will frequently have to find the process of an input or output and to find the input connected to an output. The node/vertices implementation seems to require scanning lists, which could be slow once I have thousands of processes. Other than that any pointers to how to construct networks (which go beyond mere graphs) would be much appreciated. -- Martin

3 2

Re: [Haskell-beginners] Type Inference
by Johannes Engels 17 Jan '13

17 Jan '13

Am 16.01.2013 23:48, schrieb Lemol Morais Lutonda: > > You realy need a generic indtype? what if to define it just as a Int? > > > 48 Aniversario del Instituto Superior Politecnico Jose Antonio > Echeverria, Cujae > Una obra de la Revolucion Cubana | 2 de diciembre de 1964 | > http://cujae.edu.cu > > > Consulte la enciclopedia colaborativa cubana. http://www.ecured.cu > You are right that I don't need a generic indtype in the instance declarations, the following would be sufficient for me: > instance Sliceable (UArray Int Double) Int Double > where (!) = (AU.!) > instance Sliceable (UArray Int Double) [Int] (UArray Int Double) > where (!) arr ilist = listArray (0,end_ind) [(AU.!) arr i | i <- ilist] > where end_ind = length ilist - 1 But also for this case a type annotation is necessary as in my original post. And within the class definition I think I need a generic indtype, because I want to be able to use arguments of type Int as well as [Int] as the second argument of (!).

1 0