August 2016 - Haskell-Cafe

Comparable Haskell text(s) to "ML for the Working Programmer"?
by Lawrence Bottorff 06 Aug '16

06 Aug '16

Is there a book or books that cover as much, as well as the Lawrence Paulson "ML for the Working Programmer (2nd ed)" for Haskell? What I'd like in Haskell form is pretty much exactly what Paulson covers in his brilliant book. LB

3 2

a wiki in yesod package
by Ruben Astudillo 06 Aug '16

06 Aug '16

Hi all I've been looking for a package or helpful material to do a wiki based on yesod. I seen the example on the book[1] and thought maybe somebody published something. by the way: which is the mailing list of yesod? the haskell-web is really silent. [1]: http://www.yesodweb.com/book/wiki-chat-example -- -- Ruben Astudillo

2 1

Design Question - Functions taking 'subtype' like arguments but has to be restricted to one type
by Gurudev Devanla 06 Aug '16

06 Aug '16

Hello All, I have a design question where I could use some of your thoughts and suggestions. Here is the setup. I apologize for the long email but I have resisted asking this question, since it needs a long explanation, until now. I have a set of data structures defined as follows -- An Item that can be of two types, one whose value can be changed, one whose value are frozen once created data Item = FreeToChange {freeToChangeCount:: Int} | CannotChange {frozenCount:: Int} -- The item is part of a basket data Basket = Basket { name:: String, item::Item } -- The cart can have both kind of Baskets at the start of the program, or during runtime. data Cart = List Basket You can imagine this be a shopping cart with fixed set of items. Where the count of some of the items in the basket can be changed during shopping but not the count of the items once they are tagged as frozen. Therefore, valid operation are: 1. I can create an Basket with either FreeToChange item or CannotChange item. 2. I can update the count for FreeToChange item in the Basket 3. But, once I create an instance of the Basket to contain the CannotChange item, we cannot update the count anymore or update the Basket. One approach I have taken is to throw an error if this happens by pattern matching on type. But, this is an runtime check. addToBasket :: Basket -> Basket addToBasket b = let i = item b i' = case i of FreeToChange f -> FreeToChange {freeToChangeCount = f + 1} CannotChange f -> error ("This operation is not allowed") in b {item=i'} Here are my questions: 1. Is there a way to design the above data structures in such a way I could use the type system. 2. Since, these are runtime changes, is it even a good design pattern to push this responsibility to a type system? Thanks Guru

7 11

Suggestions for an empirical master thesis
by Jon Kristensen 06 Aug '16

06 Aug '16

Hi, everyone! I'm looking for a master thesis topic that is empirical in nature (like, statistics, hypothesis testing, etc.). The work could involve analyzing either package metadata (Cabal information), code (AST), and/or data from some other source, possibly comparing similar data from some non-Haskell domain. As an example, one idea that was suggested to me was to look at usage aggregation, or like, how much of a given package another package is actually using (which could be relevant in any orphan-instance discussion). From an academic standpoint, it would be good to pick a metric that can be validated in some way. Thank you! Best, Jon

3 3

Views of data types
by Corentin Dupont 04 Aug '16

04 Aug '16

Hi guys, I have a typeclass like that: class Signal s where type SignalData s :: * The idea is that when some signal "s" fires, it returns a data of type (SignalData s). This is all fine except when I instantiate the class with: data Input a = Input [(a, String)] instance Signal (Input a) type SignalData (Input a) = a The concrete type of Input a is not known, so its difficult to store in lists, to serialize... So I need to create a "view" for the it: class (Signal s, Signal v) => SignalView s v where view :: s -> v data InputView = InputView [(Int, String)] instance Signal InputView type SignalData InputView = Int instance SignalView (Input a) InputView view (Input as) = InputView (zip [0…] (snd <$> as)) Is this a common pattern? I'm not sure I get it right...

2 1

[ANN] Compose :: Melbourne 2016 Speakers!
by Lyndon Maydwell 04 Aug '16

04 Aug '16

Hello Everyone! In case you haven't heard, Compose :: Melbourne (the new sister-conference of Compose) has now announced its program of presentations for the first day: http://www.composeconference.org/2016-melbourne/day-one-program/ Compose is a conference for functional programmers, focused on technologies such as Haskell, Scala, Clojure, OCaml, F#, SML and others. Compose :: Melbourne will run on the 29th and 30th of August (this month!) We're excited to have a wide range of talks on a diverse selection of topics including an opening keynote by Manuel M.T. Chakravarty (the author of Haskell for Mac). There will be a single track of 10 speakers from Melbourne, Australia, and around the world! ## Unconference and Workshops The second day of the conference will take the form of an unconference. The day will also feature a collection of organised, as well as spontaneous workshops including introductions to Haskell and Purescript: http://www.composeconference.org/2016-melbourne/unconference/ ## Financial Assistance If you need help attending Compose :: Melbourne, then contact us. We have put in place a financial-assistance plan: http://www.composeconference.org/2016-melbourne/assistance/ ## Volunteering As Compose :: Melbourne will be running in just a few short weeks, we're currently looking for volunteers to help assist with running the workshops on day-two. If you feel like lending a hand, then don't hesitate to get in touch! ## Sponsorship If you, or your organisation wishes to lend assistance to Compose :: Melbourne, we are accepting sponsorship. If you want to learn more, then please view our sponsorship-prospectus: https://github.com/composeconference/Compose-Melbourne/wiki/Sponsorship-Tie… ## Getting in Touch To get in touch with the organisers of Compose :: Melbourne, then email the committee at <composemel-admin(a)googlegroups.com>. For more information, see the home-page of Compose Conference: http://www.composeconference.org/

1 0

memory usage: general questions
by Christopher Howard 03 Aug '16

03 Aug '16

Hi. I've spent the last few days reading about strictness, non-strictness, laziness, thunks, and persistence, and to be honest I'm still in a bit of a fog. A few questions, which I hope might help me out: Supposing I have an expression f (f (f b))) where f :: B -> B and B is some ginormous persistent data structure. and f is an "updating" function outputting some modified copy version of it's argument So, a few questions: 1) Supposing that f does not short-circuit: Is it correct to say that it is definitely advantageous (from a memory usage perspective) for f to be a strict function? Or is this heavily dependent on whether or not b contains unevaluated data structures (e.g., a list)? 2) As far as the overall peak size of our memory graph: How will it be affected by the size of b, if f is non-strict? and if f is strict? 3) Are the "overwritten" portions of each B (i.e., those parts of the persistent data we no are no longer needed) guaranteed to be pushed out of memory at any particular point? Is this something we know will (might?) happen as soon as the GC is run? Or does it happen right when each f is applied to it's argument? -- https://qlfiles.net To protect my privacy, please use PGP encryption. It's free and easy to use! My public key ID is 0x340EA95A (pgp.mit.edu)

3 3

ANN: unicode-transforms-0.1.0.1 (unicode normalization)
by Harendra Kumar 03 Aug '16

03 Aug '16

This is for those who want unicode normalization but do not want a dependency on the heavyweight icu libraries. It has the same API as the text-icu package. It is based on the utf8proc C implementation. https://hackage.haskell.org/package/unicode-transforms https://github.com/harendra-kumar/unicode-transforms A pure Haskell implementation is a work in progress, it works for the NFD normalization form as of now. I could make it perform better than the current utf8proc C implementation and it even beats the ICU implementation in one of the benchmarks. I intend to finish it when I find time to work on it again. Contributions are welcome! -harendra

1 0

[Job] Formal Tech is hiring
by Iavor Diatchki 03 Aug '16

03 Aug '16

Hello, Formal Tech, a Galois spin-off, is looking for a mid to senior level engineer to work on CyberChaff ( https://formal.tech/products/cyberchaff/) Applicants should be proficient in Haskell and at least one other programming language. This is a full time position with benefits. If you are interested, please send your resume to careers(a)formal.tech Cheers, -Iavor

1 0

Measuring allocations with 'weigh' - what about thunks?
by Mitchell Rosen 03 Aug '16

03 Aug '16

weigh (https://www.fpcomplete.com/blog/2016/05/weigh-package) is a library measures how much allocation evaluating a function causes. It's basically a nice interface to GHC.Stats.getGCStats. Here is an example that measures *\n -> n+1*: import Weigh main :: IO () main = mainWith (func "+1" plusOne 5) plusOne :: Int -> Int plusOne = \n -> n + 1 which prints Case Bytes GCs Check +1 16 0 OK when run (64 bit machine; allocating 1 Int = 16 bytes). So my question is, what about the thunk allocated by *(+)*? My rough understanding of the heap is this: Code Heap ---- ---- 1 1 1 1 1 plusOne 1 +-------+----------+----+ +----+---+ | THUNK | reserved | --|--> | I# | 5 | +-------+----------+----+ +----+---+ 1 1 1 1 1 deepseq it +-------+----------+----+ +----+---+ | IND | | --|--> | I# | 5 | +-------+-----|----+----+ +----+---+ | 1 1 `------------> +----+---+ | I# | 6 | +----+---+ 1 1 GC +----+---+ | I# | 6 | +----+---+ So haven't we allocated 5 bytes, not two? Three for the GC'd thunk, and two for the new Int? Any clarification or pointers (heh) to some GHC reading would be greatly appreciated! Thanks, Mitchell

1 0