- Haskell - Haskell.org

RE: [Haskell] GHC / Hugs Disagree on Constraints
by Simon Peyton-Jones 04 Oct '04

04 Oct '04

Hugs does support scoped type variables bound by *type sigs in patterns*, but not by class or instance declarations. So when you say (maxBound :: b) you are really saying (maxBound :: forall b. b) and it's that universal quantification that has no Bounded b context. That's why you need -fglasgow-exts to compile it with GHC, to get the scoped type variables. I've gotten so used to scoped type variables that I couldn't see why it wasn't Haskell 98, until I tried to compile without -fglasgow-exts. Then GHC says Key.hs:99: Could not deduce (Bounded b) from the context () arising from use of `maxBound' at Key.hs:99 Probable fix: Add (Bounded b) to the expected type of an expression When checking the type signature of the expression: maxBound :: forall b. b which does at least point to the error. Simon | -----Original Message----- | From: haskell-bounces(a)haskell.org [mailto:haskell-bounces@haskell.org] On Behalf Of Dominic | Steinitz | Sent: 02 October 2004 12:04 | To: haskell(a)haskell.org | Subject: [Haskell] GHC / Hugs Disagree on Constraints | | GHC accepts this with -fglasgow-exts | | instance (Ord a, Bits a, Bounded a, Integral a, LargeWord a, | Bits b, Bounded b, Integral b, LargeWord b) => | Bounded (LargeKey a b) where | minBound = 0 | maxBound = | fromIntegral $ | (1 + fromIntegral (maxBound::b))* | (1 + fromIntegral (maxBound::a)) - 1 | | Hugs rejects it with +N -98 with | | ERROR "Codec/Encryption/LargeKey.hs":94 - Cannot justify constraints in | type annotation | *** Expression : maxBound | *** Type : b | *** Given context : () | *** Constraints : Bounded b | | Since I've already declared b to be Bounded, it looks like a bug in Hugs. | | Dominic. | | =============================================================== | ======== | | module Codec.Encryption.LargeKey | (Word128,Word192,Word256,LargeWord) where | | import Data.Word | import Data.Bits | import Numeric | import Char | | -- Keys have certain capabilities. | | class LargeWord a where | largeWordToInteger :: a -> Integer | integerToLargeWord :: Integer -> a | largeWordPlus :: a -> a -> a | largeWordAnd :: a -> a -> a | largeWordOr :: a -> a -> a | largeWordShift :: a -> Int -> a | largeWordXor :: a -> a -> a | largeBitSize :: a -> Int | | -- Word64 is a key in the obvious way. | | instance LargeWord Word64 where | largeWordToInteger = toInteger | integerToLargeWord = fromInteger | largeWordPlus = (+) | largeWordAnd = (.&.) | largeWordOr = (.|.) | largeWordShift = shift | largeWordXor = xor | largeBitSize = bitSize | | -- Define larger keys from smaller ones. | | data LargeKey a b = LargeKey a b | deriving (Eq, Ord) | | instance (Ord a, Bits a, LargeWord a, Bits b, LargeWord b) => | LargeWord (LargeKey a b) where | largeWordToInteger (LargeKey lo hi) = | largeWordToInteger lo + (2^(bitSize lo)) * largeWordToInteger hi | integerToLargeWord x = | let (h,l) = x `quotRem` (2^(bitSize lo)) | (lo,hi) = (integerToLargeWord l, integerToLargeWord h) in | LargeKey lo hi | largeWordPlus (LargeKey alo ahi) (LargeKey blo bhi) = | LargeKey lo' hi' where | lo' = alo + blo | hi' = ahi + bhi + if lo' < alo then 1 else 0 | largeWordAnd (LargeKey alo ahi) (LargeKey blo bhi) = | LargeKey lo' hi' where | lo' = alo .&. blo | hi' = ahi .&. bhi | largeWordOr (LargeKey alo ahi) (LargeKey blo bhi) = | LargeKey lo' hi' where | lo' = alo .|. blo | hi' = ahi .|. bhi | largeWordOr (LargeKey alo ahi) (LargeKey blo bhi) = | LargeKey lo' hi' where | lo' = alo .|. blo | hi' = ahi .|. bhi | largeWordXor (LargeKey alo ahi) (LargeKey blo bhi) = | LargeKey lo' hi' where | lo' = alo `xor` blo | hi' = ahi `xor` bhi | largeWordShift w 0 = w | largeWordShift (LargeKey lo hi) x = | if bitSize lo < bitSize hi | then LargeKey (shift lo x) | (shift hi x .|. (shift (conv lo) (x - | (bitSize lo)))) | else LargeKey (shift lo x) | (shift hi x .|. (conv $ shift lo (x - | (bitSize lo)))) | where conv = integerToLargeWord . largeWordToInteger | largeBitSize ~(LargeKey lo hi) = largeBitSize lo + largeBitSize hi | | instance (Ord a, Bits a, LargeWord a, Bits b, LargeWord b) => Show | (LargeKey a b) where | showsPrec p = showInt . largeWordToInteger | | instance (Ord a, Bits a, LargeWord a, Bits b, LargeWord b) => | Num (LargeKey a b) where | (+) = largeWordPlus | fromInteger = integerToLargeWord | | -- Larger keys are instances of Bits provided their constituents are keys. | | instance (Ord a, Bits a, LargeWord a, Bits b, LargeWord b) => | Bits (LargeKey a b) where | (.&.) = largeWordAnd | (.|.) = largeWordOr | xor = largeWordXor | shift = largeWordShift | bitSize = largeBitSize | | instance (Ord a, Bits a, Bounded a, Integral a, LargeWord a, | Bits b, Bounded b, Integral b, LargeWord b) => | Bounded (LargeKey a b) where | minBound = 0 | maxBound = | fromIntegral $ | (1 + fromIntegral (maxBound::b))* | (1 + fromIntegral (maxBound::a)) - 1 | | instance (Ord a, Bits a, LargeWord a, Ord b, Bits b, LargeWord b) => | Integral (LargeKey a b) where | toInteger = largeWordToInteger | | instance (Ord a, Bits a, LargeWord a, Ord b, Bits b, LargeWord b) => | Real (LargeKey a b) | | instance Enum (LargeKey a b) | | type Word96 = LargeKey Word32 Word64 | type Word128 = LargeKey Word64 Word64 | type Word160 = LargeKey Word32 Word128 | type Word192 = LargeKey Word64 Word128 | type Word224 = LargeKey Word32 Word192 | type Word256 = LargeKey Word64 Word192 | _______________________________________________ | Haskell mailing list | Haskell(a)haskell.org | http://www.haskell.org/mailman/listinfo/haskell

2 1

QuickCheck
by Dominic Steinitz 02 Oct '04

02 Oct '04

Can someone remind me how to change the number of tests that QuickCheck performs? I'm sure I knew how to do this but I can't find it in the documentation and a web search revealed no examples. Thanks, Dominic.

1 0

GHC / Hugs Disagree on Constraints
by Dominic Steinitz 02 Oct '04

02 Oct '04

GHC accepts this with -fglasgow-exts instance (Ord a, Bits a, Bounded a, Integral a, LargeWord a, Bits b, Bounded b, Integral b, LargeWord b) => Bounded (LargeKey a b) where minBound = 0 maxBound = fromIntegral $ (1 + fromIntegral (maxBound::b))* (1 + fromIntegral (maxBound::a)) - 1 Hugs rejects it with +N -98 with ERROR "Codec/Encryption/LargeKey.hs":94 - Cannot justify constraints in type annotation *** Expression : maxBound *** Type : b *** Given context : () *** Constraints : Bounded b Since I've already declared b to be Bounded, it looks like a bug in Hugs. Dominic. ======================================================================= module Codec.Encryption.LargeKey (Word128,Word192,Word256,LargeWord) where import Data.Word import Data.Bits import Numeric import Char -- Keys have certain capabilities. class LargeWord a where largeWordToInteger :: a -> Integer integerToLargeWord :: Integer -> a largeWordPlus :: a -> a -> a largeWordAnd :: a -> a -> a largeWordOr :: a -> a -> a largeWordShift :: a -> Int -> a largeWordXor :: a -> a -> a largeBitSize :: a -> Int -- Word64 is a key in the obvious way. instance LargeWord Word64 where largeWordToInteger = toInteger integerToLargeWord = fromInteger largeWordPlus = (+) largeWordAnd = (.&.) largeWordOr = (.|.) largeWordShift = shift largeWordXor = xor largeBitSize = bitSize -- Define larger keys from smaller ones. data LargeKey a b = LargeKey a b deriving (Eq, Ord) instance (Ord a, Bits a, LargeWord a, Bits b, LargeWord b) => LargeWord (LargeKey a b) where largeWordToInteger (LargeKey lo hi) = largeWordToInteger lo + (2^(bitSize lo)) * largeWordToInteger hi integerToLargeWord x = let (h,l) = x `quotRem` (2^(bitSize lo)) (lo,hi) = (integerToLargeWord l, integerToLargeWord h) in LargeKey lo hi largeWordPlus (LargeKey alo ahi) (LargeKey blo bhi) = LargeKey lo' hi' where lo' = alo + blo hi' = ahi + bhi + if lo' < alo then 1 else 0 largeWordAnd (LargeKey alo ahi) (LargeKey blo bhi) = LargeKey lo' hi' where lo' = alo .&. blo hi' = ahi .&. bhi largeWordOr (LargeKey alo ahi) (LargeKey blo bhi) = LargeKey lo' hi' where lo' = alo .|. blo hi' = ahi .|. bhi largeWordOr (LargeKey alo ahi) (LargeKey blo bhi) = LargeKey lo' hi' where lo' = alo .|. blo hi' = ahi .|. bhi largeWordXor (LargeKey alo ahi) (LargeKey blo bhi) = LargeKey lo' hi' where lo' = alo `xor` blo hi' = ahi `xor` bhi largeWordShift w 0 = w largeWordShift (LargeKey lo hi) x = if bitSize lo < bitSize hi then LargeKey (shift lo x) (shift hi x .|. (shift (conv lo) (x - (bitSize lo)))) else LargeKey (shift lo x) (shift hi x .|. (conv $ shift lo (x - (bitSize lo)))) where conv = integerToLargeWord . largeWordToInteger largeBitSize ~(LargeKey lo hi) = largeBitSize lo + largeBitSize hi instance (Ord a, Bits a, LargeWord a, Bits b, LargeWord b) => Show (LargeKey a b) where showsPrec p = showInt . largeWordToInteger instance (Ord a, Bits a, LargeWord a, Bits b, LargeWord b) => Num (LargeKey a b) where (+) = largeWordPlus fromInteger = integerToLargeWord -- Larger keys are instances of Bits provided their constituents are keys. instance (Ord a, Bits a, LargeWord a, Bits b, LargeWord b) => Bits (LargeKey a b) where (.&.) = largeWordAnd (.|.) = largeWordOr xor = largeWordXor shift = largeWordShift bitSize = largeBitSize instance (Ord a, Bits a, Bounded a, Integral a, LargeWord a, Bits b, Bounded b, Integral b, LargeWord b) => Bounded (LargeKey a b) where minBound = 0 maxBound = fromIntegral $ (1 + fromIntegral (maxBound::b))* (1 + fromIntegral (maxBound::a)) - 1 instance (Ord a, Bits a, LargeWord a, Ord b, Bits b, LargeWord b) => Integral (LargeKey a b) where toInteger = largeWordToInteger instance (Ord a, Bits a, LargeWord a, Ord b, Bits b, LargeWord b) => Real (LargeKey a b) instance Enum (LargeKey a b) type Word96 = LargeKey Word32 Word64 type Word128 = LargeKey Word64 Word64 type Word160 = LargeKey Word32 Word128 type Word192 = LargeKey Word64 Word128 type Word224 = LargeKey Word32 Word192 type Word256 = LargeKey Word64 Word192

1 0

Generalised algebraic data types
by Simon Peyton-Jones 01 Oct '04

01 Oct '04

Folks I've finished my first attempt at implementing Generalised Algebraic Data types in GHC. Ordinary algebraic data types are generalised to allow you to write an explicit type signature for each constructor; for example: data Term a where Lit :: Int -> Term Int Succ :: Term Int -> Term Int IsZero :: Term Int -> Term Bool If :: Term Bool -> Term a -> Term a -> Term a Notice that some of these constructors have return types that are not just "Term a"... that's the whole point! Now you can write typed evaluator for these terms: eval :: Term a -> a eval (Lit i) = i eval (Succ t) = 1 + eval t eval (IsZero i) = eval i == 0 eval (If b e1 e2) = if eval b then eval e1 else eval e2 GADTs are not a new idea. They go back to inductive type families in type theory and dependent types, and more recently have been studied in a programming-language context by Xi, Hinze & Cheney, Simonet & Pottier, and Sheard. Indeed, Tim Sheard's language Omega is has taken the same ideas a good deal further. He has a good bunch of links at his home page http://www.cs.pdx.edu/~sheard/ The only really new thing is that it's in GHC. Our paper about the ideas is at http://research.microsoft.com/~simonpj/papers/gadt/index.htm, and it has lots of references to related work. GADTs will be in GHC 6.4, which we'll release in the next month or two, but meanwhile it's in the HEAD of the CVS repository. Anyone who's prepared to build from source, or who can grab a development snapshot from the GHC website, can try it out. I'd like to encourage you to do so, so that you can discover the more egregious bugs before we ship 6.4, rather than after. (Also error messages are likely to be poor at first.) First-cut documentation of the feature is in ghc/docs; a typeset snapshot is at http://research.microsoft.com/~simonpj/tmp/gadt.html, There's a small collection of test programs that exercise GADTs in GHC's test suite: http://cvs.haskell.org/cgi-bin/cvsweb.cgi/fptools/testsuite/tests/ghc-re gress/gadt/ You can use these as a guide to syntax, and a source of programming ideas. You'll find red-black trees (invariants guaranteed by the types), evaluators, type checkers, and more besides. (NB: some are supposed to fail: see all.T. Most are transcribed from Tim's Omega examples.) I'd love to have more GADT tests: send them to me and I'll add them. And let me know how you get on. Simon

1 0

RE: [Haskell] optimisation of State monad
by Simon Peyton-Jones 29 Sep '04

29 Sep '04

I took a quick look. My bet is that profiling is interfering with optimisation (which it certainly does). I also used -G1 to force single-generation GC, which gives more accurate residency numbers. Without profiling, but all compiled with -O. iterR (take 100000 (repeat 1)) (Foo 0 0.0) ./test1 +RTS -K64M -sstderr -G1 Foo {foo = 0, bar = 0.0} 6,923,484 bytes allocated in the heap 3,288 bytes copied during GC 192 bytes maximum residency (18 sample(s)) iterL (take 100000 (repeat 1)) (Foo 0 0.0) ./test2 +RTS -K64M -sstderr -G1 Foo {foo = 0, bar = 0.0} 4,847,872 bytes allocated in the heap 22,032 bytes copied during GC 1,224 bytes maximum residency (18 sample(s)) iterR' (take 100000 (repeat 1)) (Foo 0 0.0) [iterR' is your iterR calling t1'] ./test3 +RTS -K64M -sstderr -G1 Foo {foo = -10, bar = 0.0} 41,046,244 bytes allocated in the heap 29,756 bytes copied during GC 212 bytes maximum residency (148 sample(s)) 148 collections in generation 0 ( 0.18s) These all seem reasonable to me. I didn't look at the state stuff. Simon | -----Original Message----- | From: haskell-bounces(a)haskell.org [mailto:haskell-bounces@haskell.org] On Behalf Of Georg Martius | Sent: 21 September 2004 11:15 | To: haskell | Subject: [Haskell] optimisation of State monad | | Hi folks, | | I have two questions to tail recursion, optimisation(ghc) and the State monad. Sorry about bothering | you with efficiency issues, but they become crusual to me since my programm needs more memory | than I have :-( | | I compiled the following small examples with | ghc -O6 -Wall -prof -auto-all -o test Test.hs (ghc 6.2.1) | and ran them with | ./test +RTS -K64M -sstderr | | > data Foo = Foo { foo :: !Integer , bar :: Double | > } deriving Show | > type Transformation a = a -> a | > | > addX :: Integer -> Transformation Foo | > addX x f = f { foo = (foo f) + x } | >t1 :: Integer -> Transformation Foo | > t1 x = (addX x . addX (-x)) | >iterR :: [Integer] -> Transformation Foo | > iterR list f = foldr t1 f list | > iterL :: [Integer] -> Transformation Foo | > iterL list f = foldl (\ f' i -> t1 i f') f listiterR (take 100000 (repeat 1)) (Foo 0 0.0) | 16,151,796 bytes copied during GC | 11,546,120 bytes maximum residency (5 sample(s)) | Productivity 20.3% of total user, 19.7% of total elapsed | | > iterL (take 100000 (repeat 1)) (Foo 0 0.0) | 3,020 bytes copied during GC | 25,904 bytes maximum residency (1 sample(s)) | Productivity 100.0% of total user, 80.0% of total elapsed | | | Okay, foldl is tail recursive and foldr not. Fine! | Now one weird thing: If I define t1 as: | > t1' :: Integer -> Transformation Foo | > t1' x f = let newfoo = foldr addX f (take 10 $ repeat x) | > in newfoo {foo = foo f - foo newfoo} | | > iterR (take 100000 (repeat 1)) (Foo 0 0.0) | 22,476 bytes copied during GC | 2,113,876 bytes maximum residency (3 sample(s)) | Productivity 60.0% of total user, 59.0% of total elapsed | | Why is the compiler able to optimise the call of t1' and not the one of t1? | | My second question belongs to the State monad: | > withFoo :: Foo -> State Foo (a) -> Foo | > withFoo state monad = execState monad state | >addXM :: Integer -> State Foo () | > addXM x = modify (\ f -> f { foo = (foo f) + x }) | >iterM :: [Integer] -> State Foo () | > iterM list = sequence_ $ map t1M list | >t1M :: Integer -> State Foo () | > t1M x = do addXM x | > addXM (-x) | | > withFoo (Foo 0 0.0) $ do iterM (take 100000 (repeat 1)) | 48,989,348 bytes copied during GC | 14,037,156 bytes maximum residency (7 sample(s)) | Productivity 15.5% of total user, 14.7% of total elapsed | | and if I define t1M as: | t1M' :: Integer -> State Foo () | t1M' x = do old <- get | sequence_ $ map addXM (take 10 $ repeat x) | modify (\ f -> f {foo = foo f - foo old}) | then the memory consumption is awfully high: | | > withFoo (Foo 0 0.0) $ do iterM (take 100000 (repeat 1)) | 172,468,996 bytes copied during GC | 48,522,520 bytes maximum residency (10 sample(s)) | Productivity 2.6% of total user, 2.5% of total elapsed | | Is there a way to optimise the State monad version? | | Any help would be appreciated. | | Georg | _______________________________________________ | Haskell mailing list | Haskell(a)haskell.org | http://www.haskell.org/mailman/listinfo/haskell

2 1

[QCQ] Haskell-QuickCheck Questionnaire
by Koen Claessen 29 Sep '04

29 Sep '04

Dear Haskell-QuickCheck user, When John and I published the first paper about QuickCheck in 2000, we released a Haskell module that was cleary marked as DRAFT. Since then, a lot of things have happened; we and other people have made extensions and improvements to QuickCheck, but only very few of the extensions are part of the "standard" distribution of QuickCheck (included in GHC and Hugs), and none of the improvements. I am right now in the process of gathering all QuickCheck extensions and improvements I am aware of, and packaging them up in a nicely structured and concise way. At the same time, we are planning to write a new tutorial based on the new QuickCheck. In order to do this well, I would like to get your feedback. If you care about the future of QuickCheck, please take a moment to fill in the attached questionnaire. In particular, I would like to know what your typical use of QuickCheck is, so that I can avoid spending time on making things backwards compatible if it is not needed. List of intended includes: - testing monadic code - automatically shrinking counter examples - support for custom counter example callbacks - programmable test coverage warnings - dealing with exceptions - dealing with non-termination - automatic coarbitrary instance - generic predefined simple test data generators - generic default definitions for shrinking Thank you for your help! /Koen ---- QuickCheck Questionnaire ======================== Please fill in your answers like this: [ ] becomes [X] ... becomes <your answer> Multiple answers allowed where applicable. Please send your filled in questionnaires to koen(a)cs.chalmers.se. Please include the tag [QCQ] in the subject of your message. Don't send back your answers later than Friday, October 8th. == General ========================================================= *1. How often do you use / have you used QuickCheck? [ ] Never [ ] A few times [ ] Often [ ] I use it in almost all my projects *2. What Haskell system do you use it with? [ ] GHC [ ] GHCi [ ] Hugs [ ] NHC [ ] Other: ... *3. How often do you use the shell script "quickCheck"? [ ] Never [ ] A few times [ ] Often [ ] I use it in almost all my projects == Original QuickCheck module ===================================== (These are questions about functions that might be dropped from future versions.) *4. How often do you use the function "check" together with your own "Config"? [ ] Never [ ] A few times [ ] Often [ ] I use it in almost all my projects *5. How often do you use the functions "label", "classify", or "trivial"? [ ] Never [ ] A few times [ ] Often [ ] I use it in almost all my projects *5a. If so, which ones: ... *6. How often do you use the functions "two", "three", or "four"? [ ] Never [ ] A few times [ ] Often [ ] I use it in almost all my projects *7. How often do you use the function "vector"? [ ] Never [ ] A few times [ ] Often [ ] I use it in almost all my projects == Other QuickCheck modules ======================================= *8. How often do you use the module Debug.QuickCheck.Batch? [ ] Never [ ] A few times [ ] Often [ ] I use it in almost all my projects *9. How often do you use the module Debug.QuickCheck.Utils? [ ] Never [ ] A few times [ ] Often [ ] I use it in almost all my projects *10. How often do you use the module Debug.QuickCheck.Poly? [ ] Never [ ] A few times [ ] Often [ ] I use it in almost all my projects == Debugging tools ================================================ *11. How often do you use HOOD? [ ] Never [ ] A few times [ ] Often [ ] I use it in almost all my projects *12. How often do you use Hat? [ ] Never [ ] A few times [ ] Often [ ] I use it in almost all my projects *13. How often do you use Buddha? [ ] Never [ ] A few times [ ] Often [ ] I use it in almost all my projects == Suggestions =================================================== *14. Please write down if you find there are functions typically missing from the current QuickCheck distribution: ... *15. Please explain typical problems you have or have had with QuickCheck and did not know how to solve: ... *16. Please suggest any extension that you would like to see in QuickCheck (not already mentioned in the "indeded includes" list in the e-mail above): ... *17. How do you think we should distribute QuickCheck in the future? (As an independent library, in the standard library tree, ...): ... *18. Feel free to write any other comments (positive ones are welcome too :-): ... == The end ====================================================== Thanks! Please send your answers to koen(a)cs.chalmers.se with the tag [QCQ] in the subject, no later than Friday, October 8th.

1 0

Call for Papers: Special Issue M4M
by Carlos Areces 29 Sep '04

29 Sep '04

Call for Papers for "Methods for Modalities" A Special Issue of the Journal of Applied Logic (http://www.sciencedirect.com/science/journal/15708683) The workshop `Methods for Modalities' (M4M) organized every two years aims to bring together researchers interested in developing proof tools and decision methods for modal logic broadly conceived, including description logic, hybrid logic, temporal logic, etc. During 2003 M4M-3, the third instance of the workshop was organized in Nancy, France. We have been invited by the Journal of Applied Logic to prepare a Special Issue containing selected publications from the workshop, in addition to other articles that might fit within the topics of interest of M4M. For that reason we are now distributing this Call for Papers. TOPICS OF INTEREST OF M4M The following list is provided as an example of suitable topics for the Special Issue. All topics should concern modal-like logics, broadly conceived. The list is by no means exhaustive and is given in an arbitrary order: * Automated theorem proving * Decision methods * Proof methods * Standard and non-standard inference problems * Model checking * Testing * Expressive power * Modelling * Applications * Successful uses of modal-like languages in other fields * Knowledge representation SUBMISSION DETAILS The final version of accepted papers has to follow the standard layout guidelines of the Journal of Applied Logic. Instructions for authors and the latex style files are available from http://staff.science.uva.nl/~m4m/SI/ Although not mandatory, we suggest to use the corresponding style files from that web-page also for your initial submission. The issue aims at high quality papers with a length around 20 to max. 30 pages. Please send your submission in postscript or pdf no later than October 31th, 2004 to areces(a)loria.fr. Notice that all papers will be refereed to ensure that their suitability for journal publication. IMPORTANT DATES Deadline for Submissions: October 31th, 2004 Reviews Due : November 21th, 2004 Final Versions Due : December 19th, 2004 GUEST EDITORS of the Special Issue Dr. Carlos Areces INRIA Lorraine 615, rue du Jardin Botanique 54602 Villers les Nancy Cedex, France phone : +33 (0)3 83 58 17 90 fax : +33 (0)3 83 41 30 79 e-mail : carlos.areces(a)loria.fr www : http://www.loria.fr/~areces ----- End forwarded message ----- -- Carlos Eduardo Areces ILLC, University of Amsterdam Nieuwe Achtergracht 166, 1018 WV, Amsterdam, The Netherlands phone : +31 20 525-6730 e-mail : carlos(a)science.uva.nl www : http://www.illc.uva.nl/~carlos visit : http://www.hylo.net -> The Hybrid Logic's Home Page

1 0

hscpp
by Wolfgang Jeltsch 28 Sep '04

28 Sep '04

Hello, there was some discussion about a Haskell-friendly CPP a while ago. What happend to this project? I desperately need such a preprocessor. CPP 3.3.4, which I use now, is not usable for me because you cannot have such a simple thing like a \\ operator with it. How do others solve such problems? I cannot believe that there isn't a better way than living with all those terrible restrictions, using CPP imposes on you. Wolfgang

5 6

Simple question about IO in hugs and ghc
by Vivian McPhail 26 Sep '04

26 Sep '04

Hi, When I run my main function I get an error: The prompt does not wait for user input and thus my parser fails with an empty head. Is there a switch to not echo to the console? Is this the problem? It happens in both Hugs and GHC main :: IO () main = do putStr "sentence? " input <- getLine case input of "quit" -> putStr "bye.\n" _ -> do let res = get_result $ proof $ do_proof $ getInput input in if res == Nothing then putStr ("-> Nothing\n") else putStr ("-> " ++ (show $ unJust res) ++ "\n") main ---- Sample output: START> $ mmccg.exe sentence? Fail: Prelude.head: empty list <END

2 1

10th Est. Winter School in Comp. Sci., 1st call for partic.
by Tarmo Uustalu 25 Sep '04

25 Sep '04

[Lecturers: Danvy, Mehlhorn, Miltersen, Morrisett, Pinkas. Place/time: Palmse, Estonia, 27 Feb-4 March 2005. Deadline for registration and submission of abstracts for student talks/posters: ** 14 Jan 2005 **.] Call for Participation 10th Estonian Winter School in Computer Science EWSCS'05 Palmse, Estonia, 27 February-4 March 2005 http://www.cs.ioc.ee/yik/schools/win2005/ Background and objectives EWSCS is a series of regional-scope international winter schools held annually in Estonia. EWSCS are organized by CIDEC, a joint initiative of Institute of Cybernetics (Tallinn), Tallinn University of Technology and University of Tartu for the advancement of higher education in computer science and information technology. EWSCS'05 is the tenth event of the series. The main objective of EWSCS is to expose Estonian, Baltic, and Nordic graduate students in computer science (but also interested students from elsewhere) to frontline research topics usually not covered within the regular curricula. The subject of the schools is general computer science, with a bias towards theory, this comprising both algorithms, complexity and models of computation, incl. cryptography, and semantics, logic and programming theory. The working language of the schools is English. Programme The schools' scientific programme consists of short courses by renowned specialists and a student session. The course list for EWSCS'05 is the following: Olivier Danvy (University of Aarhus, Denmark: A Vade Mecum to Continuations Kurt Mehlhorn (MPI Saarbrücken, Germany): Implementing Algorithms Peter Bro Miltersen (University of Aarhus, Denmark: Universal Methods for Derandomization Greg Morrisett (Harvard University, Cambridge, MA, USA): Topics in Language-Based Security Benny Pinkas (HP Labs, Haifa, Israel): Privacy-Preserving Data Mining The purpose of the student session is to give students an opportunity to present their work (typically, thesis work) and get feedback. Registrants are invited to propose short talks (20 min) or posters. The selection will be based on abstracts of 150-400 words. The social programme consists of an excursion and a conference dinner. Venue Palmse is a small settlement 80 kms to the east from Tallinn in the county of Lääne-Viru. It is renowned for a large manor that used to belong to the von Pahlen family, today hosting the visitors' center of the Lahemaa National Park, a museum, and a hotel. Tallinn, Estonia's capital, is famous for its picturesque medieval Old Town, a UNESCO World Heritage site. There are direct flights to Tallinn from London, Paris, Milan, Amsterdam, Brussels, Frankfurt, Munich, Hamburg, Berlin, Prague, Warsaw, Moscow, Kiev, Copenhagen, Oslo, Stockholm, Gothenburg, Helsinki, Vilnius, Riga, ferries from Stockholm and Helsinki. From Vilnius, Riga, the Eurolines coach service is the practical travel option. Registration and cost The deadline for registration and submission of abstracts is 14 January 2005. All registrants will be notified of acceptance to school and acceptance of their talks/posters by 28 January 2005. The participation fee 4000 EEK (260 EUR) includes course materials, full boarding at Palmse, transportation from Tallinn to Palmse, excursion and conference dinner. For some students, we can partially waive the fee. Programme committee / organizing committee Tarmo Uustalu (IoC) (chair), Monika Perkmann (IoC) (secretary), Helger Lipmaa (Helsinki U. of Techn.), Peeter Laud (U. of Tartu), Jaan Penjam (IoC), Heli Uibo (U. of Tartu), Jüri Vain (Tallinn U. of Techn.), Varmo Vene (U. of Tartu) Sponsors Centers of Excellence in Research Programme of Ministry of Education and Research of Estonia, FP5 IST project eVikings II Further information Details on the submission of abstracts, registration procedure and cost, application for fee waiver are available from the school webpage, http://www.cs.ioc.ee/yik/schools/win2005/. Questions should be sent to ewscs05(a)cs.ioc.ee.

1 0