- Haskell - Haskell.org

RE: strictness question
by Simon Peyton-Jones 02 Mar '01

02 Mar '01

Strange. You don't supply a complete program, so it's hard to test. Nevertheless, the Haskell Report (Sect 3.12) specifies that a let adds a single twiddle. Thus let (x, (y,z)) = e in b means let x = case e of (x,(y,z)) -> x y = case e of (x,(y,z)) -> y z = case e of (x,(y,z)) -> z in b And that is what GHC implements. You get something different if you add twiddles inside: let (x, ~(y,z)) = e in b means let x = case e of (x,_) -> x y = case e of (_,(y,_)) -> y etc Adding more twiddles means less eager matching. I don't know whether Hugs implements this. Simon | -----Original Message----- | From: S. Doaitse Swierstra [mailto:doaitse@cs.uu.nl] | Sent: 01 March 2001 11:26 | To: haskell(a)haskell.org | Subject: strictness question | | | I ran into a difference between GHC and Hugs. The following code: | | f (P p) ~(P q) = P (\ k -> \inp -> let (((pv, (qv, r)), m), st) = | p (q k) inp | in (((pv qv , r ), m), st)) | | runs fine with Hugs but blows up with GHC, whereas: | | f (P p) ~(P q) = P (\ k -> \inp -> let ~(~(~(pv, ~(qv, r)), m), | st) = p (q k) inp | in (((pv qv , r ), m), st)) | | runs fine with GHC too. | | From the Haskell manual I understand that pattern matching | in "let"'s | should be done lazily, so the addition of a collection of ~'s should | not make a difference. Am I right with this interpretation? | | A possible source of this problem may be origination from the smarter | GHC optimiser, but in that case the optimiser is not doing its work | well. | | Doaitse Swierstra | | | | | -- | ______________________________________________________________ | ____________ | S. Doaitse Swierstra, Department of Computer Science, Utrecht | University | P.O.Box 80.089, 3508 TB UTRECHT, the | Netherlands | Mail: mailto:doaitse@cs.uu.nl | WWW: http://www.cs.uu.nl/ | PGP Public Key: http://www.cs.uu.nl/people/doaitse/ tel: +31 (30) 253 3962, fax: +31 (30) 2513791 __________________________________________________________________________ _______________________________________________ Haskell mailing list Haskell(a)haskell.org http://www.haskell.org/mailman/listinfo/haskell

1 0

WRS'2001 - Extended deadline
by Salvador Lucas Alba 02 Mar '01

02 Mar '01

[Apologies for multiple copies of this announcement] ************************************************************************** ******* Last call for papers - extended deadline: March 12, 2001 ******* ************************************************************************** International Workshop on Reduction Strategies in Rewriting and Programming (WRS 2001) held in conjunction with RTA 2001 Utrecht, The Netherlands, May 26, 2001 -------------------------------------------------------------------------- BACKGROUND AND AIMS Reduction strategies in rewriting and programming have attracted an increasing attention within the last years. New types of reduction strategies have been invented and investigated, and new results on rewriting / computation under particular strategies have been obtained. Research in this field ranges from primarily theoretical questions about reduction strategies to very practical application and implementation issues. The need for a deeper understanding of reduction strategies in rewriting and programming, both in theory and practice, is obvious, since they bridge the gap between unrestricted general rewriting (computation) and (more deterministic) rewriting with particular strategies (programming). Moreover, reduction strategies provide a natural way to go from operational principles (e.g., graph and term rewriting, narrowing, lambda-calculus) and semantics (e.g., normalization, computation of values, infinitary normalization, head-normalization) to implementations of programming languages. Therefore any progress in this area is likely to be of interest not only to the rewriting community, but also to neighbouring fields like functional programming, functional-logic programming, and termination proofs of algorithms. The workshop wants to provide a forum for the presentation and discussion of new ideas and results, recent developments, new research directions, as well as of surveys on existing knowledge in this area. Furthermore we aim at fostering interaction and exchange between researchers and students actively working on such topics. The workshop will be held in conjunction with RTA 2001 in Utrecht (The Netherlands) on May 26, 2001. It will feature 2 invited talks, a panel discussion, and contributed presentations selected from the submissions, with ample time for discussion. The workshop is (co-)organized by U Utrecht, TU Valencia and TU Wien. TOPICS OF INTEREST Topics of interest include, but are not restricted to, - theoretical foundations for the definition and semantic description of reduction strategies - strategies in different frameworks (term rewriting, graph rewriting, infinitary rewriting, lambda calculi, higher order rewriting, conditional rewriting, rewriting with built-ins, narrowing, constraint solving, etc.) and their application in (equational, functional, functional-logic) programming (languages) - properties of reduction strategies / computations under strategies (e.g., completeness, computability, decidability, complexity, optimality, (hyper-)normalization, cofinality, fairness, perpetuality, context-freeness, neededness, laziness, eagerness, strictness) - interrelations, combinations and applications of reduction under different strategies (e.g., equivalence conditions for fundamental properties like termination and confluence, applications in modularity analysis, connections between strategies of different frameworks, etc.) - program analysis and other semantics-based optimization techniques dealing with reduction strategies - rewrite systems / tools / implementations with flexible / programmable strategies as essential concept / ingredient - specification of reduction strategies in (real) languages - data structures and implementation techniques for reduction strategies. SUBMISSIONS We solicit papers on all aspects of reduction strategies in rewriting and programming. Submissions should describe unpublished work, except for survey papers which are explicitly welcome, too. Submissions should not exceed 10 pages (however, survey papers may be longer) and be sent in postscript format to the PC co-chairs (wrs01(a)logic.at) by March 12, 2001. Selection of papers by the PC will be based on originality, significance, and correctness. Accepted papers will be included in the workshop proceedings that will be available at the workshop, and electronically on the web. Depending on the number and quality of submissions, formal publication of the proceedings is envisaged. Researchers just interested in attending the workshop may send a corresponding email to wrs01(a)logic.at by March 12, 2001, preferably together with a brief position paper (up to two pages in postscript) describing their interest and/or work in the area. However, we will also consider late requests for attendance. INVITED TALKS Sergio Antoy (U Portland State): Evaluation Strategies for Functional Logic Programming Eelco Visser (U Utrecht, The Netherlands): A Survey of Strategies in Program Transformation Systems PANEL DISCUSSION on ``Hot Topics in Reduction Strategies'' with Michael Hanus U Kiel (Germany) Tetsuo Ida U Tsukuba (Japan) Paul Klint CWI & U Amsterdam (The Netherlands) PROGRAM COMMITTEE Maria Alpuente TU Valencia (Spain) Rachid Echahed IMAG Grenoble (France) Bernhard Gramlich (co-chair) TU Wien (Austria) Salvador Lucas (co-chair) TU Valencia (Spain) Vincent van Oostrom U Utrecht (The Netherlands) Rinus Plasmeijer KU Nijmegen (The Netherlands) Manfred Schmidt-Schauss U Frankfurt a.M. (Germany) Yoshihito Toyama U Tohoku (Japan) LOCAL ORGANIZATION Vincent van Oostrom U Utrecht (The Netherlands) IMPORTANT DATES Submission March 12, 2001 (extended deadline) Notification March 31, 2001 Final version April 30, 2001 Workshop May 26, 2001 FURTHER INFORMATION workshop website http://www.logic.at/wrs01/ workshop email address wrs01(a)logic.at **************************************************************************

1 0

Re: interesting example of lazyness/ghc optimisation
by George Russell 01 Mar '01

01 Mar '01

I hope nobody actually uses this random number generator. It appears to be linear congruential with period less than 140000. The author of these benchmarks seem to have got this algorithm from "Numerical Recipes", whose code (at least, in the early editions) has come in for some very heavy criticism: http://math.jpl.nasa.gov/nr/nr.html I would take these benchmarks more seriously if "random" implied a slightly more modern algorithm, like the Mersenne Twister (which has been implemented in Haskell I believe). Julian Assange wrote: > Noticing the use of tuples and normalisation at each step of > the iteration, I re-wrote this as: > > module Main(main) where > import System(getArgs) > import Numeric(showFFloat) > > main = do > ~[n] <- getArgs > putStrLn (showFFloat (Just 12) (random 42 (read n::Int) 100.0) "") > return 1 > > random :: Int -> Int -> Double -> Double > random seed n max = norm (rand n seed) max > where norm x max = (fromIntegral x) * (max / imd) > rand n x = if n > 0 then rand (n-1) ((x * ia + ic) `mod` im) else x > im = 139968 > imd = fromIntegral im > ia = 3877 > ic = 29573

1 0

Re: interesting example of lazyness/ghc optimisation
by Ketil Malde 01 Mar '01

01 Mar '01

There's something that I've been wanting to ask sombody about, since it isn't terribly clear to me. Blatantly hijacking a function from Julian's code: > runRandom last max num > | num > 1 = runRandom (fst new) max (num-1) > | otherwise = snd new While ifs are perhaps more intuitive when there are only two choices (like here!), and pattern matching matches on the type of arguments and not (only) their values, What's the difference between the pipe-syntax, and a case statement, i.e. writing the function as runRandom last max num = case num of 1 -> runRandom .... otherwise -> snd new Is there a practical difference, and if not, why are there two ways to skin this particular cat? Apologies for coming up with a terribly naive question, but I've looked in the HR without gaining much wisdom (beyond the recipe for translating if to case). -kzm -- If I haven't seen further, it is by standing in the footprints of giants

4 3

2001 Haskell Workshop: 1st call for papers
by Ralf Hinze 01 Mar '01

01 Mar '01

============================================================================ FIRST CALL FOR PAPERS [Deadline for submission: 1st June 2001] 2001 Haskell Workshop Firenze, Italy, 2nd September 2001 The Haskell Workshop forms part of the PLI 2001 colloquium on Principles, Logics, and Implementations of high-level programming languages, which comprises the ICFP/PPDP conferences and associated workshops. Previous Haskell Workshops have been held in La Jolla (1995), Amsterdam (1997), Paris (1999), and Montreal (2000). http://www.cs.uu.nl/people/ralf/hw2001.{html,pdf,ps,txt} ============================================================================ Scope ----- The purpose of the Haskell Workshop is to discuss experience with Haskell, and possible future developments for the language. The scope of the workshop includes all aspects of the design, semantics, theory, application, implementation, and teaching of Haskell. Submissions that discuss limitations of Haskell at present and/or propose new ideas for future versions of Haskell are particularly encouraged. Adopting an idea from ICFP 2000, the workshop also solicits two special classes of submissions, application letters and functional pearls, described below. Application Letters ------------------- An application letter describes experience using Haskell to solve real-world problems. Such a paper might typically be about six pages, and may be judged by interest of the application and novel use of Haskell. Functional Pearls ----------------- A functional pearl presents - using Haskell as a vehicle - an idea that is small, rounded, and glows with its own light. Such a paper might typically be about six pages, and may be judged by elegance of development and clarity of expression. Submission details ------------------ Deadline for submission: 1st June 2001 Notification of acceptance: 1st July 2001 Final submission due: 1st August 2001 Haskell Workshop: 2nd September 2001 Authors should submit papers of at most 12 pages, in postscript format, formatted for A4 paper, to Ralf Hinze (ralf(a)cs.uu.nl) by 1st June 2001. The use of the ENTCS style files is strongly recommended. Application letters and functional pearls should be labeled as such on the first page. They may be any length up to twelve pages, though shorter submissions are welcome. The accepted papers will be published as a University of Utrecht technical report. Programme committee ------------------- Manuel Chakravarty University of New South Wales Jeremy Gibbons University of Oxford Ralf Hinze (chair) University of Utrecht Patrik Jansson Chalmers University Mark Jones Oregon Graduate Institute Ross Paterson City University, London Simon Peyton Jones Microsoft Research Stephanie Weirich Cornell University ============================================================================

1 0

interesting example of lazyness/ghc optimisation
by Julian Assange 28 Feb '01

28 Feb '01

Brian Gregor wrote a haskell entrant for the random number generation in Doug's language shootout (http://www.bagley.org/~doug/shootout) Which is as follows: module Main where import System import Numeric iMi :: Int iMi = 139968 iMd :: Double iMd = 139968.0 iA ::Int iA = 3877 iC ::Int iC = 29573 nrRandom :: Int -> Double -> (Int,Double) nrRandom last max = (newlast,(max * (fromIntegral newlast)/iMd)) where newlast = (last*iA+iC) `mod` iMi runRandom :: Int -> Double -> Int -> Double runRandom last max num | num > 1 = runRandom (fst new) max (num-1) | otherwise = snd new where new = nrRandom last max main = do ~[n] <- getArgs putStrLn (showFFloat (Just 12) (runRandom 42 100.0 (read n::Int)) "") Noticing the use of tuples and normalisation at each step of the iteration, I re-wrote this as: module Main(main) where import System(getArgs) import Numeric(showFFloat) main = do ~[n] <- getArgs putStrLn (showFFloat (Just 12) (random 42 (read n::Int) 100.0) "") return 1 random :: Int -> Int -> Double -> Double random seed n max = norm (rand n seed) max where norm x max = (fromIntegral x) * (max / imd) rand n x = if n > 0 then rand (n-1) ((x * ia + ic) `mod` im) else x im = 139968 imd = fromIntegral im ia = 3877 ic = 29573 Interestingly, ghc / lazyness is able to detect that Brian's normalisation etc at each step is uneeded, and only perform it at the end. Consequently both of these programs are the the same speed (well, almost; bizarrely, Brian's seems to be about 2% faster). Julian.

1 0

My stack overflow was due to addListToFM
by senganb＠ia.nsc.com 28 Feb '01

28 Feb '01

Hi, Interestingly given the recent discussion about foldl versus foldl' I'd like to report that the stack overflow I was seeing was due to addListToFM which is defined using foldl although addToFM is strict in FiniteMap according to ghc: addToFM :: ... {PrelBase.Ord key} -> FiniteMap key elt -> key -> elt -> FiniteMap key elt {-## __A 4 __S LSLL __U ... So, I'd like to request that FiniteMap be changed to use foldl'. Compiler implementors, I'd also like to request that your compiler emit a warning in such cases if at all possible given that this has since even the best haskell programmers seem to miss this case (the authors of FiniteMap). --------------------------------------------------------------------------------- For those of you wondering how I found this, I used the following heuristic: 1 Build your program with -prof -auto-all -caf-all 2 Run your program with -h so that Main.hp is created. 3 Then look at the various functions that are reported to use a lot of heap. 4 Grep through Main.hp for the function that happens to have a recent rapid increase of heap. 5 Then look at its code. In my case I got find_atoms/add_atom_equalities 33808 find_atoms/add_atom_equalities 186128 find_atoms/add_atom_equalities 338448 find_atoms/add_atom_equalities 490768 find_atoms/add_atom_equalities 643088 find_atoms/add_atom_equalities 795408 find_atoms/add_atom_equalities 947728 find_atoms/add_atom_equalities 1100048 find_atoms/add_atom_equalities 1233976 find_atoms/add_atom_equalities 1302328 find_atoms/add_atom_equalities 1370680 find_atoms/add_atom_equalities 1439032 find_atoms/add_atom_equalities 1507384 find_atoms/add_atom_equalities 1575736 find_atoms/add_atom_equalities 1644088 find_atoms/add_atom_equalities 1697552 just before the crash. Reasoning --------- Since GHC optimizes tail recursion, the cause of a stack overflow must be something else, eg: a you've got a non-tail recursive function b your tail recursive function has built a long spine on the heap of functions that are strict in their 2cd argument so that you need to evaluate the bottom-most element of the spine (which causes you to push the address of all the functions along the spine onto the stack) c or... I assumed b since I know my code doesn't contain many cases of a. If b is the case, since the function that is overflowing is being evaluated, we know that it's going to build up its spine quickly and then try to evalute it, find that each function on the spine is strict in its second argument and then die. Therefore heuristic step 4. You also need it to use a fair bit of heap to actually be bad enough to overflow the stack (therefore heuristic step 3). Sengan

1 0

Re: Proposal: module namespaces.
by Christian Brolin 28 Feb '01

28 Feb '01

Malcolm Wallace wrote: > > Christian writes: > > What about the module declaration? Should it be: > > module Text.Xml.Parser where ... > > or just > > module Parser where ... -- located in Text/Xml/Parser.hs? > > The former. The reason is that a compiler needs to generate a unique > linker symbol for each defined function. If the full module name is > not encoded in the source file, you will need to add a commandline > option to the compiler, which is the wrong way to go in my opinion. What?? The compiler knows the full name of the module without the module clause. If it didn't do that, it can't find the modules to compile! Does the compiler opens every file on the Internet to check whether it is the file to compile? How does the compiler find the file to compile in the first place? What should the command line option you mentioned do? > Why is e.g. Parser.f not sufficient as a unique symbol for > Text.Xml.Parser.f? Well, what if you also have Text.Html.Parser.f? > You really need the full thing. Of course, see above. > > I would also like to import modules using relative addresses, e.g. the > > file: > > My/Small/Test/Xml/Parser.hs > > contains: > > import .Lexer -- Relative path to the module: My.Small.Test.Xml.Lexer > > import ..Data -- Relative path to the module: My.Small.Test.Xml.Parser.Data > > import Text.ParserCombinators.HuttonMeijer -- Absolute address > > I'm sorry, I don't entirely follow what the differing numbers of > initial dots mean. They are used to specify relative addresses to other modules. Relative addresses is a very important concept, but You missed it in your proposal. The dots was just my suggestion of a syntax for relative addresses. One dot: Relative to the parent of this module. Two dots: Relative to this module. E.g. module A.B.C.D1 where import A.B.C.D1.E1 import A.B.C.D1.E1.F import A.B.C.D1.E2 import A.B.C.D2 import X.Y.Z would be the same as (delete 'A.B.C'): module A.B.C.D1 where import .D1.E1 import .D1.E1.F import .D1.E2 import .D2 import X.Y.Z would be the same as (delete 'D1'): module A.B.C.D1 where import ..E1 import ..E1.F import ..E2 import .D2 import X.Y.Z Move the package of modules (A.B.C.*) to (Std.AAA.BBB.CCC.*) and rename D1 to DDD: module Std.AAA.BBB.CCC.DDD where import ..E1 import ..E1.F import ..E2 import .D2 import X.Y.Z The only thing that needs to be changed is the module clause. Which of course would be unnecessary if the module clause was dropped. > > When the world realize that this is the XML parser, they won't accept > > the name and I refuse to change my implementation. The only thing that > > is needed to rename (an unused) module hierarchy is to move it. > > If you refuse to change your implementation, someone else will change > it for you! You can't have closed standards. It is not necessary to modify the modules if the module system supports relative addresses!!! The steering wheel of my car is positioned relative to my car, so it is NOT necessary to change that position when I move the car. -- Christian Brolin

4 6

Re: Proposal: module namespaces.
by Ashley Yakeley 27 Feb '01

27 Feb '01

At 2001-02-26 09:59, Malcolm Wallace wrote: >Proposal 2 >---------- >Adopt a standardised namespace layout to help those looking for or >writing libraries, I'm a big fan of the Java reversed DNS style. Whatever, I think it's important that anyone with a domain name should be able to obtain a unique namespace without any further bureaucracy. In fact, whatever you decide it's likely to happen anyway, since people will decide that for instance "Com.Microsoft.Research.MyModule" is unlikely to clash with anyone outside the appropriate domains and subdivisions. I'm assuming that module name components have enforced capitalisation, like all other Haskell identifiers. >and a "Std" namespace prefix for genuinely >standard libraries. (These are two different things.) Eeesh, let's hope ICANN doesn't register a 'std' TLD. I would prefer "Standard" for this reason and also because the abbreviation seems pretty pointless. >In addition to a standardised hierarchy layout, I propose a truly >Standard-with-a-capital-S namespace. A separate discussion is needed >on what exactly would consitute "Standard" quality, but by analogy >with Java where everything beginning "java." is sanctioned by Sun, >I propose that every module name beginning "Std." is in some sense >sanctioned by the whole Haskell community. Do you have any kind of guarantees of copyright openness in mind? In Java, everything under java.* is supposed to be owned by Sun. Will it be standard practice for versions of Standard be included with Haskell compilers? Could the Prelude make use of Standard? Could Standard become an alternative to the Prelude? If answers to these last three are all "no", an alternative would be to put it under "Org.Haskell.Standard". -- Ashley Yakeley, Seattle WA

1 0

Proposal: module namespaces.
by Malcolm Wallace 27 Feb '01

27 Feb '01

This is an annoucement of a new mailing list, and a proposal for three things: * An extended mechanism for module namespaces in Haskell. * A "standard" namespace for new libraries, common across all systems. * A social process for adding new libraries to the "standard" set. A formatted version of this proposal appears on the web at http://www.cs.york.ac.uk/fp/libraries/ The new mailing list is for the discussion of these proposals. Please subscribe if you are interested. Follow-ups set accordingly. Mailing list details -------------------- libraries(a)haskell.org The purpose for this new list is to: (a) discuss an extension to Haskell to provide a richer module namespace, (b) discuss how to partition this namespace and populate it with libraries, (c) discuss how to provide a consistent set of libraries for all compilers, and the setting up of a common library repository. To subscribe: http://haskell.org/mailman/listinfo/libraries/ Introduction ------------ Everyone agrees that Haskell needs good, useful, libraries: lots of them, well-specified, well-implemented, well-documented. A problem is that the current "Standard Libraries" defined by the Haskell'98 Report number only about a dozen. But there are actually many more libraries out there: some are in GHC's hslibs collection, others are linked from haskell.org, even more are used only by their original author and have no public distribution. What is more, there is no Haskell Committee. There is no-one to decide which candidate libraries are worthy to be added to the "Standard" set. This stifles the possible distribution of great libraries, because no-one knows how to get /my/ library "accepted". Furthermore, the existing libraries that people distribute from their own websites often run into problems when used alongside other people's libraries. A library usually consists of several modules, but often the constituent modules have simple names that can easily clash with modules from another library package. This leads people to ad hoc solutions such as prefixing all their modules with a cryptic identifier e.g. HsParse XmlParse HOGLParse THIHParse Just counting the libraries currently available from GHC's hslibs, and haskell.org's links, there are currently over 200 separate modules in semi-"standard" use. As more libraries are written, the possibility of clashes can only increase. Related to this problem, although not identical, is the difficulty of finding a library that provides exactly the functionality you need to help you write a specific application program. How do you go about searching through 200+ modules for interesting-looking datatypes and signatures, starting only from the module names? My View ------- My view is that many of these problems are rooted in Haskell's restriction to a flat module namespace. If we can address that issue adequately, then I believe that many of the difficulties surrounding the provision of good libraries for Haskell will simply fall away. Proposal 1 ---------- Introduce nested namespaces for modules. The key concept here is to map the module namespace into a hierarchical directory-like structure. I propose using the dot as a separator, analogous to Java's usage for namespaces. So for instance, the four example module names above using cryptic prefixes could perhaps be more clearly named Haskell.Language.Parse Text.Xml.Parse Graphics.Drawing.HOpenGL.ConfigFile.Parse TypeSystem.Parse Naming proceeds from the most general category on the left, through more specific subdivisions towards the right. For most compilers and interpreters, this extended module namespace maps directly to a directory/file structure in which the modules are stored. Storing unrelated modules in separate directories (and related modules in the same directory) is a useful and common practice when engineering large systems. (But note that, just as Haskell'98 does not *insist* that modules live in files of the same name, this proposal does not insist on it either. However, we expect most tools to use the close correspondance to their advantage.) There are several issues arising from the particular proposal here. * This is a surface change to the module naming convention. It does not introduce nested /definition/ of modules. * The syntax I propose (a dot separator) is familiar from other languages such as Java, but could in principle be something else, for instance a prime ' or underscore _ or centred dot · or something different again. * Of the choices of separator, dot requires a change to the Haskell'98 lexical syntax, allowing modid -> qconid where currently the syntax is modid -> conid * The use of qualified imports becomes more verbose: for instance import qualified XmlParse ... XmlParse.element f ... becomes import qualified Text.Xml.Parse ... Text.Xml.Parse.element f ... However, I propose that every import have an implicit "as" clause to use as an abbreviation, so in import qualified Text.Xml.Parse [ as Parse ] the clause "as Parse" would be implicit, unless overridden by the programmer with her own "as" clause. The implicit "as" clause always uses the final subdivision of the module name. So for instance, either the fully-qualified or abbreviated-qualified names Text.Xml.Parse.element Parse.element would be accepted and have the same referent, but a partial qualification like Xml.Parse.element would not be accepted. * Another consequence of using the dot as the module namespace separator is that it steals one extremely rare construction from Haskell'98: A.B.C.D in Haskell'98 means the composition of constructor D from module C, with constructor B from module A: (.) A.B C.D No-one so far thinks this is any great loss, and if you really want to say the latter, you still can by simply inserting spaces: A.B . C.D Further down this document, I give more motivation and a rationale for this proposal of nested namespaces. But first, two other proposals which rest on the first one. Proposal 2 ---------- Adopt a standardised namespace layout to help those looking for or writing libraries, and a "Std" namespace prefix for genuinely standard libraries. (These are two different things.) The hslibs collection of modules is a great starting place for finding common libraries that could become standards. I propose that we adopt a "standardised" namespace hierarchy, based on the current hslibs layout, into which Haskell programmers can plug their own libraries relatively easily (whether they intend to release them or not). The aim is to make it clear where to place a new module, and where to search for a possible existing module. For instance, in ASCII art, here is a small part of a suggested tree. + Data + Structures + Trees + AVL | | | + RedBlack | | | | | + Queue + Bankers | | + FIFO | + Encoding + Binary | + MD5 | + Graphics + UI + Gtk + Widget | | | + Pane | | | + Text | | | | | + FranTk | | | + Drawing + HOpenGL + .... | | + Vector | | | + Format + Jpeg | + PPM + Haskell + .... | A fuller proposed layout appears on the web at http://www.cs.york.ac.uk/fp/libraries/layout.html In addition to a standardised hierarchy layout, I propose a truly Standard-with-a-capital-S namespace. A separate discussion is needed on what exactly would consitute "Standard" quality, but by analogy with Java where everything beginning "java." is sanctioned by Sun, I propose that every module name beginning "Std." is in some sense sanctioned by the whole Haskell community. So for instance, an experimental, or not-quite-complete, library could be called Text.Xml but only a guaranteed-to-be-stable, complete, library could be called Std.Text.Xml The implication of the Std. namespace is that all such "standard" libraries will be distributed with all Haskell systems. In other words, you can rely on a standard library always being there, and always having the same interface on all systems. Proposal 3 ---------- Develop a process by which candidate libraries can be proposed to enter the Std namespace. Since Haskell'98 is fixed, and there is no longer a Haskell Committee, there is no official body capable of deciding new standards for libraries. However, we do have a Haskell community which will use or not use libraries, depending on their quality. So libraries will become standards by a de-facto process, rather than de-jure. Apart from the Haskell compiler implementers, we wanted a means to encourage the whole community to be involved in recognising de facto "standard" libraries. The mailing list 'libraries(a)haskell.org' is one contribution. We hope this will work on the same model as the FFI mailing list, which has been pretty successful at allowing a community of designers and implementers to explore their FFI needs and solidify a design that is common across at least three Haskell systems. On top of this discussion however, some final decisions will have to be made on which libraries achieve entry to the "Std." namespace. The Haskell implementers have collectively proposed a ruling troika, one representing each of the three main Haskell systems (Hugs,ghc,nhc98). These are Simon Marlow, representing ghc, and current keeper of the hslibs collection; Malcolm Wallace, representing nhc98; and Andy Gill, representing Hugs users. Some obvious criteria for entry to the "Std." namespace would be: * The interface is stable and unlikely to change significantly; * The library is written in pure Haskell'98. This criterion is likely to be the most contentious, so perhaps a better idea would be that ... * ... an implementation exists for at least the three Haskell systems Hugs, ghc, and nhc98; * The library is already in current use, so bugs in its coding and design have been ironed out; * The Haskell community recognises it as solving a common task, or encapsulating a common programming idiom. These suggested criteria need some discussion and improvement. After the initial period of deciding what belongs in the "Std." namespace, I would expect any further candidate libraries that are proposed for standardisation to spend some time in another part of the namespace hierarchy whilst they gain stability and common acceptance, before being moved to "Std.". Rationale and Motivation for Proposal 1 (nested namespaces) ----------------------------------------------------------- Scenario 1 ---------- Imagine you have just written a new library of, say, Pretty-printing combinators. You want to release it to the Haskell public. So what do you call it? module Pretty -- already taken (several times) module UU_Pretty -- also taken module PrettyLib -- already exists as well Ok, so lacking any further inspiration, you end up deciding to call it module MyPretty -- ! Surely there must be a better solution. Of course there is - namespaces. Let's classify libraries that do similar jobs together: module Text.PrettyPrinter.Hughes -- the original Hughes design module Text.PrettyPrinter.HughesPJ -- later modified by Simon PJ module Text.PrettyPrinter.UU -- the Utrecht design module Text.PrettyPrinter.Chitil -- Olaf's new design These are exactly the same Pretty libs as before, but named more sensibly. It is still clear that each is a pretty-printing library, but it is also clear that they are different. Incidentally, have you ever tried to write your own module called Pretty? You may have discovered with GHC (which has a Pretty already in the hslibs collection), that you get strange errors. This is because sometimes the compiler can be confused into reading one Pretty.hi interface file (i.e. yours), yet linking the other Pretty.o object file (i.e. from hslibs), ending in a core dump. With proper module namespaces, this confusion should never happen again. Scenario 2 ---------- You are writing a complex library that has a couple of layers of abstraction. For some users, you want to expose just a small high-level set of types and functions. Other users will need more detailed access to lower-level stuff. With namespaces, you can use the directory-like structure to make these kinds of access explicit. For instance, imagine a socket library: module Network.Socket It exports an /abstract/ type Socket for ordinary users - they only need to know its name. More advanced hackers however can play with the details of the type, because you also have: module Network.Socket.Types which exports the Socket type non-abstractly i.e. Socket(..). And of course this abstraction is easy for the library-writer to manage, because the implementation of the more abstract layer simply imports and re-exports a careful selection of the more detailed layers. Don't forget that, in terms of the actual filesystem layout, it is perfectly OK to have e.g. file Network/Socket.hs dir NetWork/Socket file Network/Socket/Types.hs Scenario 3 ---------- You are managing a software engineering project. Several people are working more-or-less independently on different sections of the program. To avoid mistakes with files, you give each one a separate directory to place their code in. But in Haskell'98 this is not enough to ensure that they invent module names that do not clash with other developers' modules. So you insist that everyone also uses a prefix-naming scheme for each appropriate sub-task. For instance, here is a sketch of the layout of the Galois Connection team's entry in the ICFP 2000 programming contest: dir CSG -- constructive solid geometry file CSG/CSG.hs file CSG/CSGConstruct.hs file CSG/CSGGeometry.hs file CSG/CSGInterval.hs dir Fran -- Fran-style animation file Fran/FranLite.hs file Fran/FranCSG.hs dir GML -- interpreter for little language file GML/GMLData.hs file GML/GMLParse.hs file GML/GMLPrimitives.hs So now the problem is that to actually build the software, you need to write a Makefile that descends into these directories. Or maybe you use 'hmake' like so: hmake examples/chess.hs -ICSG -IFran -IGML -IRayTrace -package text Note how many sub-directories you must remember to add to the command line (this applies equally for compiler options in Makefiles). Note also the inconsistency between compiling and linking /my/ modules, against using and linking a "standard" hslibs module from package text. Isn't there a simpler way? Yes. Namespaces. Prefix naming is no longer needed inside directories, because the directory name is /part/ of the module name: file CSG.hs -- re-exports everything from the CSG dir dir CSG file CSG/Construct.hs file CSG/Geometry.hs file CSG/Interval.hs dir Fran file Fran/Lite.hs file Fran/CSG.hs -- does not conflict with top-level CSG.hs dir GML file GML/Data.hs file GML/Parse.hs file GML/Primitives.hs And now, the commandline to 'hmake' (or compiler options in a Makefile) becomes simply: hmake examples/chess.hs -I. You only need to specify the root of the module tree (-I.), and all modules in all subdirectories can be found via their full namespace path as used in the source files. Note also that, whereas previously we needed to specify a package for whatever hslibs modules were used, now the compiler/hmake already knows the root of the installed hslibs tree and can use the same mechanism to find and link "standard" modules as for user modules. >From this example it should be clear that the use of module namespaces is of benefit to ordinary programs that may never become public, quite aside from any benefits we expect to derive in managing publically-distributed library code. What now? --------- Ok, so that's my proposal. The implementers of some of the main Haskell systems have seen a presentation of these ideas, and seemed to like them. Namespaces are already implemented in nhc98 (v1.02) and hmake (v2.02) if you want to play with them. I expect some discussion to refine this proposal on the 'libraries(a)haskell.org' list, to which everyone interested is invited. Once we have nailed down the precise design, we need to get matching implementations in all systems. I have rashly volunteered to implement the lexical/parsing/module-search changes in any Haskell system that no-one else volunteers for (probably ghc, Hugs, possibly hbc). But after that we will still have many more decisions to take about individual libraries, precise naming, build systems, and so on, not to mention actually writing the libraries. Get involved. Contribute. Regards, Malcolm

4 4