Changes between Version 17 and Version 18 of Status/May11


Ignore:
Timestamp:
Apr 29, 2011 8:38:34 AM (3 years ago)
Author:
simonpj
Comment:

--

Legend:

Unmodified
Added
Removed
Modified
  • Status/May11

    v17 v18  
    2626 
    2727 With the advent of GADTs and type families, the type system of the 
    28  Core necessarily has to become a bit more complicated.  For a few 
    29  years wa have been using an extension of System F, called System 
    30  FC, as described in our paper [[http://research.microsoft.com/en-us/um/people/simonpj/papers/ext-f/ FC]].  However, the way that System FC 
     28 Core had to grow a little.  For a few 
     29 years we have been using an extension of System F, called System 
     30 FC, as described in our 2007 paper [[http://research.microsoft.com/en-us/um/people/simonpj/papers/ext-f/ FC]].  However, the way that System FC 
    3131 was actually ''implemented'' in GHC's Core language was a bit unsatisfactory 
    32  so, with help from Brent Yorgey, Simon PJ is busy re-engineering it.  
     32 so, with help from '''Brent Yorgey''', Simon PJ is busy re-engineering it.  
    3333 In particular, FC has ''coercion terms'', and these will now  
    3434 be represented by their own data type `Coercion`, rather than being 
     
    3737 equivalent smaller ones. All this is described in our new paper [http://research.microsoft.com/en-us/um/people/simonpj/papers/ext-f/ !NewFC]. 
    3838 These changes will (finally) complete the type-family story by 
    39  making so-called "equality superclasses" work for the first time. 
     39 making so-called "equality superclasses" work for the first time in GHC 7.2. 
    4040 
    4141 * '''Pedro Magalhaes''' has nearly completed his implementation of the 
     
    4343   Haskell Symposium paper [[http://www.dreixel.net/research/pdf/gdmh_nocolor.pdf Derivable]].  It will be in GHC 7.2. 
    4444 
    45  * '''Edward Yang''' has done a lot of work on the long-promised new 
    46    code-generation path.  Hoopl is now fully part of GHC, and 
    47    the new path uses it extensively.  We are now working on making 
    48    the new pipeline generate better code than the old one.  Stay 
    49    tuned. 
     45 * '''Edward Yang''' has spearheaded a flurry of work on the new 
     46   code generation backend (`-fuse-new-codegen`, the rewrite of 
     47   the part of GHC that turns STG syntax into C--).  Hoopl is now 
     48   fully part of GHC [[http://research.microsoft.com/en-us/um/people/simonpj/papers/c--/ Hoopl]], and the new path uses it extensively; we’ve 
     49   ironed out most of the bugs in the backend; and now we’re 
     50   working on new optimization passes and fixing inefficiencies to 
     51   get the generated code as good (or better) than the old code 
     52   generator.  We’re still not at the point where the new code 
     53   generator will generate better code, but we’re pretty close! 
     54   Stay tuned. 
    5055 
    5156 * '''Simon Marlow: Any update?''' Previous entry: ''In joint work with Phil Trinder and his colleagues at Herriot Watt, Simon M designed implemented a new '''parallel strategies library''', described in their 2010 Haskell Symposium paper [[http://www.haskell.org/~simonmar/papers/strategies.pdf Seq]].'' 
     
    5459sleeves and help us with GHC.  Amongst those who have been active recently are: 
    5560 * Mark Lentczner and Dan Knapp have been working on cross-compilation support 
     61 * Continued work on the new I/O manager by Johan Tibbe. 
    5662 * Various improvements and build fixes for OS X, from PHO, Greg Wright, Thorkil Naur and William Knop 
    5763 * Solaris fixes from Karel Gardas and Christian Maeder 
     
    6470 * General tidyups from Matthias Kilian 
    6571 * Primop improvements from Daniel Peebles 
    66  * Some GHCi improvements from Vivian McPhail and Boris Lykah 
     72 * Some GHCi improvements from Vivian !McPhail and Boris Lykah 
    6773 * More GHCi debugger fixes from Pepe Iborra 
    6874 * LLVM development continues with David Terei 
     
    7884Here's a selection that we know about. 
    7985 
    80  * '''Jeff Epstein''' has implemented a library that brings Erlang's programming model to Haskell programmers.  In particular, you can write a Haskell program that runs on a cluster of machines that do not share memory.  It is all based on a modest but powerful language extension that makes it possible for a programmer to work with "static" functions; that is, ones consisting of pure code with no free variables.  The paper that describes all this is called "Haskell for the cloud" [[http://research.microsoft.com/en-us/um/people/simonpj/papers/parallel/ Cloud]]. 
     86 * '''Jeff Epstein''', in collaboration with Andrew Black, has implemented a library that brings Erlang's programming model to Haskell programmers.  In particular, you can write a Haskell program that runs on a cluster of machines that do not share memory.  It is all based on a modest but powerful language extension that makes it possible for a programmer to work with "static" functions; that is, ones consisting of pure code with no free variables.  The paper that describes all this is called "Haskell for the cloud" [[http://research.microsoft.com/en-us/um/people/simonpj/papers/parallel/ Cloud]]. 
    8187 
    8288 * ''David Mazieres at Stanford wants to implement '''Safe Haskell''', a flag for GHC that will guarantee that your program does not use `unsafePerformIO`, foreign calls, RULES, and other stuff stuff.'' 
     
    116122 * [FC] "System F with type equality coercions", Sulzmann, Chakravarty, Peyton Jones, TLDI 2007, http://research.microsoft.com/en-us/um/people/simonpj/papers/ext-f/ 
    117123 
     124 * [Hoopl] "A modular, reusable library for dataflow analysis and transformation", Dias, Ramsey, and Peyton Jones, Haskell Symposium 2010, http://research.microsoft.com/en-us/um/people/simonpj/papers/c--/ 
     125 
    118126 * [!NewFC] "Practical aspects of evidence-based compilation in System FC", Vytiniotis and Peyton Jones, submitted to ICFP 2011, http://research.microsoft.com/en-us/um/people/simonpj/papers/ext-f/ 
    119127