exoteric said:
Charles said:

Simon Peyton-Jones on the cutting edge and future of Haskell and C--. Secondarily, what he thinks about Phoenix.


Looking at his 2010 papers he's been busy:


New papers

  • [July 2010] Generative Type Abstraction and Type-level Computation (with Stephanie Weirich, Dimitrios Vytiniotis, and Steve Zdancewic).  This paper, submitted to POPL'11, describes how to combine Haskell's new abilities to do type-level computation with the old ability to do type abstraction using newtypes.  Rather surprisingly (to me) the two are in tension, and the solution is quite interesting.
  • [June 2010] Seq no more (by Simon Marlow, Patrick Maier, Phil Trinder, Hans-Wolfgang Loidl, and Mustafa Aswad).  A new take on the "algorithms + strategies = parallelism" story, providing a nice Haskell library to support parallel computation.  To appear at the 2010 Haskell Symposium.
  • [June 2010] Supercompilation by evaluation (with Max Bolingbroke).  This is our first foray into supercompilation.  To appear at the 2010 Haskell Symposium.
  • [June 2010] The performance of Haskell 'containers' package (by Milan Straka).  Milan did this work while an intern here at MSR Cambridge.  To appear at the 2010 Haskell Symposium.
  • [May 2010] Modular type inference with local assumptions (with Dimitrios Vytiniotis, Tom Schrijvers, Martin Suzmann).  This epic 70-page JFP submission brings together, in a single uniform framework, a series of our earlier papers on type inference for type systems involving local constraints, including GADTs and indexed type families.
  • [April 2010] Hoopl: A Modular, Reusable Library for Dataflow Analysis and Transformation (with John Dias and Norman Ramsey). This paper,completely rewritten in April 2010, describes our framework for doing dataflow optimisation on imperative C-- programs. To appear at the 2010 Haskell Symposium.
  • [April 2010] Regular, shape-polymorphic, parallel arrays in Haskell (with Manuel Chakravarty, Gabriele Keller, Roman Leshchinskiy, and Ben Lippmeier) describes a high-performance data-parallel library for regular arrays. To appear at ICFP 2010.

Would be great to talk to Simon again. That said, he works and lives in Cambridge, UK. So, we'll see if he's up for this when he next visits Redmond.