Sun 11 Sep 2022 10:00 - 10:30 at M1 - HOPE Session 1
Scoped effects as described by Wu et al are a powerful addition to algebraic effect systems in Haskell, but there is currently no effect system in Haskell that has both the power of scoped effects and an efficient implementation. We demonstrate how the efficient evidence-passing style of effect system implementation by Xie and Leijen can be altered to support scoped effects. We present our implementation of this idea, SpEff
, providing usage examples, interesting design choices, and benchmark results, in which it performs better than EvEff
in Xie and Leijen, as well as other popular effect systems. We also talk about the limitations of evidence-passing style in current GHC Haskell.
Talk slides (InScope.pdf) | 157KiB |
Sun 11 SepDisplayed time zone: Belgrade, Bratislava, Budapest, Ljubljana, Prague change
Sun 11 Sep
Displayed time zone: Belgrade, Bratislava, Budapest, Ljubljana, Prague change
09:00 - 10:30 | |||
09:00 30mTalk | Lift Inference for Lexical Effect Handlers with Second-Class Functions HOPE Marius Müller University of Tübingen, Philipp Schuster University of Tübingen, Jonathan Immanuel Brachthäuser University of Tübingen, Klaus Ostermann University of Tübingen | ||
09:30 30mTalk | Higher order programming with probabilistic effects: A model of stochastic memoization and name generation HOPE | ||
10:00 30mTalk | Effect Handlers in Scope, EvidentlyVirtual HOPE File Attached |