はてなブックマーク

Dynamic scoping is an effect, implicit parameters are a coeffect For the longest time, I thought of implicit parameters and dynamic scoping were basically the same thing, since they both can be used to solve similar problems (e.g., the so called "configuration problem" where you need to plumb down some configuration deep into a nested body of function definitions without defining them all explicit

This post is a long form essay version of a talk about PyTorch internals, that I gave at the PyTorch NYC meetup on May 14, 2019. This talk is for those of you who have used PyTorch, and thought to yourself, "It would be great if I could contribute to PyTorch," but were scared by PyTorch's behemoth of a C++ codebase. I'm not going to lie: the PyTorch codebase can be a bit overwhelming at times. The

You might be wondering: with the beta release of no-reinstall Cabal, is this functionality be coming to GHC 8.0? (Or even a new release of Cabal, since the no-reinstall feature works with GHC 7.10). Unfortunately, there is a split among the Cabal developers over whether or not the actual no-reinstall behavior should go into Cabal by default as is. Duncan Coutts, in particular, has argued that it's

Subtyping is one of those concepts that seems to makes sense when you first learn it (“Sure, convertibles are a subtype of vehicles, because all convertibles are vehicles but not all vehicles are convertibles”) but can quickly become confusing when function types are thrown into the mix. For example, if a is a subtype of b, is (a -> r) -> r a subtype of (b -> r) -> r? (If you know the answer to th

Update. Want to know more about Backpack? Read the specification So perhaps you've bought into modules and modularity and want to get to using Backpack straightaway. How can you do it? In this blog post, I want to give a tutorial-style taste of how to program Cabal in the Backpack style. These examples are executable, but you'll have to build custom versions of GHC and Cabal to build them. Comment

The fundamental problem of programming language package management Why are there so many goddamn package managers? They sprawl across both operating systems (apt, yum, pacman, Homebrew) as well as for programming languages (Bundler, Cabal, Composer, CPAN, CRAN, CTAN, EasyInstall, Go Get, Maven, npm, NuGet, OPAM, PEAR, pip, RubyGems, etc etc etc). "It is a truth universally acknowledged that a prog

Type classes: confluence, coherence and global uniqueness Today, I'd like to talk about some of the core design principles behind type classes, a wildly successful feature in Haskell. The discussion here is closely motivated by the work we are doing at MSRC to support type classes in Backpack. While I was doing background reading, I was flummoxed to discover widespread misuse of the terms "conflue

tl;dr The scope of backtracking try should be minimized, usually by placing it inside the definition of a parser. Have you ever written a Parsec parser and gotten a really uninformative error message? "test.txt" (line 15, column 7): unexpected 'A' expecting end of input The line and the column are randomly somewhere in your document, and you're pretty sure you should be in the middle of some stack

So you may have heard about this popular new programming language called Haskell. What's Haskell? Haskell is a non-dependently typed programming language, sporting general recursion, type inference and built-in side-effects. It is true that dependent types are considered an essential component of modern, expressive type systems. However, giving up dependence can result in certain benefits for othe

POPL is almost upon us! I’ll be live-Tumblr-ing it when the conference comes upon us proper, but in the meantime, I thought I’d write a little bit about one paper in the colocated PEPM'14 program: The HERMIT in the Stream, by Andrew Farmer, Christian Höner zu Sierdissen and Andy Gill. This paper presents an implementation of an optimization scheme for fusing away use of the concatMap combinator in

This Lambda the Ultimate post (dated 2010) describes a rather universal problem faced by compiler writers: how does one go about adding “extra information” (e.g. types) to an AST? (The post itself divides the problem into three components: adding the information to the data types, using the information to inform the construction of the node, and using the information to inform the destruction of a

Christopher de Sa and I have been working on a category theoretic approach to optimizing MapReduce-like pipelines. Actually, we didn’t start with any category theory—we were originally trying to impose some structure on some of the existing loop optimizations that the Delite compiler performed, and along the way, we rediscovered the rich relationship between category theory and loop optimization.

Recursion and induction are closely related. When you were first taught recursion in an introductory computer science class, you were probably told to use induction to prove that your recursive algorithm was correct. (For the purposes of this post, let us exclude hairy recursive functions like the one in the Collatz conjecture which do not obviously terminate.) Induction suspiciously resembles rec

I’d like to talk about some nitty-gritty details of GHC’s thread scheduling, discovered over the course of working on stride scheduling for GHC. Most of these choices are merely implementation details and are not part of any specification. While these choices shouldn’t be relied upon, they are worth knowing, since many of these details were accreted over the course of many performance bugs, benchm

Functions are awesome. What if we made a PL that only had functions? Objects are awesome. What if we made a PL where everything was an object? Lazy evaluation is awesome. What if we made a PL where every data type was lazy? Extremist programming (no relation to extreme programming) is the act of taking some principle, elevating it above everything else and applying it everywhere. After the dust se

Jon Howell dreams of a new Internet. In this new Internet, cross-browser compatibility checking is a distant memory and new features can be unilaterally be added to browsers without having to convince the world to upgrade first. The idea which makes this Internet possible is so crazy, it just might work. What if a web request didn’t just download a web page, but the browser too? “That’s stupid,” y

(Introductory material ahead.) Most readers of this blog should have at least a passing familiarity with applicative functors: class Applicative f where pure :: a -> f a (<*>) :: f (a -> b) -> f a -> f b This interface is quite convenient for day-to-day programming (in particular, it makes for the nice f <$> a <*> b <*> c idiom), but the laws it obeys are quite atrocious: [identity] pure id <*> v

Editor's note. I've toned down some of the rhetoric in this post. The original title was "monad-control is unsound". MonadBaseControl and MonadTransControl, from the monad-control package, specify an appealing way to automatically lift functions in IO that take "callbacks" to arbitrary monad stacks based on IO. Their appeal comes from the fact that they seem to offer a more general mechanism than

There are two primary reasons why the low-level implementations of iteratees, enumerators and enumeratees tend to be hard to understand: purely functional implementation and inversion of control. The strangeness of these features is further exacerbated by the fact that users are encouraged to think of iteratees as sinks, enumerators as sources, and enumeratees as transformers. This intuition works

tl;dr — Save this page for future reference. Have you ever been in the situation where you need to quickly understand what a piece of code in some unfamiliar language does? If the language looks a lot like what you’re comfortable with, you can usually guess what large amounts of the code does; even if you may not be completely familiar how all the language features work. For Haskell, this is a lit

A big thanks to everyone who everyone who sent in space leak specimens. All of the leaks have been inspected and cataloged by our experts, and we are quite pleased to open the doors of the space leak zoo to the public! There are a few different types of space leak here, but they are quite different and a visitor would do well not to confuse them (the methods for handling them if encountered in the

In 2007, Eric Kidd wrote a quite popular article named 8 ways to report errors in Haskell. However, it has been four years since the original publication of the article. Does this affect the veracity of the original article? Some names have changed, and some of the original advice given may have been a bit... dodgy. We’ll take a look at each of the recommendations from the original article, and al

When we teach beginners about Haskell, one of the things we handwave away is how the IO monad works. Yes, it’s a monad, and yes, it does IO, but it’s not something you can implement in Haskell itself, giving it a somewhat magical quality. In today’s post, I’d like to unravel the mystery of the IO monad by describing how GHC implements the IO monad internally in terms of primitive operations and th