Lambda the Ultimate

DISCLAIMER: I'm uneducated with PLT and don't know what I'm talking about, so please forgive any whacked terminology.

Ok, that outta the way...

I'm trying to make my way through The Duality of Computation (https://pdfs.semanticscholar.org/048f/c94be2ec752bb210c5f688cba0200c1a1f92.pdf), and this stuff, like most everything posted on this fascinating website, is way over my head. I can't follow most of the back half of the paper, but I was hoping someone might have the spare time to answer two newbie, school-level, questions...

1) (The silly one.) How on earth do you pronounce the "mu, mu with tilde" calculus. It'd literally be easier to read this if I knew how to read it. Is there some definitive guide for people who just figured out that the funny squiggle so many these papers is pronounced "eta"? :)

2) (The basic one.) The authors talk about the call-by-value side of the duality as dealing with "Environments with holes." This reminds me of PTS, with it's capital-PI lambda expressions that can close over terms or types. What's the difference between an "Environment with a hole" and a lambda over environments? And what does that mean for compiled languages, where environments aren't parameters?

Any help on either question would be appreciated.

Just saw this on Hacker News -- Project Loom: Fibers and Continuations for the Java Virtual Machine with the following overview:

Project Loom's mission is to make it easier to write, debug, profile and maintain concurrent applications meeting today's requirements. Threads, provided by Java from its first day, are a natural and convenient concurrency construct (putting aside the separate question of communication among threads) which is being supplanted by less convenient abstractions because their current implementation as OS kernel threads is insufficient for meeting modern demands, and wasteful in computing resources that are particularly valuable in the cloud. Project Loom will introduce fibers as lightweight, efficient threads managed by the Java Virtual Machine, that let developers use the same simple abstraction but with better performance and lower footprint. We want to make concurrency simple(r) again! A fiber is made of two components — a continuation and a scheduler. As Java already has an excellent scheduler in the form of ForkJoinPool, fibers will be implemented by adding continuations to the JVM.

I'm a fan of fibers, and this has quite a bit of interesting material in it for like-minded folks.

A puzzlingly named, exceedingly technical device introduced to structure the denotational semantics has by now achieved cult status. It has been married to effects -- more than once. It is compulsively looked for in all manner of things, including burritos. At least two ICFP papers brought it up without a rhyme or reason (or understanding), as the authors later admitted. I am talking about monads.

In truth, effects are not married to monads and approachable directly. The profound insight behind monads is the structuring, the separation of `pure' (context-independent) and effectful computations. The structuring can be done without explicating mathematical monads, and especially without resorting to vernacular monads such as State, etc. This article gives an example: a simple, effectful, domain-specific sublanguage embedded into an expressive `macro' metalanguage. Abstraction facilities of the metalanguage such higher-order functions and modules help keep the DSL to the bare minimum, often to the first order, easier to reason about and implement.

The key insight predates monads and goes all the way back to the origins of ML, as a scripting language for the Edinburgh LCF theorem prover. What has not been clear is how simple an effectful DSL may be while remaining useful. How convenient it is, especially compared to the monadic encodings. How viable it is to forsake the generality of first-class functions and monads and what benefits it may bring. We report on an experiment set out to explore these questions.

We pick a rather complex effect -- non-determinism -- and use it in OCaml, which at first blush seems unsuitable since it is call-by-value and has no monadic sugar. And yet, we can write non-deterministic programs just as naturally and elegantly as in Haskell or Curry.

The running tutorial example is computing all permutations of a given list of integers. The reader may want to try doing that in their favorite language or plain OCaml. Albeit a simple exercise, the code is often rather messy and not obviously correct. In the functional-logic language Curry, it is strikingly elegant: mere foldr insert []. It is the re-statement of the specification: a permutation is moving the elements of the source list one-by-one into some position in the initially empty list. The code immediately tells that the number of possible permutations of n elements is n!. From its very conception in the 1959 Rabin and Scott's paper, non-determinism was called for to write clear specifications -- and then to make them executable. That is what will shall do.

I do not know how much interest this community has in the intersection of programming languages, memory management and type systems, but for those intrigued by such topics, you might find this paper on Gradual Memory Management to be worth reading.

It proposes that a language's compiler offer more sophisticated type systems that enable a program to flexibly support multiple memory management mechanisms for improved performance, and do so with safety guaranteed at compile time. The described type systems build up from Rust's lifetime-driven owner/borrower model as well as Pony's reference capabilities (mutability/aliasing permissions). The paper also references Microsoft's experimental work on Midori.

I welcome any feedback or questions.

I have implemented programming language Theme-D, which is a Scheme-like programming language with static typing. Some properties of Theme-D include:
- Static type system
- A simple object system
- Multi-methods dispatched runtime (and also compile-time)
- Parametrized (type parameters) classes, types, and procedures
- Signature types resembling Java interfaces but multiply dispatched
- A module system
- Two kinds of variables: constants and mutable variables

Theme-D homepage is located at

http://www.tohoyn.fi/theme-d/index.html

Theme-D can also be found at

https://sourceforge.net/projects/theme-d/

I have also ported (a subset of) guile-gnome GUI library to Theme-D. Its homepage is located at

http://www.tohoyn.fi/theme-d/theme-d-gnome.html

and it can also be found at

https://sourceforge.net/projects/theme-d-gnome/

- Tommi Höynälänmaa

The Platonic Solids of Software Construction and Their Realization in C

Synopsis -

Here I try to contrive 5 (actually ends up being 6) 'Platonic Solids' of software construction - IE, the fundamental elements of programming that all programmers in all business domains end up leveraging regardless of their general purpose programming language.

As a practical matter, I then demonstrate how different aspects of each element are either supportable - or not supportable in C. A language like C is chosen partially because when we use it to encode these elements, its weak language semantics actually enable us to understand each element in a more isolated way. For discussion at this level of analysis, this turns out to be useful.

However, I warn readers that this gist-article is more conjecture than science, an emerging idea that, if accurate in its notions, is a precursor to a rigorous investigation. That is why I offer it up for evaluation and critique here.

BCS FACS - Annual Peter Landin Semantics Seminar: Compiling Without Continuations

Date/Time: Tuesday 12 December 2017, 6.00pm - 9.00pm

Venue: BCS, 1st Floor, The Davidson Building, 5 Southampton Street, London, WC2E 7HA

Speaker: Professor Simon Peyton Jones, FRS (Microsoft Research)

Cost: Free

Booking: https://events.bcs.org/book/2701/

Synopsis:

Peter Landin (1930 - 2009) was a pioneer whose ideas underpin modern computing. In the 1950s and 1960s, Landin showed that programs could be defined in terms of mathematical functions, translated into functional expressions in the lambda calculus, and their meaning calculated with an abstract mathematical machine. Compiler writers and designers of modern-day programming languages alike owe much to Landin's pioneering work.

Each year, a leading figure in computer science will pay tribute to Landin's contribution to computing through a public seminar. This year's seminar is entitled “Compiling Without Continuations” and will be given by Professor Simon Peyton Jones, FRS (Microsoft Research).

Programme

5.15pm Coffee

6.00pm Welcome & Introduction

6.05pm Peter Landin Semantics Seminar

Compiling Without Continuations

Professor Simon Peyton Jones, FRS (Microsoft Research)

7.20pm Drinks Reception

Seminar details

GHC compiles Haskell via Core, a tiny intermediate language based closely on the lambda calculus. Almost all GHC’s optimisations happen in Core, but until recently there was an important kind of optimisation that Core really did not handle well. In this talk I’ll show you what the problem was, and how Core’s new “join points” solve it simply and beautifully, by effectively allowing Core to express control flow as well as data flow; there are strong links to so-called “continuation passing style” (CPS) here.

Understanding join points can help you are a programmer too, because you can write code confident that it will optimise well. I’ll show you a rather compelling example this: “skip-less streams” now fuse well, for the first time, which allows us to drop the previous (ingenious but awkward) workarounds.

A thought experiment I am too lazy to do so I'll ask you folk.

Define SK, define a reduction relation, ask whether two terms are equal/reduce to similar terms.

Can you do this in Prolog? (Asked because of interest in current unification based languages like miniKanren.)

It came up in an old LTU thread about regular expressions. There was an argument whether tagged FA invented by Ville Laurikari can support POSIX disambiguation semantics. It turns out, they can: it is possible to construct efficient Laurikari TDFA with POSIX semantics, as well as with leftmost greedy semantics (details in this paper).

Back in 2007 Chris Kuklewicz suggested an algorithm which is implemented in his Regex-TDFA Haskell library. He also wrote an informal description of his algorithm, but never fully formalized it. Some ten years later I stumbled upon this thread when I was trying implement fast submatch extraction in the open source lexer generator re2c. I revised both the original algorithm by Laurikari and the modification by Kuklewicz and found a number of improvements and bugs in Regex-TDFA.

Are there any special programming language features ("superpowers") that are made possible by specifically exploting a tracing JIT compiler?

It seems to me like tracing JITs are often developed for existing languages (JavaScript, Python, Lua, etc) and tend to focus on efficiently compiling the existing code bases that are written with their traditional programming idioms. I am interested in the opposite approach: what programming idioms and language extensions can help programmers to make the most of their tracing JIT compiler?

I can offer one candidate. Tracing JIT compilers can choose to specialize machine code based on the exact values of parameters that are only available at runtime, and the language can facilitate "hinting" whether a certain variable should be specialized on its value (e.g. generate machine code for the specific case that i = 42) or on its type (e.g. generate machine code for the general case that i is an integer.) Often a certain operations can be compiled much more efficiently when a parameter is constant, for example the size argument to a memcpy or memcmp, and so a potential "superpower" is to apply these optimizations automatically without having to statically generate code that anticipates each possible value.

Is that a legitimate superpower? what are some others? what tricks can tracing JIT users exploit to leave C/C++ hackers eating their dust? :-)