(Post)Functional

Now that all contemporary languages have become functional languages,we have to shed the old ways of programming and bow the big lambda.

Day of week:

Wednesday

Location:

Seacliff A/B

Host:

Erik Meijer

Presentations

Deconstructing Functional Programming

4:05pm - 4:55pm

By: Gilad Bracha
Co-author of Java, Creator of Newspeak

Functional programming (FP) embodies two major ideas of great value: the elimination of side effects and the use of first class functions. However, many other ideas are strongly associated with functional programming: laziness, Hindley-Milner style types, ML modules, monads, currying, pattern matching. In this talk we'll focus on the negative aspects of some of these ideas and try to separate the good parts of FP from its unnecessary cultural baggage.

Survivors of the talk will be better positioned to distinguish FP hype from reality and to apply key ideas of FP in non-FP languages.

(Post)Functional Open Space

11:45am - 12:35pm

By:

Reactive REST

1:35pm - 2:25pm

By: Jafar Husain
Senior Developer at Netflix

Increasingly we live in a world of big data and small clients. In the world of distributed systems, all developers must play the role of magician. Developers must create the illusion that all of the data in the cloud is sitting on our user's devices, regardless of hardware or bandwidth limitations. The most powerful trick we developers have at our disposal is the cache. Caching can dramatically improve the perceived performance of a distributed system. However introducing a cache can add a lot of complexity to our applications. How do we avoid caching multiple, divergent copies of the same object? Furthermore how can we add caching to an existing system without requiring large changes to our code? REST (Representational State Transfer) is an architecture that provides answers to these questions. By providing a set of idempotent operations for working with your models, REST makes it easy to add cache layers to distributed systems. In this talk, you'll learn how Netflix uses Reactive programming to build and consume REST endpoints. You'll also learn how to Netflix works around the limitations of the HTTP protocol to create high-performance REST APIs.

Presentations

C++: The Good Parts

5:20pm - 6:10pm

By: Jordan DeLong
Software Engineer at Facebook

Although C++ originated with the primary aim of adding Object Oriented features to C, it has evolved from the OO paradigm into a multi-paradigm language with strong support for various functional programming idioms. C++11 adds several language features, most notably lambda expressions, that are overtly functional. But even prior to C++11, key parts of the standard library sported a design that has fundamentally more in common with functional programming than with OO. This talk gives an overview of the past, current and near future "good parts" of C++'s functional side, through the colored lens of the speaker's biases.

Reactive Extensions in the Cloud -- Clojure, Java, and All the REST of That

10:30am - 11:20am

By: Brian Beckman
Physicist, Symbolic and Functional Programming, Amazon

In bringing up Netflix's RxJava Clojure language adapter, I discovered parsimonious use for all of Clojure's agents, atoms, refs, futures, and promises. The only one of Clojure's concurrency constructs that didn't naturally arise in my application was Dynamic Var. Before this experience, I might have characterized Clojure's collection of concurrency primitives as a grab-bag; after, I realized that it's a minimal, complete set of interoperable constructs (physics analogy: a complete set of commuting operators). In this live-coding talk, I walk through my discoveries in thematic contexts: growing and shrinking collections, implementing distinct and distinct-until-changed, working famous exercises. I end with speculations on mixing Rx with mini-Kanren (a.k.a., clojure.core.logic) and on hosting large reactive computations in the cloud.

The Secret Life of a Mathematica Expression

2:50pm - 3:40pm

By: David Leibs
Senior Researcher at Oracle, VMs and Compilers

Mathematica is an expensive platform for technical computing used by applied mathematicians and engineers. While Mathematica is an immensely powerful and productive system it isn't on the radar of most practitioners of the computer arts. Sadly, the computer science crowd never gets to see the beautiful programming language that is the heart of Mathematica. Of course the Mathematica Programming Language is a functional programming language and a dynamically typed one at that.

Because we are interested in "(Post)Functional" matters I will also expose some of the other features of the language that are like nothing else in the world of programming. I hope to give enough of a taste of Mathematica to encourage software developers to give Mathematica a try for their prototyping and algorithm design.