Code Mesh 2015
who it's for
backend engineers, functional programmers, language designers, systems programmers, game developers, site reliability engineers, devops engineers
topics
Julia JIT specialization; D for low latency finance; Erlang message-passing games; conflict-free replicated data types; PouchDB/CouchDB sync; event sourcing and CQRS; container schedulers (Kubernetes); core rope memory and AGC software; bounded queues and backpressure; unikernels and hyperelastic clouds; HDF5 with nanomsg; Riak and Erlang event analysis
| title | author | length |
|---|---|---|
| How Julia Goes Fast | Leah Hanson | 43m |
Claims that Julia hits near-C speeds by combining JIT compilation, aggressive specialization and best-effort type inference and demonstrates concrete tools: a… | ||
| D: a Practitioner's Perspective - Productivity _and_ Performance | Laeeth Isharc | 43m |
Argues that D is practical for finance: prototypes compile fast with DMD, production uses GCC/LLVM back ends, and processed tens of GB without choking where… | ||
| Synchronizing Game Components | Robert Virding | 48m |
Argues that scaling games needs isolated processes and message-passing rather than shared mutable state, implements it in Erlang with a small L-like language… | ||
| From Irrational Configuration System to Functional Data Store | Rob Martin | 52m |
Begins with 42 Oracle tables, 5200+ configuration values, a 1.5 million–line Java legacy and onboarding that runs up to 10 months and claims anyone who touches… | ||
| Keynote: Grace Murray Hopper: The Original Pirate Hacker | Melissa Pierce | 52m |
Claims that Grace Murray Hopper builds the first operational compiler in 1952, writes the Mark I manual, invented the patch to fix punched tape, popularizes… | ||
| The Thrilling Adventures of Lovelace, Babbage and the Analytical Engine | Sydney Padua | 43m |
Demonstrates 3D animations and a worked‑out elevation of Babbage’s Analytical Engine, claiming 50 decimal places of storage, three punch-card types (number… | ||
| An Introduction to Schedulers | Kushal Pisavadia | 37m |
Schedulers are presented as concrete, repeatable systems: a queue a lock and a selection algorithm that implement concrete steps (schedule, allocate, bind)… | ||
| CRDTs in Practice | Marc Shapiro, Nuno Preguiça | 50m |
CRDTs give always-accept updates with deterministic merge rules and actual libraries (React implements an observe-remove set used in production at a major… | ||
| The Charming Genius of the Apollo Guidance Computer | Brian L Troutwine | 49m |
The Apollo Guidance Computer runs at ~1 MHz with 15‑bit words (one parity + 15 data), about 248 words of RAM and ~36K words of ROM implemented as core rope… | ||
| Putting You Back in Charge of Your Data | Matthew Podwysocki | 47m |
Thali claims to put data back on the phone by shipping a peer-to-peer stack that uses Bluetooth, BLE and Wi‑Fi Direct for discovery, Node via JXcore + Cordova… | ||
| Keynote: Propositions as Types | Philip Wadler | 55m |
The principle propositions as types asserts a tight isomorphism: natural deduction ↔ typed lambda calculus, proofs ↔ programs, normalization ↔ evaluation… | ||
| Everything Will Flow | Zach Tellman | 45m |
Claims that unbounded queues are fundamentally broken and that back pressure plus clear limits beat opaque buffering. Shows tools and code: Clojure libraries… | ||
| Spreadsheets are Code | Felienne Hermans | 44m |
Claim: spreadsheets are Turing complete — a full Turing machine is implemented in Excel formulas and visualized, and a working selection sort in Excel is built… | ||
| What NOT to do when building large scale systems in Erlang | Chandru Mullaparthi | 41m |
Practical anti-patterns and concrete fixes for large Erlang systems: amnesia fails on write‑heavy replicated clusters and net splits cause data loss unless… | ||
| Coordination-Free Designs for Mobile Gaming | Christopher S. Meiklejohn | 44m |
Claims that coordination can be minimized by designing with CRDTs and a dataflow/lattice model, implemented as the Lasp programming model embedded in Erlang… | ||
| Accidentally Concurrent | Evan Czaplicki | 38m |
Claim: nearly every program becomes a concurrent system unless its interfaces prevent hidden channels. Uses message-passing concurrency (channels + spawn) to… | ||
| Function-Passing, A New Model for Typed, Asynchronous and Distributed Programming | Heather Miller | 40m |
Claims a concrete runtime model called function-passing that inverts actors by keeping immutable data stationary in SiloRefs and shipping typed, serializable… | ||
| Events Analysis with Riak, Perl and Erlang at Booking.com | Damien Krotkine | 46m |
Booking.com stores >50k events/sec, keeps 8–10 days hot and ships older data to Hadoop, aggregates per-second event blobs serialized with Sereal serialization… | ||
| Exercise Analysis | Jan Machacek | 44m |
Muvr builds a full pipeline: wearables (Apple Watch, Pebble) stream accelerometer and heart data through mobile to an Akka-backed cluster and Spark training… | ||
| The Road to Running Haskell at Facebook Scale | Jon Coens | 44m |
Facebook runs Haxl, a Haskell DSL that replaces a 100K-line rule engine, and migrates that codebase into GHC over about 15 months of rollout. The system hits… | ||
| Heresy! Combining Type Systems with PHP | Josh Watzman | 44m |
Explains how Hack actually puts a strong, static type system on PHP, with concrete mechanisms like unresolved types, fake locals, flow-sensitive tracking and a… | ||
| Contemporary Approaches to Data at Scale | Ben Stopford | 42m |
Explains that storage performance hinges on sequential vs random I/O and compares three concrete engines: in-memory indexes (fast until indexes exceed RAM)… | ||
| Concurrency + Distribution = Scalability + Availability (...) | Francesco Cesarini | 45m |
Erlang/OTP is an enabler not magic: prototypes and numbers decide. When nodes are fully connected performance and monitor overhead start limiting clusters… | ||
| Panel Discussion | Tony Hoare, Bruce Tate | 51m |
A ten minute historical run through claims that added language constructs aim to push errors to compile time and isolate them at run time leads into a panel… | ||
| Communication, Expression, and Types: Your Compiler as an Ally | Allele Alelle | 34m |
Shows concrete Haskell work: Haskell live coding in GHCi with GHCi + Emacs + ghc‑mod and fast type‑check cycles around 100 ms. Builds types (Person, Color)… | ||
| VR Best Practices | Andreia Gaita | 50m |
Covers concrete constraints and fixes: Unity integration and Gear developer mode for building and profiling, device specs like 2160×1200 (1080×1200 per eye)… | ||
| Registers: Systems of Record with Guaranteed Integrity | Philip Potter | 40m |
Prototype registers provide a single source of truth with cryptographic integrity for structured UK government data they model everything as an append-only… | ||
| Modern Abstractions and the Web, and avoiding Pitfalls because Life is short | Irina Guberman | 37m |
Compares concrete tradeoffs: Erlang style process isolation gives each request its own heap so a misbehaving task can be killed and restarted, Elixir + Phoenix… | ||
| A Vision for Relational Programming in miniKanren | William E. Byrd | 44m |
Implementation details dominate: miniKanren code runs in a web-based veneer REPL, defines a lambda-calculus syntax checker that also generates every syntactic… | ||
| An Evolution of a Betting Engine at William Hill | Peter Morgan | 44m |
Claims concrete scaling numbers 464 bets/sec, 5,000,000 price changes/day and 160TB/day and a working stream architecture built in Erlang with Apache Kafka and… | ||
| Practical Lambda Calculus | Einar W. Høst, Jonas Winje | 42m |
Builds actual Church numerals and implements successor, addition, multiplication and a predecessor via pairs that counts down; constructs lists… | ||
| The Free and The Furious: And by 'Furious' I mean Codensity. | raichoo | 44m |
Claim: applying the Codensity transformation fixes specific left-associated bind blowups in Free Monad programs. The talk demonstrates replacing naïve… | ||
| My Little Pony | Darach Ennis, Sylvan Clebsch | 50m |
Deep dive into Pony and the Erlang ecosystem with concrete results: microbenchmarks show Pony beating C (C ~34% slower) and Java (~7.8x slower), tail-recursive… | ||
| Reactive Design Patterns | Roland Kuhn | 42m |
Argument: reactive systems demand different design — partition state, decouple components, and make communication message-driven. Concrete claims: elasticity… | ||
| Into Production | Jamie Winsor | 40m |
Built an Elixir based online game backend for Moonrise running on Elixir 0.9.0, with routers, lobby/world/combat services and a Postgres datastore sharded into… | ||
| Simplifying Scalable Cloud Application Development with Go | Mandy Waite | 46m |
Claims Go makes large cloud systems simpler and faster, citing Docker and Kubernetes and parse.com moving from Rails to Go to "save sanity." Concrete demos… | ||
| Depending on Types | Stephanie Weirich | 45m |
Argues that Haskell already supports a lot of dependently typed programming by using GADTs, datatype promotion and type families to encode invariants… | ||
| Functional Programming on the Web | Becky Conning | 47m |
Claims that moving from JavaScript patterns to PureScript yields clearer, purer effects and fewer bugs, and demonstrates an actual workflow: compile-to-JS… | ||
| What's New and Exciting in Julia | Stefan Karpinski | 43m |
Live demos show concrete claims and numbers: Julia 0.4 introduces callable objects (overloading call), docstrings, fast package loading and a Jupyter workflow… | ||
| Boot my (secure)-(portable) clouds! | Nicki Watt | 50m |
Claims ability to spin up secure, portable environments across clouds and re-create a full tooling stack after failure in about 40 minutes. Implementation uses… | ||
| Transactions: Myths, Surprises and Opportunities | Martin Kleppmann | 52m |
Defines ACID precisely, shows read skew and write skew with concrete examples (bank transfer, ambulance on‑call), and demonstrates why two‑phase locking is… | ||
| Beyond Lists: High Performance Data Structures | Phil Trelford | 45m |
Claims concrete wins: replacing naive arrays/lists with an unrolled linked list boosts throughput from ~100 trades/sec to 10,000 trades per second while… | ||
| The Functional Tao of Bash | Garrett Smith | 48m |
Claims Bash isn’t a functional language but functional techniques apply: the presenter refactors a 140‑line CI.sh, extracts blocks into named functions… | ||
| JVM Profiling under the Hood | Richard Warburton | 44m |
Profiling claim is that sampling reveals hotspots but default JVM TI safepoint sampling hides them and can massively mislead developers, for example 63%… | ||
| Harmful GOTOs, Premature Optimizations and Programming Myths (...) | Alvaro Videla | 45m |
Argues that well‑worn maxims hide context and cites concrete evidence: 3% of code accounts for most runtime, so profile before optimizing; historical papers… | ||
| Adventures with Concurrent Programming in Java: A Quest for Predictable Latency | Martin Thompson | 47m |
Claims concrete latency and scalability tradeoffs, measured not promised: 167 ns baseline for a lockfree one-to-one ring buffer, blocking condition-variable… | ||
| Unikernels and Hyper Elastic Clouds | Amir Chaudhry | 44m |
Claims that unikernels cut attack surface and boot fast with concrete demos: a MirageOS static site becomes a deployable unikernel via Mirage skeleton and… | ||
| Keynote: Why Functional Programming Matters | John Hughes, Mary Sheeran | 58m |
Claims that functions can replace booleans and integers via Church encodings, that lazy infinite lists let producers supply whole values on demand for… | ||
| FRP and Functional Game Programming | Elise Huard | 26m |
Explains concrete results: a casual game built in Haskell + Elerea, running a 60 fps game loop with signal generators, dynamic per-bullet signals and level… | ||
| How to destroy your work (and why you should) | Dan Hett | 39m |
Rapid prototyping leads to concrete claims and failures: a cardboard-first prototype turns into a carpentered scribbler animatronic pencil that prints tweets… | ||
| nothing matches | ||