Promise of Better Concurrency in Functional Programming: Big Win or BS?

Are you ready for the death of Moore's Law?

Or will you be like this guy, threatened with extinction?

http://penanusantara.files.wordpress.com/2010/10/macan-tutul.jpg

Your software no longer gets automatic speedups. But what about the competition? What are they doing with Multicore disruptions? And what about Cloud Computing that brings its own set of headaches?

FP promises solutions that not only make the transition easier, they abstract away some of the worst pains, e.g. callback hell.

You face two choices. You can ignore FP and risk being eaten by your competition. Or you can start learning about it now, to spread your tuition payments.

This talk tells the story of the Before and After of FP concurrency solutions in real-world problems. Examples will be drawn from a swathe of languages including Go, Javascript, and Haskell.

Come and evaluate for yourself whether it's all hype, whether there are real gains to be scooped up, and if so, where are the side-effects and hidden gotchas.

Details

• The event starts with a motion of thanks to Traveloka and a word from our hosts.

• For the next 30-60 minutes we'll investigate what parallelism and concurrency means and how they differ from each other. Examples will be drawn from a swathe of languages, authorities, and web resources. This is active learning -- bring your opinions, sift through the facts, arrive at a deeper understanding.

• Whew! After all that work, let's decompress by taking a look at the controversy between two top engineers in their respective fields.

• On the right we have Bryan Cantrill, awarded for his work on DTrace at Sun, now CTO of Joyent.

http://photos4.meetupstatic.com/photos/event/d/0/2/2/600_413453282.jpeg

• On the left we have Simon Peyton Jones, type hacker and Haskell researcher:

http://photos3.meetupstatic.com/photos/event/d/1/d/a/600_413453722.jpeg

• At the heart of the debate is "How To Do Concurrency Right." The issues raised include: Is traditional lock-based programming too difficult? Is S/TM inefficient, restrictive, overhyped? Who's right?

• Let's take a simple vote at the end to wrap up. No fist fights please :)