show and tell for hackers in DC.

Round 19: Semi-Automatic Semaphore

19 Apr 2015

Here are the Hackpad presentation notes from Round 19!


Speaker #1 - Reed Spool

Created a new esoteric programming language: Concrete. Named came from design, there is nothing abstract about it. I wanted to make a programming language that would be similar to other programming languages and be practical. Usually they make programming more abstract and less practical. Whitespace is important, but not quite like in Python, Written completely in Node.js

Q: Is it Turing complete? Don’t know but trying to show it.

Other Programming Games: http://www.kongregate.com/programming-games

My favorite of those: http://www.kongregate.com/games/jahooma/jahoomas-logicbox

Speaker #2 - James Leopore

Self-driving ATV

Processor: Brix Pro All-in-One, everything else is USB communicating via serial commands. Did the electronics in Upverter. Designed custom PCB, emailed the design to a factory and got it shipped in 2 weeks, only cost $14 to produce a stack of the first boards. $100 for the newer, fancy ones.

Speaker #3 - Chrissy Ziccarelli

Wrote an awesome twitter Bot, @findnickisverse Tweet at the bot with the name of a song Nicki Minaj is featured on and the bot will tweet back with a youtube link starting at the correct timestamp. Handles remixes, uses python, the tweepy module, and is hosted on webfaction in a virtual environment.

@findnickisverse in action:

Speaker #4 - Travis Hoppe

Automagic-API!

Slides are here.

Got tired of doing repetitive web scraping, so he wrote a Python program to automate scraping; connected it to d3 to do visualization to help you build the API so you can then query an API that you’ve then created. Look at the Jeopardy tree!

Right now it creates a giant nested JSON list – better than a page full of HTML, but there’s still work to do in flattening the list. Run it on fixed static HTML, remove components that you don’t want, use visualization to make it easier to select the information to keep, then it creates the API.

Speaker #5 - Fletcher Heisler

Q-learning: value iteration AI. Taught my computer to play tic-tac-toe. Wrote in python. Computer learned that sometimes the bast strategy when playing to win is not to play at all (with min-max strat).

Q: Train it on real human games? Maybe someday…


Thanks to everyone who presented, everyone who attended, @LauraNLorenz for the pictures, and of course thanks to our favorite WeWork for hosting! Round 20 is already scheduled for Monday May 11th, 2015, so RSVP and sign up to present!