Skip to content

Instantly share code, notes, and snippets.

/* A facade for the Web Worker API that fakes it in case it's missing.
Good when web workers aren't supported in the browser, but it's still fast enough, so execution doesn't hang too badly (e.g. Opera 10.5).
By Stefan Wehrmeyer, licensed under MIT
*/
var WorkerFacade;
if(!!window.Worker && !iCanHasFastBrowser()){
WorkerFacade = (function(){
return function(path){
return new window.Worker(path);
@gleicon
gleicon / zemitter.js
Created June 9, 2011 02:16
ZeroMQ and Node.js pub/sub emitter
// ZeroMQ PUB/SUB + Multicast (PGM) enabled event emitter for node.js
// gleicon - 2011
var util = require("util");
var events = require("events");
var zeromq = require("../node_modules/zeromq"); // check the right path
var sys = require("sys");
function DistEventEmitter(name, remote_node) {
events.EventEmitter.call(this);
@jpantuso
jpantuso / osx_lion_rail_setup.md
Created July 27, 2011 19:51
Setup OS X 10.7 w/ homebrew, oh-my-zsh, rvm, rails, and MySQL
@mreid
mreid / INSTALL-VW-OSX.md
Created January 29, 2012 21:59
Install Vowpal Wabbit on Mac OS X Lion

The INSTALL instructions that come with Vowpal Wabbit appear not to work on Mac OS X Lion. Here's what I did to get it to compile. You will need the developer tools that come with the XCode installation.

The only dependency VW has is the boost C++ library. So first, download and install Boost

To install Boost, do the following:

$ cp ~/Downloads/boost_1_48_0.tar.bz2 ./
@yusugomori
yusugomori / RBM.scala
Last active December 15, 2015 08:29
Restricted Boltzmann Machine
// $ scalac RBM.scala
// $ scala RBM
import scala.util.Random
import scala.math
class RBM(val N: Int, val n_visible: Int, val n_hidden: Int,
_W: Array[Array[Double]]=null, _hbias: Array[Double]=null, _vbias: Array[Double]=null,
var rng: Random=null) {
@lexander
lexander / gist:5751836
Created June 10, 2013 20:09
YUIDoc overloaded method example.
/**
* A test class file.
* @class TestClass
* @constructor
*/
function TestClass () {
}
/**
@debasishg
debasishg / gist:8172796
Last active July 5, 2024 11:53
A collection of links for streaming algorithms and data structures

General Background and Overview

  1. Probabilistic Data Structures for Web Analytics and Data Mining : A great overview of the space of probabilistic data structures and how they are used in approximation algorithm implementation.
  2. Models and Issues in Data Stream Systems
  3. Philippe Flajolet’s contribution to streaming algorithms : A presentation by Jérémie Lumbroso that visits some of the hostorical perspectives and how it all began with Flajolet
  4. Approximate Frequency Counts over Data Streams by Gurmeet Singh Manku & Rajeev Motwani : One of the early papers on the subject.
  5. [Methods for Finding Frequent Items in Data Streams](http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.187.9800&rep=rep1&t
@harith
harith / shellmarks.sh
Last active September 18, 2022 07:50
Utilities to let you easily reach frequently visited but deeply nested directories.
# Utilities for quickly accessing frequently used directories in bash.
# Usage:
# $ cd /path/to/project/src/
# $ mark code # Will create a new shortcut.
# # Becomes interactive if a shortcut already exists
# # m is an alias for mark. You can also `m code`
#
# $ code # From now on, running this anywhere in the shell
# # will put you in /path/to/project/src/code
<!-- a tm / sublime snippet to go with https://github.com/abeisgreat/-_-.js -->
<snippet>
<content><![CDATA[ಠ_ಠ( ${1} )]]></content>
<tabTrigger>oo</tabTrigger>
<scope>source.js</scope>
</snippet>

using reputation systems to create shared function-critical datastructures in open networks

Search engines, spam filtration, and p2p protocols - all need to rate the value of information. Search engines need it to provide good results; spam filtration needs it to exclude noise; and p2p networks need it for security and efficiency.

What is "value?" I'll use two dimensions:

  • Quality (how useful the information is). The more permissive participation is, the greater the need for quality-ranking. Term-frequency search alone fails for the Web, because the Web doesn't prefilter for quality.
  • Trust (how safe the information is to use). The more function-critical the information is, the greater the need for trust-ranking. If a DHT's host-lookup can be manipulated to flood an unsuspecting computer (a DDoS) then the DHT is unsound.