#!/usr/bin/env python # -*- coding: utf-8 -*- import string, random goal = "To be or not to be, that is the question" gen_size = 100 gen_count = 100000000 bases = ''.join(set(list(goal))) # Our Adam/Eve seed = {'gene': ''.join([random.choice(bases) for i in xrange(len(goal))]), 'fitness': 0} print 'Start: ' + seed['gene'] + "(" + str(seed['fitness']) + ")" # For each generation for i in xrange(gen_count): children = [] # Make each baby for j in xrange(gen_size): child = {'gene': '', 'fitness': 0} # Copy each letter for k in xrange(len(seed['gene'])): nextbase = seed['gene'][k] # Mutation! if random.random() > 0.999: nextbase = random.choice(bases) # Add base to gene and calculate fitness boost child['gene'] += nextbase if nextbase == goal[k]: child['fitness'] += 1 # Append each child to the generation children.append(child) # Choose a winner newseed = max(children,key=(lambda x : x['fitness'])) if newseed != seed: print '#' + str(i) + ' ' + newseed['gene'] + \ "(" + str(newseed['fitness']) + "/" + str(len(goal)) + ")" seed = newseed # Finish if we're done if seed['gene'] == goal: break

Start: iTmZ bReNyaQDtJWPbBkZSQmJPYNjCerARSqEPNE(0) #0 iTmZ bReNyaQDtJWPbBkZSQmJPYNjCerARSqEPNE

<site title="John Doe" template="default.tmpl" css="main.css"> <page link="" name="Home"> <header>My Home Page</header> <content> <p>Hello World!</p> </content> </page> <page link="Contact me"> <header>To contact me</header> <content> <p>choose a medium</p> </content> <subpages> <page link="Email"> <header>Email</header> <content> <p>superhappyfuntime@me.com</p> </content> </page> <page link="IM"> <header>IM</header> <content> <p>IM superhappyfuntime</p> </content> </page> </subpages> </page> </site>

There's nothing here!

My Home Page

Hello World!

To contact me

choose a medium

superhappyfuntime@me.com

superhappyfuntime

MatTex - Embedded Matlab in LaTeX

This was a quick project started and finished at a Super Happy Dev House. The gist is, that I (used) to do a bunch of problem sets which required a combination of computation and explanation. If I can do something on a computer, I will, so I end up writing most of my problem sets in LaTeX. I thought it would be great if I could combine the two into one process, minimizing the amount of copying and pasting required.

PHP's inlining into HTML is quite convenient for quick and dirty jobs (as much as I hate actually coding in PHP), so I thought I would come up with a similar scheme for Matlab and LaTeX and MatTex was born. Here's a contrived, but illustrative example:

\documentclass{article}
\begin{document}
    Hello World.  Recently I found out that 1 + 1 =
<?ml
1 + 1
?>
\end{document}

A particularly handy function in Matlab is “latex”, which formats symbolic expressions in LaTeX. One example of this is printing a matrix:

\documentclass{article}
\begin{document}
    \begin{displaymath}
<?ml
latex(sym([1,2,3,4;5,6,7,8;9,10,11,12]))
?>
    \end{displaymath}
\end{document}

Matrix Example

For anyone who’s done matrices by hand in LaTeX, this is a lot less painful.

It's all quite a hack - It’s more or less a bash script which runs some perl regular expressions to parse out the Matlab, runs the Matlab, uses python to paste the Matlab back into the LaTeX (I would use perl, but my knowledge does not extend beyond regex one-liners), then runs your LaTeX distro of choice (I use TeXShop on my MacBook).

If you want to take a look yourself, I’ve put the code up at github under an MIT License. A word of warning, I coded it amidst heavy socializing all afternoon, so its not exactly my best work. At the moment it requires that the ‘<?ml’ and ‘?>’ reside on their own lines. You’re welcome to fix it though and commit said fix to github!

The syntax is as follows:

./mattex.sh [your tex file]

ImpTran - A smith chart in a script

ImpTran is a little script I wrote a few months ago for impedance transforms. What the heck is an impedance transform you ask? Let me attempt to explain in a completely non-technical way:

Say you are standing outside of a large croud of people and you want to get inside the crowd for whatever reason. It’s a lot easier to move around outside the crowd than inside the crowd simply because there are more people inside the crowd. In the first attempt to get into the crowd, you run at it only to be bounced back, left still outside the circle and frustrated. What could you do to make it easier to get into the crowd? Well you could request that people near the edges leave, making the gradient into the crowd softer. This however would require people leaving, which is no fun at all. You could, however, require that people shuffle around and allow people into the crowd at a specified rate, this way people entering the crowd would face no resistance in the transition in.

Now, there are a lot of inconsistencies in that anology, but the general picture is there. Often an analogue electronics you want to reduce reflections between certain circuit elements (the crowd and the not crowd), and you want to do this by dissipating as little power as possible (ie. not asking people to leave). So what you can do is, instead of using resistors, which burn power, you can use special configurations of capacitors and inductors to “match” the impedance (resistance with an imaginary component) between different elements.

Now normally, figuring out how to construct these capacitor and inductor circuits required the use of a Smith Chart. A Smith Chart allows you to do a lot of complex math (and by complex I mean imaginary, not necessarily hard) by just tracing some arcs. It will tell you, for example, when to add an inductor in series and a capacitor in parallel. Trouble is, when doing actual impedance matching in lab, things never work. You measure the impedance of an antenna input to be 25 - 20iΩ, and you trace some lines and solder some capacitors and inductors, remeasure the impedance of the match and invariably it has destroyed some other match somewhere else in the circuit. Next you iteratively, match back and forth countless matches until you magically arrive on one that works and doesn’t disturb the other matches.

In the world of perfect models, you wouldn’t see such unpredictable changes, but the world isn’t perfect and so me, again being lazy, decided to write a script to do the Smith Chart stuff automatically given an impedance. It took me a few hours, and in the end there is some serious code reduplication (thanks to matlab’s insistence on making a file for each function, which I refuse to do), but it works. One future feature would be to make the script perform a match with only a single part, because often that’s all you have room for and/or really need.

At any rate, the syntax looks as follows for the above 25 - 20iΩ impedance:

EDU>> imptran(25,-20)
    ans =
First add inductor in series of 2.864789e-07H,
 then add capacitor in parallel of 1.273240e-10F
OR
First add inductor in series of -3.183099e-08H,
 then add inductor in parallel of 3.183099e-07H

You can find the code here.

Ben elsewhere on the internet

You can find me on	Twitter
	Facebook
	HN
	Stack Overflow
	Yelp
	Reddit

Ben is	co-founder of Aria Glassworks, an augmented reality startup
	formerly part of Yelp's Mobile Team, the creator of Monocle AR
	a graduate of the EE BS program at Stanford University
	formerly co-artistic director of Stanford Swingtime
	an alumnus of Microsoft Research Asia
	an alumnus of SwayLaw
	an alumnus of the Rotary Youth Exchange Program to Taipei
	on twitter @newhouseb

Past Talks
2/6	CS193P	iPhone Application Development: Yelp Monocle (Stanford iTunes U link)
4/21	ARConf	Sci-Fi to Sci-Fact: How Computer Vision Will Change AR and the World
4/24	VTM	iPhone Sci-Fi: What's Possible and How with Augmented Reality
6/2	ARE	Building Yelp's Monocle

Source		Replacement
<div id="foo" />		<div id="foo">bar</div>
<foo />		bar

The Problem

The Solution

A Hello World

Features

History

Only three easy steps!

If You say: But I can only use static HTML!

A Static Example: My Stanford Web Space

Ground Rules

Document Structure

<site>

<page>

<subpages>

<*> (a.k.a variable tags)

Parsing Stages

Variable Initialization and Scoping

Automagic Variables

Recursive Variable Substitution

An Example