See the effects that small changes in environment have on evolution

I’ve been studying artificial intelligence for many years. One of the AI constructs that has fascinated me the most has been Genetic Algorithms (GA). With a GA, one “evolves” a solution to ones problem. A “gene” is a candidate solution to a problem, and individual alleles in that gene are individual parameters to the function that attempts to solve the problem. The output of the function is evaluated for “fitness”–essentially, how well it solved the problem–and good candidates are intermingled for the benefit of future generations, while particularly bad candidates are discarded; “survival of the fittest” in its most incredibly literal sense.

This little sketch was written for a couple of reasons. First, I’ve been seeing some AI code that I felt like has been overly complicated. A lot of principle AI algorithms are quite simple to implement, once you understand them. But also, I wanted to demonstrate the flow of a GA and how it tends to find intermediate solutions, improving over time.

This sketch will require some basic programming knowledge to be able to alter the fitness function and make the GA do different things, but I think it can be an exciting and compelling exploration into code. There are actually a few simple things one can do to alter the course of the algorithm. Each “gene” represents a row of pixels in the image being drawn when you click “start”. As it currently stands, it will “evolve” solutions converging on the color red. If you change “if((i%4)===0)”* to read “if((i%4)===1)”, it will converge on green. “if((i%4)===2)” converges on blue. There is a 4th opacity component at #3, but it basically just ends up showing black. It isn’t necessary but I just didn’t feel like fixing the code to get rid of it, and maybe someone else will find it useful.

Instead of “if((i%4)” you replace the 4 with another number–say a prime number like 31–you get some interesting results as well. There are a lot of things that you could do here, it’s just a matter of whether or not you wish to pluck at it.

* the percent sign, ‘%’, is known as the modulus operator. For positive numbers, it returns the remainder of a division operation. So, “13 % 5 ” return 3 because 13 divided by 5 is 2 remainder 3 (i.e. 2 x 5 + 3 = 13).

A few years back I received a birthday card that, when opened, played a 15-second clip of The Final Countdown. Before tossing it out, I tore the little board and speaker from the card. I knew nothing about electronics at the time, but I just thought it was worth keeping around.

The audible gates of Hell.

Well I found the little device in one of my junk containers last week and immediately recognized it as something to be circuit bent. The board had an IC, a capacitor, and two resistors. As I learned from Nicolas Collins’ excellent Art of Handmade Electronic Music, one of those resistors was the freakout resistor. Following his “laying of hands” technique, I licked my finger and ran it around the board while the circuit was on. This actually worked very well as I immediately found the resistor that caused the sound to go all wonky.

The basic idea here is that the circuit operates at many cycles per second, at a rate determined by some resistor. If you change the value of the resistor, that rate will change and the audio clip will either speed up or slow down; this is the basic idea behind circuit bending. When you run a damp finger across the board, you’re making random connections and changing how components interact with each other. When you short out the clock resistor, it sounds like hitting the circuit’s main nerve.

So now with the clock resistor found, the next step was to replace it with a potentiometer so that the clock speed could be manipulated with a twist of a knob. After removing the resistor from the board, I measured its value as ~400k and set out to find a potentiometer in the same range. I was very happy to find a 2 megohm pot in the parts hole at the space since it would slow the clock quite dramatically, producing some really cool roaring depths-of-hell sounds.

Trimpot between output wires.

The board had an integrated 3V watch battery which I replaced with a 2-AA holder. There was also a mechanical contact on the board to turn the circuit on when the card was opened; I replaced this with a simple toggle switch. I also thought it would be cool to switch back and forth between normal mode and circuit-bent mode, so I used a DPDT switch to toggle between the original clock resistor and the circuit-bending potentiometer.

Lastly, I wanted to add 1/4″ output to the circuit. That way it could be amplified, distorted and subject to all sorts of effects and processing. I cut the wires leading to the little speaker and soldered them to the signal and ground leads of a 1/4″ output jack. This arrangement alone would have been far too much to run into an instrument amplifier, so I soldered in a trimpot to act as a volume control (and/or load resistor taking the place of the speaker) between the output wires.

Guts.

I neglected to consider an enclosure for this project. However, I did notice an alarming surplus of queso dip at the space. Since apparently nobody likes queso, I emptied a jar, cleaned it out (probably the hardest part of the whole project) and drilled some holes in the lid to accommodate my components. A coating of white spray paint covers the Tostito’s logo for extra class. I actually really like this au-natural enclosure since you can see all the components and wiring just hanging there. Since nothing’s fixed down, I hot-glued all my solder connections to the board. I think I’d like to experiment with glass jar enclosures again, perhaps with some lighting involved next time.

In the process of building and testing I noticed that when I would touch the third, unused lug of the pot, I could get some cool scratchy AM Radio static sounds from the circuit, especially with the pot wiper all the way over in satan mode. So as a last minute little addition, I soldered a bare wire to the lug and ran it outside of the jar to retain this touch-sensitive flavor of noise.

And that’s how you make Dan’s homemade queso dip.

A few of us will also be presenting at this event! We will have live 3D sugar glass printing, exhibitions of Brendan’s boom cases, Dan’s 8 mm RockBox, PJ’s electronics, Corrie’s textiles and artwork, Chris Terrell’s wood burning, and maybe a few more things.

We hope to see you there! Deets and directions below.

Deets:
Free Refreshments (wine, food, music) will be provided at the event!!
APS Requests your RSVP HERE: museum@amphilsoc.org
Thursday, June 6th, 2013
5:30-7:30 pm
APS Museum in Philosophical Hall
104 S. Fifth Street, Philadelphia, PA

View Larger Map

But it is a little rough around the edges. While it has a lot to make it a very easy language to learn, it is ultimately meant for computer scientists, those in training and those in working. There is an underlying feeling to everything that “this easy thing will eventually get harder.” As I see more and more inside the Racket community itself, I know that that is not their intention, and that they hope to be able to bring the joy of programming to everyone, regardless of their background.

A tool that has brought the joy of programming to everyone has been Processing. Processing, in a round-about way, brought a lot of features of the Racket[1] environment to Java. There are some very, very clear parallels between the DrRacket programming environment and the Processing editor. If this was unintentional, then it at least clearly indicates the superiority of the form as a pedagogical tool, as two independent environments have both evolved the same feature. If it was intentional, then there are ways in presentation in which Processing has grown past Racket that I think could be brought back to Racket-land.

Racket and Processing

At a very high level, you can see the similarities between the DrRacket editor and the Processing editor. I have intentionally made the code samples different for each, as they represent more of the canonical methodology for each language[2].

A checkboard. Left: DrRacket. Right: Processing

As an artist with a very strong background in computer science, there are things about Racket that appeal to me that Processing will never be able to do[3]. There are things about Processing that are very awesome and very nice that Racket could do, if someone with a very strong background in computer science took the time to make it so.

Where Racket succeeds over Lisp and Scheme, and where Processing–and thus Java–succeeds over all, is that enough is provided for you to not require you to learn the more advanced language features to get good results. Racket calls it “batteries included”, and it is, I believe, the most important concept that Processing borrows from Racket. Beginners don’t want to learn about “public static void mains”; to this day I still don’t know why that garbage is so necessary.

My hope is to try to bring some of the batteries of Processing back to Racket. The potential is there for Racket to be every bit as popular with artists as Processing is, today. The things that it lacks in comparison to Processing are relatively trivial to develop, in relation to the things that Processing lacks in comparison to Racket. By bringing artists into the much more powerful environment of the Racket programming language, once they have mastered the fundamentals of programming with the training-wheels mode, Racket can be molded by artists, for artists, to let people think and program in ways that the programmer-artist chooses, not in ways that the original language designer chose for them. If there are two groups of people that I know hate being told how to think more than anyone, it’s artists and programmers.

I’d like to hear from the Hive76 readership, especially those who have experience with Processing (though all are welcome to comment), about what you like and don’t like about Processing, about programming, about art, about putting code down in text files, etc. This certainly applies to the Arduino crowd, as well, as Arduino is based on Processing. In the meantime, I have a few thoughts on where the work could begin.

Make the website prettier

First impressions are important, and most programming language websites are prettier than Racket’s. I would more readily crib from Ruby than from Processing here. Ruby’s website is very clean and concise, very inviting and pleasing. Everyone cares about taste, especially artists.

Racket’s website is built in Racket, with a sub-language called Scribble, specifically designed for building documents and documentation. It’s an excellent language, but its default styling is kind of ugly. Restyling the site is a low-hanging fruit that could bring in more people, if only because they don’t jump ship prematurely.

Make more dynamic examples for beginners

The Racket documentation has its heart in the right place, but ultimately feels like its drawing examples are just pat examples. The drawings aren’t very interesting and its not immediately obvious how one gets to more interesting things. See this page for an example. It’s easy to draw a smiley face in Racket, it’s much harder to draw a good looking smiley face. And before it becomes apparent, the documentation dives too quickly into the computer science topics thereafter. In contrast, Processing starts off with examples that are extremely visually compelling. As people are visual creatures by nature, this seems the superior methodology. Note that the Processing example is the introductory page to the documentation–one click off of the front page–whereas the Racket example is several pages into the documentation–two clicks and several screenfuls of scrolling off of the front page. Much more compelling examples in Racket are much more deeply hidden.

Finally, make more artistic things with Racket

The Processing website is built around showcasing what people have done with Processing. While Racket’s package system can be considered the same thing, to a degree (and it is already in the process of getting better), it is again not as visually compelling. Racket needs an elevator pitch, a high level view that shows the answer to “what will you do with Racket?”.

Footnotes:

[1] Historical Note: Racket has been around since 1994, when it was originally called PLT Scheme. They figured they lost more people being immediately associated with Scheme and Lisp than they gained, so in 2010 they changed the name to underscore the fact that there is so much more to Racket than the name “Scheme” implies. Suffice to say, for beginners it doesn’t matter, other than knowing that their is history to the project and it isn’t going away anytime soon.

[2] There are problems in both examples; both are overly simplistic and both would not scale well for large applications. But the gist of the difference is there, Racket is about functions that define what we want to happen, Processing (because it is based on Java) is about instructions. The differences to the non computer scientist are difficult to explain, so for now you will just have to take my word for it that the former is a generalization of the latter and is therefore more robust and open for greater expressiveness.

[3] For the computer scientists reading, succinctly: macros. You either know what I mean or you don’t, this is not the place to get into a technical description of macros. For the non-computer scientists, the best analogy I can make is magic. Magic isn’t real in our world, but I can think of nothing that comes closer to resembling it than the massive productivity gains that macros can afford, once mastered. But neither is mastery of macros in Racket required to do great work. Processing is what happens to languages like Java that do not have macros. In Racket, something like Processing would stay a part of Racket and not be a separate project.

I got some time to explore the Raspbian distro.

After seeing all of my efforts, Morfin couldn’t wait to give it a shot.

Eventually we got my favorite light-weight print controller github.com/kliment/printrun running an active 3D print. It really was incredible to have a $40 computer connected to the interwebs and sending gcode with a full GUI over python->USB-serial. It’s a bit too slow for computational slicing, but would probably be GREAT for a bot-farm. Note that you should also use pianobar instead of full-blown pithos for pandora audio. Note that the audio worked great after we ran the apt-get upgrades mentioned above.

And remember to grab our desktop background! It’s only 1.2 MB.

With RAMBo bypassing the stock motherboard we can drive the ShopBot to scary speeds (10x faster in XY and 100x faster in Z.). Precision should also be ~10x better than belt-driven motion, but needs more fine tuning.

and did i mention it was freaking awesome?

My ShopBot RAMBo Marlin firmware branch is available via GitHub (of course). Follow along in the GitHub log to understand our process.

Thanks to ShopBot and Ultimachine for all your help and schematics!!

So…I put Sean’s Harrow through it’s paces, and here are some of my newest desktop backgrounds.

First you get the basic render down. I use 32-bit color with OpenEXR file format, saving z-buffer info too, just, you know, in case you need it later. You should get something like this… kind of flat when viewed on a crappy computer screen, but i assure you there is a ton of color info there for fine tuning later.

With all that extra color depth, you can easily fine tune the contrast, like so.

Then you need to come up with the shadow version… Andrew Price from BlenderGuru does this in a new Scene. I like the stark contrast… when you look carefully the sharp edges tell you that you are looking at a perspective view.

Put it all together, and you get a softly-back-lit Logo. WIN!

Making things to make music.

On Thursday April 25, our series of events for Philly Tech Week continues as we open our doors for DIY Music Night (5pm-???). If you’re into music, making music, or making things that make music, you won’t want to miss it! If you’ve been to the space before, you’ll know that we run on a steady diet of tunes. And on Thursday, we’ll have all our audio and music-centric projects out in what is sure to be the loudest night of PTW. Come by and see the space, make some amplified noise, hang out, or share your own projects.

We’ll have guitars, amps, synthesizers, sequencers, speakers, fuzzboxes, tremolo pedals, signal generators, oscillators, speakers, drum machines, pickups, karaoke machines and probably alot more – all made at Hive76.

Plus we’ll have a handful of contact microphones to give away! We’ll help you turn anything into an amplified electric instrument in 10 minutes flat.

On Wednesday, April 24, Hive76 will be transforming our typical Wednesday open house into a special Tech Week Open Hack!

Over the years, we’ve managed to accumulate a plethora of electronic detritus, as one might expect from a hackerspace. We’d love to invite you to come to our space and help us clean it put the junk to good use! We’ll have some microcontrollers on hand and several members around to help you along your way. The idea is to split attendees out into workgroups that will concentrate on designing and building objects from parts we have available for you at the space! Wanna throw together a bot that responds to table tapping? How about something that plays like a musical instrument? Dig making LEDs respond to external stimuli? Just wanna learn how some of these things might be possible? Slightly annoyed with how often I ask questions in my posts?

—Please bring a laptop along if you have one available. Some programming experience is a plus, but definitely not necessary.—

Wednesday 4/24 5pm – 10
Gratis and Libre (free)

RSVP by commenting below.

***ALSO: If you haven’t been to Hive76 and want to see what we’re about, stop by on Monday April 8th for our Monthly Monday Microcontroller Madness (7-10PM) or on any (other) Wednesday night for our weekly open houses (7-11PM)***

Our grand finale/blowout/party for Philly Tech Week will take place on Saturday, April 27. Everything from the previous week’s events will be on interactive display, and more. It’s an all-day exhibition of everything that goes on at Hive76 and everybody’s invited!

• 3D Printing/OpenSCAD interactive demo
• Fighting robots
• DIY musical instruments, effects, snyths, amps, circuit bending, everything
• Continuous open hack/project marathon: make whatever you want with whatever we’ve got
• Open soldering workstations
• Member’s signature projects on display
• Gaming, arcade and pinball machines
• Music
• Karaoke!
• Libations
• Free food

We are located in the Arts Building at 915 Spring Garden Street. Stop by at any time during the day/night and dial 0519 at the callbox. See you there!