I want to play around with ARToolKit. My plan is to take the pattern detection information and ship it off to places for further processing. I figured the easiest way to do that would be to use Open Sound Control. Since ARToolKit is a C framework, I though it would make sense to use an OSC library written in C. oscpack looked like a good candidate.

I’m using a Intel-based Mac. To build and install the oscpack library I needed make a few changes to the Makefile. I thought I would share these changes in case anybody else tries to do this.

Step zero is ONLY NECESSARY IF YOU ARE RUNNING ON A POWER PC-BASED MAC. Sorry to shout, I just want to emphasize this. I didn’t need to do this, but I’m pointing it out anyway in case it helps somebody. At the beginning of the Makefile, change

ENDIANESS=OSC_HOST_LITTLE_ENDIAN

to

ENDIANESS=OSC_HOST_BIG_ENDIAN


This next step, step one, is only necessary if you don’t want to create (or have not already created) a /usr/local directory. It doesn’t exist by default, and if it isn’t there then the installation process will fail. You need to change

PREFIX = /usr/local

to

PREFIX = /usr


The thrid step involves using the right compiler syntax to set up a shared library. Change this part of the code

@#GNU/Linux case
$(CXX) -shared -Wl,-soname,$(LIBSONAME) -o $(LIBFILENAME) $(LIBOBJECTS) -lc
@#Mac OS X case
@#$(CXX) -dynamiclib -Wl,-install_name,$(LIBSONAME) -o $(LIBFILENAME) $(LIBOBJECTS) -lc

to

@#GNU/Linux case
@#$(CXX) -shared -Wl,-soname,$(LIBSONAME) -o $(LIBFILENAME) $(LIBOBJECTS) -lc
@#Mac OS X case
$(CXX) -dynamiclib -Wl,-install_name,$(LIBSONAME) -o $(LIBFILENAME) $(LIBOBJECTS) -lc


Now you can run sudo make install and the libraries and headers will be installed on your system. Awesome! Now try to build one of the examples, like examples/SimpleSend.cpp. Don’t forget to point to the include files (-I) and the shared library (-l).

cd examples
g++ I/usr/lib/oscpack -loscpack -o SimpleSend SimpleSend.cpp

Now if you listen to port 7000 and run the program you should see the message.

A while back, there was a poll on createdigitalmusic.com where people were asked about mobile music making devices that they would like to see the website cover. A few people said that they wanted to see coverage of calculators. It was treated as a joke, but I thought it might be fun to make it a reality.

I have an HP 48 GX. The HP 48 actually has a speaker that can be controlled programatically. This is fun for little beeps, but it isn’t very flexible. The calculator also has a serial port. This seemed more promising. I decided to try to figure out a way to convert the signal from the serial port into MIDI messages that could then be sent to a MIDI keyboard.

I used a Parallax Basic Stamp to read serial data on one pin and send MIDI data on another pin. Stamp Basic provides some nice functions for reading and writing serial data. The code on the Stamp looks like this:

note var byte
velocity var byte
MIDINOTEON con 144
MIDINOTEOFF CON 128
MIDICHANNEL CON 0
MIDIVELOCITY CON 127
MAINLOOP:
SERIN 3, 167, 16780, [note, velocity]
SEROUT 4, 12, [MIDINOTEON + MIDICHANNEL, note, velocity]
GOTO MAINLOOP

Nothing too special here. We read the data in on pin 3, and write out on pin 4. The incoming serial data needs to be coming in at 2400 baud. Also, the incoming message only specifies the note and velocity. The output should go through a MIDI output circuit.

I wrote some code on the calculator that takes keystrokes and turns them into MIDI noteon messages. The code is broken into two functions. The first is called FLIP. It takes a list and an index into the list, and returns the list with a “not” applied to the value at the given index, as well as the original index and the new value. I use this function along with a list to keep track of whether a note is on or off, so that when the key is pressed the corresponding note can be turned on or off. Here’s the code for FLIP:

<< -> L P << L P GET NOT DUP L SWAP P SWAP PUT P ROT >> >>

You caught all that, right? And now, here’s the code that watches for a keypress and sends the appropriate message:

<< '0*X' 'X' 0 127 1 SEQ
WHILE 1 REPEAT
IF KEY THEN
1 - DUP 10 / FLOOR 6 * 40 + SWAP 10 MOD + FLIP 64 * MIDIOUT
END
END >>

Ah, right, and the MIDIOUT function:


<< CHR SWAP CHR SWAP XMIT DROP >>


All of this is really just a fancy way to use the built-in XMIT function to send two bytes out the serial port.

So, after the coding and the circuit building, here’s what I got.

I had the microfiche machine at Maker Faire. And I said that I would post some information about it here, so that’s what I’ll do.

The microfiche machine takes light (in this case, from the screen of the microfiche reader) and converts it into MIDI signals that can then be sent to devices that take MIDI input, like synthesizers or computers. We use a phototransistor on the end of an antenna to read the light level from the screen. That information is then processed by a BS2 and converted into a MIDI signal.

In addition to the phototransistor, the microfiche machine has several other inputs.

• two dials and a button to set the highest and lowest note played
• a stop/star button that stops and starts the MIDI signals without actually turning the machine off
• two buttons that are used to calibrate the highest and lowest light signals
• a dial that is used to set the key of the notes that will be played

The BS2 does not have an analog to digital converter, but it does have a function called RCTIME that can be used to detect changes in resistance by determining the time it takes for an RC (resistor-capacitor) circuit to change its voltage. This article talks about how to use the RCTIME fuction and how to construct circuits that can use it. I use the function to measure the light levels from the phototransistor, and to measure the settings of the dials, which are just potentiometers (variable resistors).
Here is the BS2 source code that I use. If you have any questions, feel free to post them.

Update: The IEEE Spectrum website has a little writeup and a video showing the microfiche machine in action.

I just got home from Maker Faire a few hours ago. I’ve been there for the last two days presenting some of my projects, such as:

• Unicorn Vs. Dolphin
• my stomp pad
• $2 multitouch instrument (based on the water-in-a-bag design by Erling Ellingsen)
• Rock Band Keynote
• the Microfiche Machine (light to MIDI converter for the band Microfiche)
• the LED VU Meter

I talked to a lot of great people, and I got a chance to see a few neat projects, but I spent most of the time at my booth explaining what I was doing. Some of the highlights:

• families playing Unicorn Vs. Dolphin together
• a 14-year-old talking to me about how he had built some things that were similar to the stuff I had on display (I’m pretty sure his stuff was better than mine)
• talking to some high school kids about the projects they were working on in Python
• Darwin Grosse of Cycling ’74 telling me he though the Microfiche Machine was cool
• looking like an idiot by telling Mr. Grosse that I used Pure Data for a lot of my projects
• giving a short talk about the multitouch instrument

I need to give a few special thanks:

• Heidi, for all her help and support, and not yelling at me for the apartment getting messy, and covering for my complete lack of organization, and worrying about things when I wasn’t worried but should have been, and everything else she always does for me
• to Mike and Tim for including me on the VU Meter project
• to Adam and Amanda for watching my booth while I watched Microfiche play their set
• to Dan Mode for lending me his projector so that I could show Unicorn Vs. Dolphin in it’s full glory
• Noah Thorp for inviting me to give a short talk about the multitouch instrument
• the guys from Tube Time and NYCResistor for being awesome neighbors and for loaning me the stuff that I was too stupid to bring

So, if I talked to you at the Maker Faire and sent you here with promises of source code and diagrams, I will try to get those up tomorrow. Tonight, I need sleep.

I spent some time today trying to write a script to mark all of the spam comments that I’ve gotten over the last few months as spam. I tried to use curl to log in to the blog and delete them, but I kept having trouble hanging on to my session. It seems like if you don’t follow the right sequence of pages, you have to re-validate your session. And the, when you finally have the right sequence figured out, you have to visit more pages than you should. Finally, I realized that I could just edit the messages in bulk, and mark them all as spam. Sure, it took twenty pages of marking, but I finally got it finished.

So the point is, you should always look for an easy way to do things before committing yourself to the hard way.

But hey, if you’re interested, here’s what I was working with:


curl --cookie-jar /tmp/cj 'http://pillowsopher.com/blog/wp-login.php'

curl --cookie /tmp/cj --cookie-jar /tmp/cj 'http://www.pillowsopher.com/blog/wp-admin/edit-comments.php' --data 'log=aturley' --data "pwd=$PASSWORD"


Some of the –data arguments are unnecessary, but I’m tried right now, so I don’t feel like figuring out which ones.

I like boardgames. I’m probably not what you would consider a board game geek, but the statement “KIRK I TRADE WHEAT FOR SHEEP” makes me laugh. I like to think that I can enjoy any reasonably good game.

Except Monopoly.

When I was but a we lad I had a friend who loved to play Monopoly. I would spend the night at his house on weekends, and I always knew that before we built forts or watched movies or played video games or rode bikes, that we would first play a game of Monopoly. And by “play”, I mean that I would be dragged through a game of Monopoly. It would take hours. It seemed interminable. After the third hour I would begin to beg him to take all my properties at a deep discount and let me simply remove myself from the game. Only occasionally would he take pity on me.

So it was with great interest that I read this article about Monopoly. In it, the author claimed that the reason most people hate Monopoly is because most people play it with house rules that cause the game to go on forever. He flat out says, “Monopoly does not take substantially longer to play than most other board games.” He lays out his thesis rather well.

Being a self-professed fan of games, and priding myself on a willingness to admit that I was wrong (sometimes), I though it might be fun to have another go at Monopoly. After all, some people do seem to like the game, so there must be something to it.

I mentioned this to my boss, who is also a big gamer. A few days later he told me that he told me that he had found  a copy of Star Wars Monopoly in the bottom of the trunk of his car. We decided to get a game together soon.

“Soon” came last night. My boss, his wife, my girlfriend, a coworker, and me. We went to my boss’ place and got going. At first, things looked like they were going pretty quickly. Money was changing hands, but we weren’t flooding the game with cash. Property was being bought up left and right. I was impressed. My coworker went out first. I followed him a short time later. Monopolies had been established. When I went out, I left a massive pile of property and some cash in my boss’ hands. I had been mortgaged to the hilt, and couldn’t pull it off. Oh well. A few more trips around the board and my boss’ wife went out. He also got her property. Now it was down to my boss and my girlfriend. My girlfriend had one monopoly (the orange ones, just before Free Parking) and two other properties. My boss held most of the other property on the board, but almost all of was mortgaged, except for the three yellow properties just before Go To Jail. And that’s where the death march began. My girlfriend was able to elude my boss’ property often enough to no go broke. He was unable to get enough money to get his properties out of mortgage. Around and around they went. It was funny at first. After 15 minutes it was amusing. After 30 minutes it became tiring. After about 45 minutes of this I was quite alarmed. What happened to the short game I had been promised? What was going on here?

I have no answers. The combatants were finally persuaded to call a draw (no small task considering that both of them take their games very seriously) and everybody went home. I had hoped to have a fresh start with Monopoly. I had dreams of adding another game to my usual rotation. Those dreams evaporated last night, melting away like fog in the morning light.

Monopoly sucks.

Well, the final nail in the coffin of last week was not bike-related, it was car-related. Some crackhead got into our garage, busted out my driver’s-side window, and took my iPod. They also took the charger/FM tuner that I used to listen to the iPod, and an old Sony Discman and tape adapter that I had the glovebox. Not a very good haul when you consider the following:

• The iPod was engraved with “To Andrew Turley, from your friends.” Something tells me that’s going to drop the resale value.
• The iPod was pretty beat up. There’s this obsession that people have with keeping  iPods in pristine condition. If the buttons work and I can read the display, I’m happy.
• You can buy a NEW portable CD player for something like $20. Who the hell is going to give you any money for a used one?

I spent 30 minutes cleaning out the glass from the interior and the parking space, 2.5 hours and $190 getting the window replaced, and I’m out about $350 if I want a new iPod and tuner. Mr. Crackhead is probably up $60 at most. Doing the math, that leaves us (the crackhead and me) with a net loss of (190+350)-60=$480. Looking at numbers like that, I’m ready to just start hand out $20 bills to the neighborhood crackheads and begging them to not break into my car. Actually, what I should do is try to start helping out at CATS (Community Awareness and Treatment Services), just a few blocks from my house. They probably need volunteers and donations. I have to imagine that would be more productive than revenge fantasies.

I wondered what the bike gods had in mind. Now I see their diabolical plan. Rain.

Let me start by saying that I don’t mind riding in the rain on the proper bike. In my case, the proper bike is my fixed gear. What makes it the proper bike? Fenders. Fenders are the difference between riding along enjoying the sigh of a wet world, and riding along with dirty water pouring down your legs and into your shoes. Unfortunately, my fixed gear is at home with a flat tire. Today I’m on my touring bike, which lacks fenders. (And yes, I know, any respectable touring bike has fenders. Mistakes were made.)

So, do I tough it out and hope the rain isn’t too bad? Do I go over to Mike’s at lunch and pick up some fenders and put them on before I ride home? Do I completely chicken out and take the Caltrain all the way to San Francisco? There just aren’t a lot of good options. Oh yeah, and it’s supposed to continue to rain tomorrow, so I also need to figure out what I’m going to do about that.

I’ve done something to make the bike gods very angry.

• Monday: I biked to work from San Francisco down to Palo Alto. About seven miles from work I tried to clip in to my pedal after a stop light, and my cleat wouldn’t catch. I rode for about a block and a half trying to get it in, and finally pulled over. It turns out the spring holds the blades of the pedal in place had broken. I rode the rest of the way with a more or less useless pedal.
• Tuesday: I took the fixed gear to work. Riding home on Market Street I notice a low rumble and some bumpiness and a wobble in the bike. I pull over and sure enough I have a flat tire. I was only a few blocks from home, so I just walked. It was windy and cold, and it sucked. I still haven’t figured out what happened, but the tires are pretty old, and when I looked at them without any air in them it became pretty obvious that they need to be replaced.
• Wednesday: I replaced the pedal on the Trek thinking, “It’s a good thing I have a backup bike.” At work, just after lunch, I looked at my bike and noticed that I HAD ANOTHER FLAT. It turned out to be a big staple through the tire. It didn’t take too long to fix, but it was annoying.

Let’s see what today has in store.

John Backus, the man who headed the team that created Fortran, has died.

Fortran helped usher in the modern computer era. It gave us a way of separating software from hardware by giving us a language that could (in theory) run on any platform for which there was a compiler. Fortran is still with us today in places like high-performance math libraries.

This is my 21 gun salute to a man who changed the world.

      program MAIN
      integer i, n, sum

      n = 21

      do 10 i = 1, n
         write(*, *) 'bang'
  10  continue
      END