see a puffin eat a fish

We’re over in SF staying at the Unicorn Precinct, I fly back tomorrow. The first half of the week we were at MUM 2006 presenting our paper and met a lot of Fins from Nokia. The Keynote, Mark Davis, one of the founding members of Yahoo Research, did a really great talk on ZoneTag and the power of community based tagging. He should a really neat semantic map that allowed you to see where people think places ARE versus where a static map says they are. We also bummed around Palo Alto and Mt. View and stayed at Yelena for a while.

The picture here is of the Tele-Actor a potential interesting use of WayMarkr.  A wearer of WayMarkr could participate in bidirectional communication.  The viewer of a wearer’s perspective could inform and instruct the wearer for what to do next.  This of course would only work if the photo upload lag for WayMarkr was reduced.

I also have a new appreciation of the WayMarkr but have found out that my Razr won’t support it.  My phone only supports CLDC 1.0 and a Motorola developer has confirmed that i’m stuck.  But I want to start using ZoneTag so maybe i’ll give the Nokia another chance.  I think the appeal of ZoneTag initially for me is not even the new development of community tagging but that I can take photos, they get buffered + uploaded later and automatically appear in the rich Flickr UI.  I could see myself documenting more.

I’ve been thinking about directions lately. People have an emotional attachment to web services or certain knowledgeable friends when asking for directions. Some people swear by MapQuest, others always call Shawn. Without knowledge of the back end systems I wanted to see how these different methods stack up. Requesting directions creates an ad-hoc, single direction network with variable hop distances. Although we are given all the directions at once, we can imagine a scenario where each node has addressing information for the next node. So far humans are doing a slightly better than job than computers in navigating the west side of town.

Although all the different online mapping services picked the same route, individuals had a varied set of paths.

The de Stijl movement was wrapped up in two kinds of beauties, a sensual or subjective one and a higher, rational or objective, “universal” beauty. Using live video and the geometries inspired by Mondrian, I am collecting live action video and generating Mondrian style paintings in real time. The idea is to distill the chaos of real life to its core, primary elements.

The above Mondrian is of sunset outside of my apartment. I am capturing 320×240 pixels, but my laptop doesn’t have enough cycles to keep track of 76,800 pixels. So i’m averaging them down. For every region of the Mondrian, the R, G and B values can be only one of 5 values giving me a total of 75 possible colors. I’m still playing around with these numbers to see if bumping it up to 216 colors (6^3) gives me enough variation to make it worthwhile. Also worth mentioning is JPGs of these images compress really well!

The painting above is the camera trained at Anna’s desk. Is this something? What else can I do here? How can I break down the live image to it’s core colors and essential visuals? Will changing the geometries help? Adding more colors? A nicer camera will definitely help as the iSight I use is terrible for capturing color. I know there has to be something else here, or is the canvas I am using too limited?

I finally got to play with my GPS this week. I got it a few weeks ago but haven’t really had a chance to try it out. Anna and I really want to mess around with GeoCaching too, it seems like the kind of geekery that’s right up our ally. I don’t know how well it’ll work in the city but there are plenty of caches right down the street from us. I also started thinking about using the body as a paint brush. Our gestures, our movements through time and space as the pallet for a new composition.

Geodrawing is using the earth as a canvas, making a drawing out of a collection of GPS markers. It’s using the body’s location to draw.

Being out of the city this past weekend was a great opportunity to try out the GPS, no “urban canon” effect. We drove way into Connecticut, close to the Rhode Island border. I kept the GPS on and it drew a fairly accurate line of our location, very cool. Here’s a screenshot and a link to the GPX file. Open it in Google Earth for fancy effects.

I wanted to see if I could make something that involved time and motion to paint using the natural movements of the body. To see if I could make a collage of the body’s movements at different intervals. I’ve evolved an application with a fairly straightforward interface. It happened organically which I think is very cool. The program paints with video and time, that’s the working description anyway.

It starts out by capturing a certain number of frames to be played back at video speeds. The diversity of the painting pallet will depend on the quality of the captured footage. Once the video is captured, it can be played forwards or backwards with the arrow keys. If the video hits the end or the beginning it loops around. The space bar toggles the size of the brush. It’s either the whole window or a resizable circle. As you move the brush around, only the portion in its view is animated forward and backwards. As soon as the mouse is moved, the brush freezes that area in time and begins to work on a new drawing. Here’s a sample image:

It’s pixilated because due to performance reasons it’s 320×240 blown up. That could be fixed later.

Motion Box came by school yesterday to show off their video editor and recruit. Motion Box a simplified final cut for the web that allows you to upload, splice and share your content using a Flash application. Although there isn’t much feature length content on the web, most of YouTube is three minute clips of someone falling out of a boat, Motion Box contends that even those clips are way too long. They have a point. In extreme attention deficit disorder land where we all now live, the 2 minute buildup to the 3 second clip of someone falling out of a boat is way too long to wait. Now in the two minutes where nothing happened, you know right before the guy fell out of the boat, you can have a highlight reel of hundreds of people falling out of boats. That’s a lot of wet people.

In all seriousness, the interesting idea that Motion Box brings to the table is the de-contextualization of video. Not only can you splice together your content, you can splice together any public content on the site. This is exactly why they don’t yet allow to download spliced up video, there’s a big issue with attribution. Everyone’s heard the famous Amen Break story but what if you could do the same thing with video?