"Enkin" introduces a new handheld navigation concept. It displays location-based content in a unique way that bridges the gap between reality and classic map-like representations. It combines GPS, orientation sensors, 3D graphics, live video, several web services and a novel user interface into an intuitive and light navigation system for mobile devices.

This project is a submission for the first round of the Google Android Developer Challenge and should not be considered a final product.

Enkin from Enkin on Vimeo.

Bruce Sterling from Innovationsforum on Vimeo

So for the last 50 years or so, we've seen the rise of weekly media moments. Those water-cooler, event television moments, where instead of gathering around a warm fire, we all huddled round the blue glow of our TVs.

But now that broadcast is crumbling, and I can watch shows whenever I want. The question becomes, does the loss of the VHF ties that bind us impact our society adversely?

Its not like people used to read Moby Dick at 10PM on Wednesday's right?

One might think that the United States' nuclear weapons — the cornerstone deterrent in the country's arsenal — would be treated with the utmost precision.

This comfortable illusion was shaken on Aug. 31, 2007, when crews loaded six live nuclear warheads onto a B-52 bomber and flew from Minot Air Force Base in North Dakota to Barksdale Air Force Base in Louisiana, cruising over the nation's heartland. Each warhead was 10 times more powerful than the atomic bombs dropped on Hiroshima and Nagasaki during World War II.

The last day in August [2007], Air Force personnel loaded the nuclear warheads on a routine repositioning of weapons stocks, believing them to be cruise missiles.

The article goes on to cover what the USAF can do to prevent this lovely state of events in future. However, I'm really motivated now to start saving up for a doomsday bunker. Maybe I can hide it somewhere on Hampstead Heath…

Image models::

James Hayes and Alexi Efros: Scene Completion, here, they created a 10k library of images, and used that to mask and replace parts of images. This didn't work so well till they had Millions of pictures.

Finding Canonical Images: Jing, Baluja, Rowley @ Google: Most image searches use key words from meta data. In this case the algorithem looks for like features in images over thousands of images, and asks which features are most like the others to sort page rank style, this allows the system to find images in a better way.

Image Represenation: Using EigenFaces, a blurred representation, to find like images, basically, looking for averages, or major similar features (roughly).

Text Models: How do we segement text into understandable words? Segmentation, ask for the probabilty that a chunck of letters is a word…. Needs billions of words in a library of some sort to identifiy those words.

Google Sets: Searches for related words based on stuff around the words you searched for across Google's vast index. So, if you were to type "lions, and tigers and bears" you'd get a list of other animals returned.

Machine Translation: Statistical Machine Translation, look for alignments. First line up two sentences in different languages, then look for statistical probability that the words in sentence 1 appear in similar combinations across your index of text. Repeat for line two, now look for the relationship between those combinations between sentence one and two.

Going over Repeatable Actions, Patterns & Processes

Show everything

Collecting until there is a pattern:

- In the News: Grabbed news from google 4 times an hour, and show changes over time
- Oakland Crimespotting: Oakland's GIS only stored info for 30 days, so Stamen started caching it for longer periods to get better understanding
- Dig Ark, keeps the center clear for the subject, the information flows around, to create a contextual understanding of what's happening: You don't need to feed your viz precompiled data all the time you can also think about ways of firing up the viz and feeding it data over time…

Color:
- ColorBrewer: Generates pallets for geografic data
- CPT-City: Catalogues color palates over maps!
- Dopplr: Generates city for every city by taking the MD5 Hash (function in most languages) of every name…
- IBM's History Flow does the same thing. It uses the hash of the string changed in a Wikipedia entry.

Find the "Objects" and Connect them:

Look for Second Order Connections:
- Get information out of what people are doing anyway…
- Check out Oreilly Book called: Programming Collective Intelligence.

Put Sliders on it:
- You've shown everything, now how do you dial it back????
- YUI has good one
- Measure Map slider from Adaptive Path, includes histogram…
- Willett & Heer: Scented Widgets:@ Improving Navigation with scented cues

Add a playback button & Animate it:
- If you've already got a scruber on a slider, its only one more step to animate your data

Give Everything a URL:
- I should be able to link straight into stuff: yes, true, but why can't i do that to call up a feature, or instance of something within a visualization?
- Paparazzi: desktop app that can take a snapshot of a page at an interval of X, which is great cause then you can string those snapshots together….

Nasa Blue Marble Tiles: Freely available sat imagery from NASA!

Seek Forgiveness, not Permission:
- In the case of Oakland Crime Spotting: It took a week to write the script to scrape the data off of the city public mapping site, and it took 6 months to ask for the permission to get the data in an easily manipulatable form. Nutty.

- For example: Everyblock.com - scraping the web for all stuff related to a place. Screenscrape your way to wealth of information, the problem with this is that this produces "data friction," as John Udell puts it.

Use Direct Manipulation:
- An old example: Film Finder

Flare Visualization Toolkit: Flash based set of libraries to help create viz in flash…

Start and End with Questions:
- A good visualization asks questions, not necessarily answers questions difinitively, it should open up more questions.

Over our pre-session caffine boost, Michael let me know of a new development in the progress of OpenMoko that was announced at 6 am today. OpenMoko has just released the CAD files for the physical design of both the first edition of their device, the Neo 1973 and their soon to be released Freerunner (GTA 02).

The idea is to see what the community starts to do with the physical details of their device. Having just started working with product designers myself, this is a particularly interesting development.

The key here, is really what's always made me excited by FIC's backing of OpenMoko. Have a look at Dash, a new in-car navigation system. Dash is built on both the hardware and the software that makes up OpenMoko. Now things are really getting interesting, as other companies begin to build on top of OpenMoko as a mobile platform. Now we can begin playing with the look and feel of our devices in the same way we tinker with code.