Insight VR

I had a great time at PyCon and was gratified to see the number of people that showed up for the wiiMote open space and the enthusiasm that they showed. Oh, and the patience too, given that it took much longer that I had expected to get the demo to actually work. Sorry about that.

I know that I owe everybody new downloads. I’m working on it. I want to make sure that things are packaged up nicely and are relatively easy to use. So expect a post with detailed instructions and downloads soon.

Also I’ve done some more visualizer work. It isn’t ready for download yet, but here is a preview of a new visualization mode for Manifesto:

Manifesto Demo: Happy Up Here from InsightVR on Vimeo.

The pwn2own contest is going on.  Safari on Mac OS X fell in 5 to 10 seconds, generating headlines.  There’s a sense in which these sorts of hacking contests are meaningless, since everyone knows that there are holes in each system.  The 5 seconds headline is especially meaningless since the exploit had taken over a year to prepare.  Once it is all packaged and ready to go it can look elegant and swift, but some articles give the perception that the guy walked up the the machine and hacked into it by writing an exploit from scratch in a matter of seconds.

What actually happened is that over the course of a year (doubtful that it was full time) Charlie Miller wrote an exploit.  Prior to the contest he packaged it up nicely and put it on a server.  He gave the URL to the contest organizers before the contest.  Then at the contest he simply clicked on the link he had given them and demonstrated that he had control of the machine.

So the headlines give the impression that the Mac is not as secure as Windows and other such wild extrapolations.  Security is a complex combination of not only the technologies at use in the OS and the carefulness of the implementation, but also of the economic incentives that exist to attack a set of systems.  Most stories on this event give only a nod to the complexity of the situtation.  But ZDNet has an interview up with Charlie Miller which gives some insight into the hows and whys of the exploits.

(continue reading…)

I saw a post on BoingBoing yesterday about high resolution Tetris. This reminded me that long ago I wrote a PacMan Java applet that got progressively harder. The intent wasn’t to create an exact duplicate of PacMan. Instead I wanted to have similar PacMan movement dynamics (you’d be surprised how many PacMan clones mess this up badly) random mazes, and an intelligent ghost.

Click through to play…
(continue reading…)

My wife and I are in the Bay Area this weekend for our 10th anniversary. We wanted to visit Stanford so I looked online to see if there were any activities that would be fun. I didn’t find anything “fun” but I did find out that there would be a lecture on Data Visualization. Given that I do data visualization for work and music visualization for fun I thought it would be worth dropping in.

The talk was part of CS547, organized by Terry Winograd. I had CS147, intro to human computer interactions, from Prof. Winnograd about eleven years ago. I took it because I had to for my CS degree and some friends were taking it as well. I remember enjoying the class, working really hard on a remote helicopter control software presentation, and getting hammered on the final. I still think that my answers were good but unexpected rather then wrong.

I was excited to attend the presentation. It was in the basement of the Gates buiding, room B01, which I spent a lot of time in 11 years ago. At the time the room was brand new and very high-tech. Given the pace of technology it has aged pretty well. I don’t know if the cameras are high definition now (useful when trying to read a chalkboard on a TV or computer screen) and the projector could be brighter, but overall there isn’t much I would change.

As for the talk itself, Jeff Heer gave a 90 minute presentation on the history of data visualization. I thought it was fascinating.

Those that are really interested will want to watch the movie I just linked to, but here are the things that I found most interesting:

• The ability to reorder data allows you to see patterns more easily. Jeff showed a picture of a mechanical loom that was created in order to be able to more efficiently move strips of paper with graphs from a data set on them. We’ve got it relatively easy today.
• John Tukey’s work was something I was unaware of. One of his goals was to separate data analysis from statistics. I had never thought of these as separate camps, but the situation at the time was such that visualization was looked down upon as being imprecise. While this may be true, it doesn’t mean that visualization and data analysis are not useful. Seeing a variety of trends and other graphical information can allow one to quickly look for insights that might not be obvious from a table of medians and standard deviations.
• The concept of interactive brushes to filter data dynamically was new to me. This is more easily demonstrated than explained (see the video) but Jeff showed a system which displayed a number of small scatter plots derived from data from the same group of items. Highlighting a grouping in any of the scatter plots caused the same items to be highlighted in all the other graphs. This allowed a user to quickly see similarities and differences between the selected items.

Towards the end Jeff talked about moving towards a building block model for visualization so that non-specialists can interactively build and manipulate visualizations. Given my experience I think that this is a lofty goal, but even lowering the bar to building visualizations a little will get many more people into this space. He demonstrated Protoviz, which is his unreleased javascript visualization language which can be used to (relatively) quickly and easily build visualizations. I’ve got a love/hate relationship with javascript, especially on IE, but it is the defacto language of the internet and is widely used, so this seems like an interesting project.

The other thing he demoed was http://sense.us which is a tool to visualize US Census data. This showed another way of lowering the bar to visualization which is to allow for interaction between users. Users of the site can leave comments and annotations so that others looking at the same data can see the questions it raised and some suggested explanations. This also enables using meta-data about who looked at what items. So you can then see what data is popular and look at why that is. You can also see which paths haven’t been explored yet and be something of a trailblazer by sifting through virgin data.

After the talk I went up and thanked him and explained my own difficulties in trying to get users to understand and use treemap software. He was sympathetic. He stated that the the New York Times, which has some great visualizations, is loath to use scatterplots because readers can’t understand them. If the do use a scatterplot it is because the feel that it makes an essential point and they simplify and annotate it like crazy. He also mentioned Hans Rosling’s TED talks and said that people come away thinking, “What a great visualization!” while what Jeff thinks is, “What an amazing presentation and explanation to make such a complex visualization accessible to so many people in so short a time.”

So I didn’t get a solution to my treemaps vs users issue, but I certainly have a lot to chew on for now.