I flew, with colleagues, to Portland in Oregon for the Diagrams 2010 conference. We arrived on a Saturday and departed the following Thursday. This makes it the shortest period that I’ve ever taken a long-haul flight for. So today I’m feeling rough: approximately one part dead, four parts like I have some painful communicable disease and six parts hungover. I’m counting on a smoked rasher sandwich to right all the ills in the world.
As I previously posted, Portland is fantastic. We partook in many of the local cultural delights such as visiting Powell’s book store, sampling ale at the Deschutes micro-brewery and visiting the Bite of Oregon festival. I spent $10 at Powells and picked up a copy of Flatland and a copy of Wittgenstein’s Brown and Blue book. To my lifelong shame, I made a beeline for the maths section, but totally forgot to visit the computing section! The conference banquet was in the local art museum, which was excellent. They have a collection of American abstract art which Paolo Bottoni kindly explained to me. They also have some Rodin sculptures and a sculpture by Picasso, a Monet water-lilly, an early Van Gogh and several other impressionist works by artists such as Cezzane. All very beautiful.
The conference itself was an excellent event. I thoroughly enjoyed Randall Davis’ keynote. He demonstrated the breadth and depth of research that you might expect from an MIT lab. His assertion that we should be having a conversation with our computers has already influenced how I think some of our sketch recognition work should proceed.
There were lots of excellent papers. Two in particular will, I think, influence future directions of my own research. The first is by Koji Mineshima, Mitsuhiro Okada and Ryo Takemura titled “Two Types of Diagrammatic Inference System: Natural Deduction Style and Resolution Style”. In this paper they do lots of interesting things, including quantifying free rides by comparison of Euler diagrams and certain sentential forms used in reasoning. The second is by Mathias Frisch, Jens Heydekorn and Raimund Dachselt titled “Diagram Editing on Interactive Displays Using Multi-Touch and Pen Gestures”. They demonstrated some really well done software for drawing certain types of diagrams. I’m particularly interested in taking the approach described in Davis’ keynote, the design and technology described by Frisch et al. and the constraint based layout described by Tim Dwyer in “Hi-tree layout using quadratic programming ” to inform our sketch recognition stuff.
There were some other great papers. David Landy spoke about his “Toward a physics of equations” and demo’ed his very, very cool iPhone/iPad application that helped him capture his findings. The papers that won the best paper prizes were, unsurprisingly, interesting and well presented.
All in all, I had a great time in Portland. I enjoyed the beer and the food was better than food I’ve eaten anywhere I’ve visited in the US. And thankfully, the rasher sandwich seems to be doing its thing. I might even get some work done later