/* ---- Google Analytics Code Below */

Tuesday, April 26, 2005

Electric Universe

Read David Bodanis' Electric Universe, The Shocking True Story of Electricity .... which is a set of essays on the history of electricity. The are disjointed, hardly forming any sort of sort of coherently written view of the subject, which I was looking for. The longest essay is a popular history of radar in WWII, which though interesting, is out of place in this book. The individual topics are of interest, but scantly covered. Some of the writing too left me a bit puzzled. The author heavily overuses specific mental images, e.g. clouds of electrons for example. There are some emphatically negative reviews in Amazon, and I checked the references on a few, and the mistakes could be attributed to interpretation, but I would expect a book like this, to be read by non-scientists, to be more clear. This is another example where it is very hard to believe that the well-known reviewers on the back cover ever read the book. The Amazon reviews led me back to Robert Thompson's 1947 history of Telegraphy: Wiring a Continent. Still considered a classic.

Monday, April 25, 2005

Mathematicians and Tapestries

Interesting article Capturing the Unicorn A story of digital imagry, reconstruction and mathematics. Its also about stitching images. For some new work on this idea, for decidedly less complex sets of images, see here. An interesting, seemingly simple topic, with lots of complexity.

Sunday, April 24, 2005

Examining Playful Invention Company (PICO)

Saw a presentation by Elisabeth Sylvan on the Playful Invention Company (PICO) this past week. This has come out of the Lego supported work at the MIT Media Lab ... inspired by Froebel and Papert. This first inspired the Lego Mindstorm project, and now PICO plans to sell programmable bricks or crickets, which can be programmed and used to design systems. I am particularly intriqued by the tactile aspect of this, and its use to help engage collaborators on tough problems.

Monday, April 18, 2005

Sunday, April 17, 2005

Bobby Fischer Goes To War

Reading Bobby Fischer Goes To War: How The Soviets Lost the Most Extraordinary Chess Match of all Time. I followed this event closely when it was news. Now Fischer is back in the news due to the recent granting of Icelandic citizenship to him. An avid chess-player at the time, I had no idea what Fischer was doing outside of the moves of the game, it was my first introduction to the notion of gamesmanship, bizarre as his form of that was. Finally a writeup of what occured, and hopefully a full explanation of the motivations.

Update/Review: You really have to be very interested in this topic to want to read this. It did fill me in on some of the events, which I only barely remembered via NY press coverage at the time. It most made me wonder why Iceland wanted to take him back in. Reinforced my impression of Fischer as an embarassment ... Although it seemed like Fisher-Spassky progressed interest in chess at the time, it has progressed very little since then.

Wednesday, April 13, 2005

Gershenfeld's FAB

Have not read Gershenfeld's FAB book yet, but its on the list. His previous popular books were thought-provoking, and his modeling texts are well worth taking a look at.

Sunday, April 10, 2005

Shaping the Future - Robust Simulation

Brought to my attention by Steve King at Future Now an article: Scientific American.com: Shaping the Future -- On Robust modeling. Interesting view of the problem of robust modeling under uncertainty. I will comment there later this week. Brought to my attention: The Copenhagen Consensus and the fact that the poorly performing Club of Rome still exists. Here is a link to the PDF of the full RAND Pardee Center report.

Friday, April 08, 2005

Affective Computing example

Here Software Detects Unhappy Callers is an example of interacting with consumer emotions to classify calls. Note commercial example of work suggested by Rosalind Picard in her book: Affective Computing.

A Trail of DNA and Data

Paul Saffo article: A Trail of DNA and Data

Wednesday, April 06, 2005

Beer Games and Engagement Models

Some work activities today made me re-visit the MIT Beer Game. I have some fond memories of participating in this game back in the 90s. Also recall some work being done on an agent-based advanced, multi-SKU version being developed by members of the Santa Fe Institute, have to look that one up. A good example of a simple game leading to complex behavior. This is required reading for folks new to supply chain. This brings to mind what I call 'engagement models', used to introduce folks to a domain, usually very visual, but which are not intended to be predictive. I have seen these to be quite valuable in industry. The Sim models are examples, but perhaps more complex than they need to be to represent this modeling approach.

Monday, April 04, 2005

Niklas Luhmann

Niklas Luhmann's Social Systems Theory . - In the latest issue of the Journal of Artificial Societies a post on the use of agent-based models to implement Luhmann's societal theories, which has led me to investigate them further. Further, these papers may be useful as examples of the use of simulation for agent based models and related techniques.

Physics of Society

Futurology gets a little more exact: Guardian article on Philip Ball's: Critical Mass: The Physics of Society. (Published under a different title in US ...: Critical mASS: How One Thing Leads to Another ) Covers issues of modeling people's group behavior, most recently suggested in agent-based technologies. Do not have the book as yet, but it looks to be interesting. Is it possible to model people in a way that is statisically meaningful? Yes thats been done fairly well,depending opon the measures you care about. Is predictive at some individual level? Not clear. Can provide true emergent behavior? In this too the jury is out ... I have not seen it as yet. This book appears to be non-technical, for a technical view of the complexity of models of people see: Journal of Artificial Societies ... See also Ball's article in PhysicsWeb.

Saturday, April 02, 2005


Picked up Patrick Keefe's Chatter, recently, all about sigint surveillance, in particular the Echelon project. We have all heard about term 'chatter' by now, we have a feeling for what it means, but how is it measured? That question is never answered. I pose the more general question in a post at Future Now. I know a little about the subject and worked on Darpanet many years ago, have also looked at text mining approaches .. so I thought there would be some interesting detail here. Pretty early on Keefe makes it clear that he doen't know much about the topic, or technology in general. He provides uninteresting descriptions of his standing outside of facilities looking in through the fence, or interviewing other journalists with similarly limited knowledge. While visiting a purported, abandoned NSA site ... he is amazed that the computer room had a raised floor ... that there was a cable tunnel between two buildings on the site. He is in awe of the radomes that protect antenna complexes. He shows shock that an employee at a UK base is prosecuted when she steals classified material to support her own political agenda. He lost my respect fairly quickly, but I did slog on. He is further amazed that the NSA's public web site does not contain a full directory of their employees (no company does this today). He describes the criminal trashing of Admiral Poindexter and his neighbors to make a point. There was no faint suggestion of loss of privacy or potential data mis-use that he does not cover in depth, but there was not much mention of the potential value of these capabilities. His discussion of 'mining' in particular was weak; you got the impression it was just about finding words. The one part I did find interesting was some of the statistics regarding human intelligence ... but can I trust them? Find this at the library, but beware.