Reality-BasedTM Political Numbers from Nicholas Beaudrot
Electoral Math
Reality-BasedTM Political Numbers from Nicholas Beaudrot
Friday Shuffle! | Home |
This past week I got to sit in on Edward Tufte's one-day seminar on ... the presentation of technical information, or something to that effect. I can tell you that Tufte likes Galileo, dislikes Power Point, is not a huge fan of marketing, and cares deeply about user interface design. For some of his more groundbreaking work, take a look at his chapter on sparklines (huge page with lots of big images; broadband recommended), which, if you can produce enough of them quickly, promise to drastically increase the amount of content on a page.
Most fascinating was Tufte's work on the commissions responsible for investigating the root causes of the Columbia and Challenger accidents. In both cases, there had been some discussion before the accident that a catastrophic failure might occur. In the case of the Challenger, engineers responsible for desigining the rocket boosters correctly identified that the cold temperatures at launch would cause a failure, but could not present this evidence to NASA executives in an effective manner. In the case of the Columbia, power point slides presented by Boeing obscured the lack of data on the effect of large chunks of foam landing on the shuttle; tests had examined the effect of 3 cubic inches of foam hitting the shuttle, while the estimated size of the foam chunk was around 1920 cubic inches.
These two accidents illustrate the tremendous challenges of delivering quality information to executives in large organizations, as well as the difficulties of avoiding the influence of political pressure in high risk technical decisions (engineers were told to "reexamine" their calculations in the Challenger accident after recommending "no go"). Sadly, there aren't many silver bullets to be had in this department either.
| | technorati
Last
updated by Nicholas Beaudrot on 08:43 01 August 2005
