Saturday, February 1, 2003

Six boys and a girl from a private school near Calgary were killed in a huge avalanche on Glacier National Park, British Columbia. They were on an annual school tour guided by school staff. The avalanche danger was rated "considerable." I find it difficult to understand why they were out in that place at this time. The Revelstoke disaster a few weeks ago, the "considerable" rating, the well-documented weak layers, a large group, would seem to suggest a "no go" decision. Especially with kids. Here's the opening of the relevant avalanche bulletin: Another 10cm of snow overnight is adding to the light, yet steady snowfalls that have accumulated since midweek.  The new snow has fallen under mild and at times windy conditions in a warm southwest flow.   Thus, a total of 30 cm load of storm snow is contributing to the stiffening of the slab over a surface hoar layer responsible for last weeks extensive avalanche cycle.  Snowprofile work done yesterday concentrated on West and North aspects both around the 2100 meter range.  Stability tests showed consistent failures on the Jan 20 surface hoar with scores in the moderate to hard range.   The new snow instability described yesterday, now over 30 cm deep, was also observed on the west aspect.   Given the recent and past wind events, these are the layers of highest concern at this time, especially on lee slopes or near large terrain features.  Ski tests easily released the Jan 20th layer on unsupported rolls on the north aspect. Notice that last sentence. They were on a North face. 10:49:11 PM

A record executive and his son make a formal case for freely downloading music. The gist: 50 million Americans can't be wrong. [Salon] Very interesting thoughts from an industry insider, with a link to an excellent Tim O'Reilly analysis, which I hadn't seen before. I have had similar discussions with academic journal publishers. They are as hated by academics as the music industry is by music lovers. For the same reasons. They think that their product they distribute, which they must protect at all costs, rather than the selection of content, the metadata. At the same time, they pay little, sometimes even charge, the actual creators. There is a valuable, honorable business in helping match creators with audiences. But instead they focus on fleecing both sides and preventing many potential matches that new technology would allow. 9:26:09 PM

Unlike politicians and the public, people working on the shuttle know it's still experimental, an expert on the Challenger disaster says. [Salon] In a way, much of our technology is experimental, in the sense that it is too complicated for any analysis or testing procedure to be able to provide accurate estimates of risk. Controlled experiments are rarely possible. Instead, we learn what we can from the debris of tragedy. This is a fundamental difference between science and modern engineering. It is natural to ask why we could not predict and avert disaster, given all the powerful quantitative analysis methods that science has given us. It is hard to accept that the world and our machines have far too many parts, and too many unknowns, for certainty or even good probability estimates. But we should know that from the rest of life, so unpredictable in its pain and wonder. 9:13:19 PM