While at school in Cambridge, MA. I took a course in decision analysis. One of the examples given in class was the case of the Hancock Building. This is how I remember it: A model of the Hancock Building and the surrounding structures was tested in a wind tunnel. The wind direction in the initial tests was incremented by 45 degree intervals. The model behaved well for these tests. Later, after the problem occurred on the real structure, more testing [at a finer mesh] revealed very narrow bands in wind direction in which resonance did occur. The 45 degree increments were too coarse to pick out the resonant zones. (I believe that their initial tests were done informally, as a kind of favour, and so were not very rigourous.) *This is all from memory, so my apologies if I didn't get it quite right. A. Remy Malan
This is from memory, and it's late, so bear with me: A very recent Smithsonian (June 86?) had an article on flight simulators -- the same month as the Scientific American article. In it, the chief instructor for one of the airlines related that, a few months ago, he flew as the flight engineer on a commercial flight. The plane encountered a wind-shear situation on take off. The instructor, from his flight engineer's position, reminded the pilot that the correct recovery for wind-shear is opposite to the correct recovery for a stall (which has a similar appearance to the pilot). Hope this reassures your pilot subscribers. By the way, accident investigation reports are usually summarized in Aviation Week and Space Technology. Martin Minow firstname.lastname@example.org
Personal item, no documentation known: I once purchased a used USPS station wagon at GSA auction for $350. While cleaning it out, my wife and I found well over a hundred pieces of undelivered mail. We trashed all but the first class - and dropped 30 or 40 pieces into the nearest mail box. Some were over five (5) years old. We watched the paper for days, but saw no items about late mail. Only relevance to RISKS is that people will _always_ be imperfect. - Mike [And how often do we assume that a system will work properly in the face of that statement?! PGN]
>From: Graeme Hirst <gh%ai.toronto.edu@CSNET-RELAY.ARPA> >Subject: Re: Laserprinter dangers Sean Malloy dealt with the ease of forging with laser printers. A more general point is that forging ANY computer-produced item, be it a hard-copy output or a message on a wire, is easier than forging old-style pieces of paper, etc., because:- 1. The machinery involved is cheap - bytes on a wire which have come from a cheapo toy computer just look just like expensive DEC or IBM bytes. (Coiners need expensive metal presses) 2. You can realistically attain a 100% perfect forgery - my bogus bytes look just the same as real ones. (Coiners presumably have difficulty making the right alloys, but worse, have to copy the shapes on the coin - how do they know when their product is "good enough"?) 3. The skills required are, more or less, the same for producing ordinary software as for producing forgeries - software is software, whether legal or otherwise. (It is also true that an engraver uses his same skills whether he is forging banknotes or producing a bookplate; the big difference however, is in the widespread distribution of skills needed for forging - there are very few qualified engravers, but lots of "qualified" programmers). In summary, a lot of people are finding themselves in a position they were never in before - not only have they all the skills and equipment necessary for a particular type of crime, but increasingly they are being presented with opportunities to commit those same crimes. Ergo ...
In the Washington Post, July 30 1986, pg A-23, columnist James J. Kilpatrick discusses the nomination and confirmation of Daniel Manion as appellate judge. He laments at length the lack of support for the individual, notes that a keen sense of justice is not all that important for appellate judges anyway if they have a good knowledge of the structure of law which is what they really rule on. He goes on to note that the analysis of pertinent law and the detailed writing will likely be turned over to law clerks anyway. The last paragraph of the article is the clincher and source of humor. "In sum, I fear not for the republic, or for the 7th Circuit, when Manion joins the club. Give him an intelligent clerk and a good word processor, and the gentleman may look forward to many happy years on the bench." Do you suppose it could be called an application of AI, when software offsets presumed deficiencies of appointed officials? Are things such as this off-the-cuff suggestion an early step of having software front for the performance and/or the beavhior of public officials? And with what unseen, possibly unknowable, risks?
The following article appeared in yesterday's Observer, and is reproduced here without permission: Robert Stroud, Computing Laboratory, University of Newcastle upon Tyne. ARPA robert%cheviot.newcastle@ucl-cs.ARPA UUCP ...!cheviot!robert ======================================================= "Chernobyl report blames turbines" p.6 Observer, Sunday 10 August 1986 by Robin McKie and Laura Veltman (c) Observer Newspapers Soviet operators who experimented with turbines and alternators at the Chernobyl plant are to be blamed for the nuclear disaster there. Western experts who have recently visited Chernobyl say that the full Soviet accident report which is expected to be published this week, will blame 'human error' and 'misuse' of turbines for the chain of events that led to the disaster in April. But many believe the explanation is inadequate and that it is being promoted mainly to protect the country's nuclear construction programme. 'The theory moves the source of the accident from the reactor itself to the turbines which are housed separately,' said Mr Peter Potter, a British nuclear expert who has seen many Soviet reactors. 'By maintaining that human error and turbine problems were really to blame, the Russians could say that their reactors have no serious design flaws. They could then avoid calls for closures of other reactors or for the implementation of drastic redesign work.' The Soviet theory argues that the Chernobyl accident was caused by a total loss of electricity supply to the pumps which circulate cooling water through the heated reactor core. One Western scientist, Professor Leslie Kemeny, of the University of New South Wales' nuclear engineering group, does believe that an accident with the electricity-generating turbines - which are worked by steam heated in the reactor - triggered the disaster. Prof Kemeny, who took detailed samples of air, water and soil contamination during a recent visit to the Chernobyl area, said: 'The loss of electricity to the pumps was due to human error. During the night of 25 April, the turbo-alternator linked to Reactor 4 at Chernobyl was undergoing a "run-down" experiment. In effect, this meant that engineers were studying the behaviour of the turbines while they were being run down. Throughout the hour of the experiment, alternative energy sources should have supplied replacement power for the pumps. But this did not function, and the reactor was left uncooled.' Normally, the reactor's own electricity should have been used to run the cooling pumps. During a run-down, an alternative source should have been switched on automatically. It was this which failed at Chernobyl. Without cooling water, the reactor's temperature was sent soaring - with dire effects on its uranium fuel, zirconium cladding and graphite core. First the remaining water inside the reactor heated up, forming steam which began to react with the zirconium to produce hydrogen. The pressure of the steam and the hydrogen eventually cracked the reactor core's outer tube. Finally, when air mixed with the hydrogen, it exploded and set fire to the graphite in the core. The result was an inferno which sent radioactive debris puring over much of Europe. Despite his support for the accident theory, Prof Kemeny criticised the Russians for failing to build pressure domes over the reactor core. 'I stand by my belief that the Chernobyl reactor was safety-deficient,' he said. 'American, German, French and British reactors have pressure vessels and strongly reinforced concrete structures to contain such radiation releases.' But other nuclear experts cast doubt on the turbine theory. 'I don't think it is the whole story,' Mr Potter said. 'The explanation begs some questions. Why didn't the alternative back-up power supples switch on automatically, and what caused the power surge which the Russians say occurred at the time of the accident? I think there was another factor - concerned with the reactor itself - which was involved but which the Russians do not want highlighted for political reasons. They would find it very inconvenient if it was shown that there were serious generic design faults in all their RBMK reactors, the ones like the Chernobyl reactor. They are not going to let that idea spread'
Please report problems with the web pages to the maintainer