According to the CBS Evening News, the National Weather Service issued a report that Rockford Illinois was destroyed by a killer tornado this morning. The report was picked up by the media and reported as fact. Rockford is still there, the NWS was just testing a new reporting mechanism. The report should not have been issued. The NWS blames faulty computer software.
The Number 4 ESS system in Dallas went down for much of the day on Wednesday, 25 February 1987, blocking most long-distance calls in and out of area code 214. Both the main system and the backup system failed. One smart company was Fidelity Investor Information, which was able to reroute incoming calls (presumably through an 800 number?) to phone centers in Boston and Salt Lake City. Multilevel layers of redundancy seem like a good practice. [Source: Austin American Statesman, 26 Feb 87, p. D11, courtesy of Steve Smaha, by SnailMail.] [Although presumably not computer related, a highly toxic fire broke out at 3 a.m. on 18 Feb 87 in a Brooklyn NY Tel central office, downing 5 exchanges and 41,000 customers. Because of the toxicity levels, repair personnel were not allowed in the building until after 5 p.m. During the same week, a Chesapeake & Potomac switching center also experienced a toxic fire, forcing evacuation on two consecutive days. See Management Information Systems Week, 23 Feb 87, p. 31 and 54 for details.]
The computer complex at the FAA's en-route traffic control center in Houston went down at 7:13 a.m. on Tuesday, 24 February 1987. Primary radar was restored at 7:45; the manual backup system was in effect throughout the outage. The computer system came back up at 10:40 a.m. Delays of 90 minutes for commercial flights were reported, affecting airports in the surrounding multistate area. [Source: UPI, from SF Chron, 25 Feb 87, p. 3.]
In this decade there have been at least a dozen dock crashes in the Puget Sound ferry system (the largest such system in the USA) that were attributable to onboard computer failures. The damages for one crash alone (12 September 1986) cost an estimated $750,000 in repairs to the Whidbey Island dock. The $17 million mid-sized Issaquah ferries [100 cars, 1200 passengers] came on board in 1980 with the slogan, "Computerized propeller systems make the ferries more fuel efficient." The state sued the ferry builder (the now bankrupt Marine Power & Equipment of Seattle), which agreed to pay $7 million over 10 years. The state's recommendation now is to spend an extra $3 million cutting 6 ferries over to MANUAL CONTROLS. [Source: An article by Deeann Glamser in USA Today, 25 Feb 87.] [It is disappointing that the fix is to bypass the computer systems, rather than to make them work. Nevertheless, accepting reality is clearly a good idea. Although they did not have a gift horse in whose mouth to look, perhaps Seattle still believes in the truth ferry.]
Proper use of suid is easy to characterize: don't use it, use sgid instead! If you need complete security, set up a separate group for each separate application, make the files it needs access to writable by that group, and you're set. [with sgid] Jef Poskanzer unisoft!jef@ucbvax.Berkeley.Edu ...ucbvax!unisoft!jef
I had the good fortune to tour General Electric's Grove City, PA diesel engine manufacturing plant on Friday. The plant manager, who was conducting the tour, was especially proud of the highly automated machine tools and the computerized engine testing cells. They are so confident of the process- control computers' ability to detect problems that the employees in charge of watching the process are allowed to take a break while things keep running. I found this appalling. The fact that the test cells were made of reinforced concrete to shield the rest of the facility from an engine explosion did not make me feel any better. The plant is currently running at less than one third of capacity. I wonder what surprises they are in for if and when it starts running at or near capacity?
The March 2, 1987, issue of the `New Yorker', has a discussion of `people meters' in its editorial column. The two major television audience-rating companies, Nielsen and AGB, are each going to switch from a paper-and-pencil diary system of recording viewing samples, to an automatic electronic system that is connected to the viewing family's television sets and VCRs. There will be some measurement effects: ` ``Here's something that causes us concern, '' Mr. Dominus (a vice-president of CBS) stated. ``To install this system, a man has to wire your house. Let's say you've got two sets and a VCR. He has to literally solder stuff to your equipment. When you walk into the room and turn on the set, you have to punch in, and when you go out of the room you punch out. I would say there's a personality bias toward people with a high-tech style. Now, some people are technology- adverse -- I'm one of them, so I ought to know. They say, `I don't want to do this.' How do you adjust for that mind-set?'' ' Apparently the advertising agencies will want `a money-back guarantee that a given commercial would reach a givena number -- and type -- of viewer.' The networks, because of the unknown nature of the measurement effects, want to avoid giving such guarantees, particularly on $3.7G worth of business, the amount of up-front advertising that was sold last year. They would like to forego guarantees this next year to `save the networks a fortune in unfairly assumed risk.' Toward the end of the article it is revealed that the actual system under discussion is a `real-time electronic diary', instead of a true `people meter', which would function in a totally passive way, leaving no room for human error (such as forgetting to punch in). `Computerized voice identification' and `miniature radio transmitters built into the family jewelry' are mentioned as research directions... Bill
Occasionally, one encounters a truly lovely algorithm. Often they can be recognized by their simplicity. A friend of mine discovered such an algorithm on the Burroughs 6700, lo these many years ago. It all came about because he was debugging a database manager. One day, it attempted to use a somewhat random number as in index into a data file. Now, my friend had a budget, and received bills monthly from the computer centre. The next bill was shocking, and in fact, wasn't even believable. He had been charged for more disk space than the centre owned. It was obvious that the billing software didn't really know how large the files were. Instead, the biller trusted each user program to end at the end of its file. In the true spirit of experimental science, my friend changed his program so that it would always finish by accessing at index zero. And indeed, on the next bill, he was charged precisely zero for disk space.
Please report problems with the web pages to the maintainer