What about program interface changes?

greenspun.com : LUSENET : TimeBomb 2000 (Y2000) : One Thread

This is an extract from my book "The Year 2000 Millennium Bug Report" it follows on from a rather large expose' of the FAA problem (believe me it is a problem) the point of the question is this. What happens when we start to change data / date field settings in a program, when that program interface's with a network? The following article will give you a clue as to how dangerous Y2K remediation process's can be.

Timothy Wilbur

August 11th 1997, New York Times, Foreign Desk - no author sited Software Error Crippled Radar In Guam Crash, Officials Say A software error crippled an airport radar system that might have prevented last week's deadly crash of a Korean Air jet in Guam, Federal investigators said today. A device called the radar minimum safe altitude warning normally issues an alert if a jet is flying too low, and officials on the ground inform the pilot. But federal investigators said the system -- at Anderson Air Force Base -- was modified recently (y2k fix?)and an error apparently was inserted into the software. (Bad luck for the 226 people killed; author) "This is just one piece," said the lead investigator, Gregory Feith. "Yes, it would have helped, but this is not as we know it the cause of the crash" (denial or the big silence?)The warning system was not the only piece of equipment not operating on Guam at the time of the crash. The airport's glide slope, which helps guide the pilot to the runway, was out for regular maintenance. (Getting de-bugged, maybe? author)

An indication of things to come, or an isolated incident never to be repeated? Remember what we are facing here, worldwide computer re-programming. With a documented shortfall of more than 800,000 programmers worldwide, an immovable deadline. In addition, the most time consuming segment to come ie, testing. Add to this scenario management pressure on programmers to finish on time, hours of testing and re-testing and testing again, do not forget to throw in the fallibility of fatigued technicians. What if - a few bugs get through; or are inserted by mistake; will they be as deadly as the Guam bug?

-- Timothy J Wilbur (timkaz@nor.com.au), July 23, 1998


If the brakes on a truck fail, or the driver doesn't see a bend in the road in time, that's deadly.

If the road has a high-strength barrier on the bend that can stop a runaway truck, that's good for the driver and anything in the way. If there's no barrier there, you can hardly blame that for the crash.

This is exactly the same as what happened at Guam.

BTW the pilot could hardly have not known that the glide slope indicator was out (as you say, it was scheduled and notified in advance and in any case would have shown its absence on the instruments). That should not have been a problem either.

-- Nigel Arnot (nra@maxwell.ph.kcl.ac.uk), July 24, 1998.

Moderation questions? read the FAQ