Testing for Y2K compliancegreenspun.com : LUSENET : TimeBomb 2000 (Y2000) : One Thread
We produce a number of embedded products which use real time clocks that, on an individual basis, may or may not be Y2K compliant. Admittedly, these products may not be as complex as those utilized by utility companies or the federal government. However, by simply forcing the clock (to 1/1/00, for example) on a sample of the products in question, and then exercising the standard Functional Test Procedure (or QA Test Procedure) specified for the product, any Y2K bugs (in our systems, at least) are rather apparent.
Perhaps I am understateing the potential impact of the Y2K problem because I can not realistically hope to understand all the applications which can be affected by Y2K, and the capability of all the engineering organizations which administer these applications. On the other hand, it could also be that those who are proclaiming Y2K Armageddon also have the same lack of information but choose to react by overstating, rather than understating, the potential impact. Doesn't the truth probably lie somewhere in between? It seems in fact that no one really knows what the actual impact will be. And for that reason, while caution is indeed warranted, I am as wary of those who say that there is no problem, as I am of those who predict a millenial catastrophe. Don't forget that there were those of us who, back in 1992, proclaimed, "The Decline and Fall of the American Programmer", only to recant that same apocalypctic prediction just four years later.
Paul Bender Senior Electronics Engineer Pulse Electronics A Westinghouse Air Brake Company firstname.lastname@example.org
-- Paul Bender (email@example.com), April 16, 1998
Mr. Bender, I found your comments very interesting. I love it when people with credentials share their opinions on this forum, it helps develope objective data. Here is the rub. You stated that a fairly simple proceedure can be used for testiung the RTC's in the components your firm makes. These tests would easily identify y2k problems. OK, then what? Do you have compliant parts to replace the non-compliant parts? How many of the non-compliant ones have been sold over the last 20 years? How will they be tested and replaced? Do you (the firm) even know where all of them are? Is there time to do the simple testing you described on all the devices with the RTC's you described that are currently in use and production? Please don't misconstrue my list of questions, they are genuine. I believe it would be a great service to many of us trying to get a handle on the embedded systems issue if you were to candidly answer these questions. Of course, all of the usual disclaimers would apply.
-- P. Larson (firstname.lastname@example.org), April 17, 1998.
In addition to the questions asked by Mr. Larson, I would like to ask:
How does one test a system containing one of your controllers when it is embedded in a black box that does not allow for setting the clock to any value? It is my understanding that there are a number of such system (perhaps not using your embedded products, in which case the question becomes more general, but if you could answer it, I would appreciate it.)
I am afraid that the combination of a few failing embedded systems together with cascading failures could bring much of our infrastructure to its knees: air, rail, electricity, water and sewer. I'm a little bit more than somewhat worried, and any encouragement based upon fact rather than wishful thinking would be most welcome.
-- George Valentine (email@example.com), April 17, 1998.
While I have been programming computers in many languages since 1961, I will admit that I have no knowledge about embedded chips. I would be interested in knowing exactly how one would go about testing such for compliance, for example in an elevator, a heating and ventilating system, a traffic light or in an oil refinery. Approximately how long would it take to locate the chip, test it (in place or after removal?), re-insert it or replace, etc? Are we talking minutes or hours per chip? And are there really 50 billion of them? And what kinds of people have the technical skills to do th
-- Dan Hunt (firstname.lastname@example.org), April 17, 1998.
Sematech, the semiconductor consortium, has a very involved test for the Y2K. If I remember correctly it involves 13 or so dates. About 15 pages in all. I don't know if they have published it or if they consider it proprietary. Sematech used to be partilly government funded but I don't know if that is still the case. If it is still government funded I would think anyone could get a copy.
-- Hud Dunlap (email@example.com), January 31, 1999.