Silicon Investor (SI) -- The First Internet Community

STOCKTALK

We've detected that you're using an ad content blocking browser plug-in or feature. Ads provide a critical source of revenue to the continued operation of Silicon Investor. We ask that you disable ad blocking while on Silicon Investor in the best interests of our community. If you are not using an ad blocker but are still receiving this message, make sure your browser's tracking protection is set to the 'standard' level.

Technology Stocks : Compaq -- Ignore unavailable to you. Want to Upgrade?

To: ed who wrote (25992)	5/10/1998 11:25:00 PM
From: rudedog	Respond to of 97611

Ed - It may need little fix on the software on any statement regarding the time These systems don't HAVE any statements regarding time, or any statements at all in most cases. These are not general purpose computers we are talking about. There is no code base to search looking for time statements. I don't think you are giving sufficient weight to the size of the problem EHF is referring to. Let's just look at GM. They have more than 100 plants making everything from bumpers to radios. Each plant has on average 2000 lines (some more, some less). A line is a group of machines which perform a series of operations which take some component from one state to the next in an assembly process. An example of a line is the spindle assembly line at New Departure Hyatt (the bearing division). A spindle assembly line takes a completed spindle subassembly and presses on the bearings. There are several spindle assembly lines for each type of car GM makes. A line has a number of stations, places where a part stops and gets something done to it. some lines have only a few stations, some as many as 50. the average is about 30. The line itself must have controllers at each station to make sure the line is working correctly and moving things along. Likewise the things at the station that handle the parts (referred to as 'tooling') must be monitored and controlled. Finally, the actual station operation is controlled by some kind of computer system. The average number of embedded systems per station is about 5. So ignoring the many systems which monitor the overall plant, heating and cooling, electrical usage, etc. and just looking at hardware that directly touches the parts which eventually end up in GM cars, we have 30 MILLION embedded controllers from thousands of different vendors. Ed, almost none of these are 'standard' computers. They are not PCs, they are not minicomputers. They are black boxes made to a GM specification sometime in the last 25 years. The companies which made them may or may not still be in business. the programmers who designed them may or may not still be available. Just tracking down what these boxes are and what they do was a major effort which occupied many thousands of EDS people for more than 10 years. That knowledge base does not extend to the inner workings of the machines in many cases, and even where it does, there is no common coding standard, programming language, or documentation method. So even though any competent programmer could look at the date problem on any one of these machines and fix it in 10 minutes, THAT IS NOT THE PROBLEM. The problem is that there are not enough competent people to do the digging into the documentation to even find out what machines have a problem, let alone why the problem exists and how to fix it. Many are programmable logic controllers of one type or another. Those should all be easy to fix right? NOT!! PLCs have some of the most idiosyncratic and arcane programming of any devices in the industry, and there is not even any way to search the programs, and no easy way to tell that a series of counter loops and event timers are actually driven by assumptions about the date made late at night by a programmer who was 20 years away from the Y2K problem and didn't give it a thought. And this is just one example.