This analysis of the embedded chip problem appeared on Peter de Jager's forum on August 22.
* * * * * * * *
The problem with infrastructure (embedded systems) is real and potentially huge. In my opinion, it's worse than the mainframe problem. Shavers and VCRs aside, there are real and substantial problems with medical equipment, electronic locks, power generation/transmission/control systems, manufacturing process control systems, etc. What we don't need on January 1, 2000 is a bunch of manufacturing plants or electric power generation plants destructively stopping. Whether they will or not is still a function of public and company awareness of the YR2K potential and TESTING of each system.
The major problems in infrastructure are 1)it is so easy to change the design/microcode/ interfaces/chips/etc. in manufacturing and installing and maintaining embedded systems; 2)that there are so many chip manufacturers; 3) there are NO STANDARDS for manufacturing something like a motherboard so everyone can use any chip any way they want to; and 4) every design engineer KNOWS that their installation is different than anyone else's and works hard to make sure it will be. Since it is so easy to make systems different, each one of the BILLIONS of embedded systems in use worldwide must be individually checked and tested to see if THAT system has a YR2K problem. You CANNOT test one "typical" system and then extrapolate the results to all others from that manufacturer or installer. Having the resources to actually touch every system and the knowledge of what to look for is where the cost comes in. If you are interested in seeing what many people have learned about infrastructure systems, check out my overall evaluation presentation at www.year2000.unt.edu, topic 11 page. I also have several additional presentations available consolidating test results on systems like computer peripherals.
Not trying to flame anyone, but too many people are dismissing the potential infrastructure problems with offhand comments about shavers, VCRs, TVs, etc. There are problems in these systems, but such problems are symptomatic of much more serious problems in the fabric of our lives. We are surrounded and maintained by embedded systems and if even 1% of those billions have YR2K problems, the impacts will be significant. Especially if no one has checked to see and multiple failures catch us unaware. Infrastructure YR2K awareness is where mainframe YR2K awareness was two years ago. But we don't have those two years to bring people up to speed and get the problems fixed. That is also assuming that manufacturers come up with "YR2K compliant" designs.
Dave Hall . . . My opinions only, and usual disclaimers apply
To visit the Web site he recommends, go to: