CHEEKY KID: This is so, so silly ... you're really groping on this CHIPS deal.
Alistair Stewart, a senior advisor for Giga on embedded systems, said that ONLY ABOUT 3% OF CHIPS have been found to have minor problems, typically requiring resetting the date or restarting a device. The PERCENTAGE OF CHIPS that experience outright failures is "so small as to be statistically insignificant," he said.
Stewart's an expert on embedded systems, but his comment was about CHIPS. Maybe this will help you understand:-)
BLACK BOXES: EMBEDDED SYSTEMS
A piece of equipment may appear as a black box to an end-user, but contains upon inspection a multitude of "black boxes" (smaller embedded systems) from different vendors inside it. Furthermore, inside some of those smaller "black boxes," they may contain themselves a multitude of tinier "black boxes" from still different vendors and so on.
Each "black box" can have up to ten layers of technology: Chips & microcode; pre-manufacture custom functionality; post-manufacture custom functionality; interfacing of devices; drivers; operating system; vendor-supplied application library; user-defined functionality; user integration of systems and devices; and business process associated with system use.
From a design standpoint, to find out if a suspect device is Y2K compliant, many people have to be consulted for EACH "black box" component in order to cover these layers of technology. These consultants include microprocessor architecture specialists, firmware engineers, operating system engineers, device vendors, and end-users who oversaw the installation and/or modifications of the "black box" in the next larger "black box." The TransAlta document [see below] covers these points in depth and visually.
ourworld.compuserve.com
Vendor statements are seldom available or reliable, either because of ignorance or falsehoods. In addition, the available engineering information is not always pertinent to Y2K compliance requirements or noot available for all vintages (variations) of the equipment over the years (decades).
Also, vendor statements can rarely address end-user implementation compliance issues. That is, a multitude of made-Y2K compliant devices might not be able to work in sync because they were made compliant in different, incompatible ways.
Consequently end-user testing is required. See the GM DOCUMENT, Figures 5.2.4 Year--2000 Component Test Report Form and Figure 5.2.7 -- Year 2000 Combined Component Test Report Form.
ourworld.compuserve.com
The first has up to 48 TESTS to report on for individual component Y2K testing, also called by others "stand-alone component compliance testing."
The second has up to 43 TESTS to report on for combined component Y2K testing, also called by others "implementation compliance testing."
Yet, the GM document cautions: "This limited set of tests CANNOT prove a Component/System to be Year 2000 compliant, but using them will help identify several frequently observed problems. These test procedures are written as general instructions. Specific knowledge of the systems or components under test is required in addition to apply these test cases." Repeated testing is required to rule out fluke successes or fluke failures.
Even in less rigorous Y2K embedded systems testing environments, the number of 30 TESTS per device is commonly recommended.
ENTIRE SERIES ON EMBEDDED SYSTEMS [by R. Martin] AVAILABLE:
y2ktimebomb.com
CHEEKY:
Alistair Stewart, a senior advisor for Giga on embedded systems, said that ONLY ABOUT 3% OF CHIPS have been found to have minor problems, typically requiring resetting the date or restarting a device. The PERCENTAGE OF CHIPS that experience outright failures is "so small as to be statistically insignificant," he said.
A multitude of customized CHIPS are found in each of these "black boxes". Only 3% have problems? Where are they? Which black box? Who's the vendor? Where's the documentation? How easy is this to fix?
Remember, while people/companies have been aware of Y2K software problems for many, many years. And, major corporations have been working on software for many years. Awareness of Y2K embedded system problems first came up in early '97. Only 2 years ago.
Cheryl
FORTUNE Magazine Article - MUST READ!
“Catastrophic on every plant Floor” – General Motors CIO
pathfinder.com
In March '98 the automaker disclosed that it expects to spend $400 million to $550 million to fix year 2000 problems in factories as well as engineering labs and offices.
Figure has now increased to $700+ million. Almost a billion bucks.
|
