just heard on NPR this a.m. about the incidence of unexploded US munitions in Laos running "between 10 and 30% of those dropped";

Point :

historically, munitions are designed and manufactured "to work";

this is done by using simple, proven designs, and simple, proven manufactured and assembled parts and components.

For there to be such a high rate of performance-failure of these 'products' is not a good indicator of the quality of manufacturing of the time - early 60's to mid-70's.

Of course, they did say that some did not work as designed because they were dropped from too low an altitude; regardless of that, the balance still represent an alarmingly high rate-of-failure.

[by comparison, I believe today, Motorola is claiming a failure-rate of less than 5 units for every million cell phones they build.]

My question is: if the rate-of-failure of embedded systems, or other manufactured elements involved in the IT-chain was similar to the percentage-range cited above for munitions, how might that color our perception of the near future?

-- Perry Arnett (pjarnett@pdqnet.net), November 20, 1999

Answers

One problem with the analogy would be that IIRC the munitions being used were WW2 and Korean War stock. Not sure that it was manufactured in the 60's and 70's. Still think the question is good though. Cheers, AGF

-- drac (greenspanisgod@frb.giov), November 20, 1999.

Perry,

I believe you are comparing apple and oranges. The important thing to remember is: the failure rate of time and date sensitive chips in mission critical positions could theoretically be 100%. This is the reason why remediation efforts have been so important, and why the United States Federal government has spent $100 - $114 billion to repair them.

The question I'm asking myself today is, "Of the 1.8 million chips in the world in mission critical areas, how many of those are in the United States, and how many of those have been remediated?"

I have to believe that utilites have been working on this for years, and have replaced the vast majority of those chips/computers. (Did you catch that? I'm trying to convince myself. I'm tired of having nightmares.)

-- Laura (Ladylogic46@aol.com), November 20, 1999.

Hi Laura,

Having worked on embedded systems for utilities, I'm not losing any sleep over large scale utilities failures (in my area) in the short term. They've got a lot of money to throw at it, and I don't think the lights will go out.

What worries me is the mid term effect of the UK's reliance on overseas oil and food, and the local employment effects of non- critical systems going wobbly in the likes of bottling plants.

Here's hoping it won't even come to that.

-- Colin MacDonald (roborogerborg@yahoo.com), November 22, 1999.