Embedded Systems

greenspun.com : LUSENET : TimeBomb 2000 (Y2000) : One Thread

My greatest concern with the y2k problem was allways the embedded systems that {I was led to believe ] could shut down water, electric, pipeline, train swithing, etc. systems within a short time after the rollover. It is apparent now that none of this happend. And from the news I hear third world countries that have spent little time on the problem came through as well North America.

Question? How did so many people get it wrong? This must not have been a problem in the first place. I would realy appreciate some feedback on this subjet. Thanks, Search

-- search (search@conclusion.com), January 03, 2000

Answers

I'm confused about these chips being no problem all over the world as well. Although the experts disagreed on the per centage of chips that would fail (0.1% on the low end, 5% on the high end), they all agreed that many would fail.

A couple of concerning examples: 1) Sen. Bennett stated in the Spring of 1999 that when they rolled the date over in a Colorado water facility, a chip caused the system to over-chlorinate the drinking water to the point that it was poisonous. 2) When they rolled over the dates at a sewage plant in Van Nuys, Ca. somewhere in the neighbourhood (Canadian spelling) of 3 to 4 million gallons of raw sewage spilled down a boulevard and into a park around June, 1999.

Then you got the Jim Lord & "Mr. CEO" spiel about two or three weeks ago. You got oil engineers predicting massive cut backs in global oil production.

Something doesn't add up somewhere . . .

-- Think It (Through@Pollies.Duh), January 03, 2000.


Think It,

On the contrary, it ALL adds up!! North charging $129 a year for his newsletter. Multiply that by 15,000 subscribers and see what you get. Lord charged $99 and probably has 5000 readers. How many books did Yourdon and Hyatt sell? Hundreds of thousands, that's how many! And not a damn one of the bunch had ANY idea what they were talking about. And we fell for it! And on the other thread you get angry because somebody suggest writing letters to North's publisher . You accuse them of trying to take away North's First Amendment rights! Then you come hear and cry about it not adding up!

And North is reading this and laughing at you. Bonehead!

-- Forget It (computergeek75@hotmail.com), January 03, 2000.


IMHO :: As I always predicted the actual date functionality of embedded chips (and their behaviour in unexpected conditions) was wildly over stated. People leant firmly on the side of pessimism without using enough common sense.

"We don't have concrete data absolutely PROVING such systems CANNOT fail" is NOT a valid argument for TEOTWAWKI.

Programmers, IT managers, Engineers - not stupid people. Clearly they have a CLUE that something will fail at certain boundary conditions WHEN it is created. Clearly they either assume it will be replaced or out of service when such a date boundary condition occurs - BUT - they do know about it.

Go and have even a cursory glance at the IBM web site. Look at some of the dates - some of the data that has been available for a long long time now with regards to potential problems.

I think it is fair to say that 99% of all y2k date functionality problems WERE flagged, and organisations were able to leverage this data in their remediation programs.

I never did like the argument that since x% of all software projects fail then x% (or more due to the fixed deadline) of all remediation attempts regards y2k would fail.

This is a rampant over-simplification. Remediation is not as difficult as large software (and hardware) system creation. It is NOT fraught with all of the same problems. Perhaps it is fraught with SOME of the issues such as time management - again wildly over stated.

IMHO :: The pessimists of the computer science community were loud and listened to (and made some bucks). Most of us shut the hell up and fixed things is my guess.

PS. NB// You don't fire nukes when a clock fails.

So far as date calculation - it's pretty clear where it IS present in embedded systems. The data was available. The problems (even if they do occur) are 99% harmless.

All my opinion of course. I got sick of being howled down as an over optimist months ago ;)

[as a side note huge credit to Ed Yourdon should be afforded, I seem to recall a lecturer first making me aware of the problem based on some of his work years ago - guys like Ed should be congratulated for at least making more people aware of the problems.]

ALSO -> As the above note should prove, I am no old time software engineer. In software engineering (and computer science) terms I am a youngster. This COULD well mean that I am just an over optimist who can't see how lucky we all are ;)

-- Fletchman (fletch@picknowl.com.au), January 03, 2000.


First of all, I'm not crying. Far from it - I'm rejoicing. Secondly, the sewage spill really did occur in Van Nuys. Thirdly, Sen. Bob Bennett (R-Utah) did state that the chlorine mishap occurred. You can try to blame it all on the profiteers (they never got any money from (save on Hyatt book I picked up in an airport)) if you want but that doesn't explain the IEEE's Congressional testimony, the sewage spill or Bob Bennett's (who spent the last year trying to calm the American public) statement. It just doesn't add up . . .

-- Think It (Through@Pollies.Duh), January 03, 2000.

Search, If you look down through the archives under Embedded Systems, you will find hundreds of posts describing why embedded chips could fail, the number of embedded chips out there, and the percentages that may have date problems.

However most people read these posts as being: why embedded chips would fail, the number of embedded chips out there, and the percentages that have date problems.

There are posts that say "Do the Math" and multiply all of these figures together effectively showing that the world would crash.

Unfortunately most of these post are NOT be experts, but are merely reports of comments made by a friend in some anonymous industry, or senators or other people who are good at spin. Any post by true experts in the field like Cherri or The Engineer were quickly dismissed as being Polly hype.

The most accurate take on the whole issue was by the Gartner group, where they broke the embedded issue down to types of system, the probability of failure, and the potential impact. I have never seen anyone say that there would be no failures, but just there would would be very failures, and that what did fail would be manageable.

IMHO the whole issue to be faced now is not to lay blame for mis- information that has occurred in the past, but to sift the present information carefully in order to ride out the errors that will occur in business systems.

-- Malcolm Taylor (taylorm@es.co.nz), January 03, 2000.



Fletchman, A couple of points: 1) You're probably right (and you'd know better than me) that 99% of the errors where recognized and engineers leveraged this info for their testing, remediation, replacement, whatever. But surely Gambia wasn't the only country who didn't act on it. Maybe they were, but that sure isn't what the CIA, World Bank, United Nations, et al. were saying last week. 2) Yes you do fire missles when computer systems fail. That's why the Pentagon was so intent on having Russians in Colorado and Americans in wherever Russia or Siberia. 3) A lot of people at the very top with far better access to info regarding the threat acted very concerned. Million dollar bunkers/war rooms constructed all over the world, National Guard on high alert, Red Cross recommending preparations.

-- Think It (Through@Pollies.Duh), January 03, 2000.

First and foremost, I think that the issue is that we still have very little understanding about the machines and systems that we have created, therefore the faulty analysis. Secondly, just because there was no explosion and the lights stayed on doesn't mean that there were no important embedded failures. For example, the entire Iranian oil industry adopted a fix on failure approach. If their equipment failed, it will take a little time for it to be felt. If the railroads get screwed up like they did during that last merger causing coal deliveries to be screwed up, you'll feel it later.

-- Chris Tisone (c_tisone@hotmail.com), January 03, 2000.

I agree with you Think It. I can't understand how the IEEE, MIT, and NIST could all have written articles saying SOMETHING was going to happen, and then have absolutely NOTHING happen.

Not even in the places where it wasn't fixed. The argument that "we fixed it" doesn't hold water when it wasn't even broken in Gambia and Malaysia and Italy!

-- Ariana (pergados@yahoo.com), January 03, 2000.


Moderation questions? read the FAQ