R2O
>>When someone says 'not Y2K compliant' and 'problems' does this mean that 1. the systems have specific KNOWN defects and 2. that the specific KNOWN defects WILL cause (possible) loss of life
or
does it mean 1. the systems (may or probably) have Y2K defects and 2. the system controls things that may, in some failure modes, cause a loss of life.<<
--------
This may muddy the waters further but it appears that there is a real problem with definitions. The following excerpt is from:
The Y2K Crisis: A Global Ticking Time Bomb?
csis.org
MR. WEBSTER: There is a fundamental problem here, and it's captured in an old joke from another field. It comes from the early days of personal computers. There's an old joke that went what's the difference between a used car salesman and a computer salesman?
Difference is a used car salesman knows how to drive and knows when he's lying.
There is a problem that a lot of the reporting is done by organizations that have no clue how to assess the state of their IT organization. They don't know what the truth is. They don't know when they're lying.
And beyond that, you have the simple human incentives that, even if they did know, they would not report the truth unless they felt it was (a) safe and (b) to their advantage, and right now they see it as neither.
DR. RUBIN: Yeah, I was just going to say my field is metrics, and there's a real basic measurement problem. Keith pointed it out.
Number one, it's definitional. I don't want to knock Horn's report card there, but when you read that as an outsider, the figures show what percent of stuff is done.
Does done mean it's done? Does done mean it's done in operation? Does done mean it's like kindergarten done but plays now well with friends who are done? And so there are issues with that.
And number two, people like -- when you ask companies where they stand, they like to say, you know, we're 35% complete. Well, how do you tell if that's good or bad? Maybe they should have been 80% by this week, right?
So you can't tell. So companies need different metrics. They need progress metrics to show where they stand or where they should be. They need operational metrics to show if things are going to operation.
They need interconnectivity to test their thread. So there are a lot of problems with the reporting. And as Ed and the others will tell you, there are implications when people put the stuff down in writing in their annual reports to the SEC.
The SEC stuff for the banks -- or the Fed. stuff to the banks are still in these boxes about your progress. We don't care how you got the numbers.
----------------
And from that same document:
----------------
It depends again on what you mean by compliant. If you're trying to get down to every piece of equipment, you're talking about a very large number. If you're trying to get down to where you can stay in business, then you've bounded it.
The problem that I have with any of the figures, whether it's $5 billion or $50 billion or $100 billion, or a trillion or however much money is being discussed, is that I do not understand what the baseline is.
I do not understand what the Department of Defense's baseline for compliance is, what the Department of Agriculture's baseline for compliance is. I can't find across the Government the following definition: what is a system?
Some people will talk about a system in terms of all the computers that make up some acronym that ends in S -- something, something, something system. Then some people are talking about the equipment that's sitting on somebody's desk.
And then some people are talking about just the applications. Now, trying to make those intersect with one another and understanding that I have X amount of money at such and such a value here and I'm at this point of progress is very difficult to do.
That's the problem that I have with any of the estimations. |
