The Central Mystery Continues

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

The central mystery, of course, being why countries and organizations which did little to remediate seem to have done fairly well.

People (and Ed Yourdon is one) are now writing that we don't know why, and then go on to try to analyze other aspects of the situation. My view is that if we did know why, this would color our analysis of everything, including our evaluation of organizations where a great deal was spent.

I recently went on the Debuker's forum (God help me, fool that I am) in the vain hope of getting a little insight. The reason is that LadyLogic's Mystery Man, whom she wants to get on this forum to debate Steve Heller, actually made some points about use of the system date which were, I felt, reasonably respectable. I wanted to know if he had anything to say about hard-coded dates in programs.

Now there are theoretical reasons for stressing the importance of hard-coded dates, which I won't go into. Suffice it to say that just politely asking the quiestion got me attacked by the harpies from hell, I guess because they knew that I had posted on this forum. CPR demanded to know what I could possibly know about hard-coded dates. The ever clever Mr. Polly told me to stick my head in a toilet and flush. CPR said that he had called my bluff, you moron, that I had come over to set up a strawman.

Well, to get back to my main point, I think we should make a strong distinction between (1) fixing instances that needed fixing and (2) finding them in the first place. For Y2k remediation, I think fixing them is probably routine and easy, and Capers Jones style metrics have led us astray. Finding all of them is what's hard, one reason (not the only one) for my emphasis on hard-coded dates.

I remember a knockdown drag out verbal debate I had on this forum, with either "vtoc" or "a", who were arguing about having to go through every line of code to find... My counterargument used an example from Ed's book involving the variable GEEZER. As I recall, Ed's point, well taken, was that we didn't know without considerable digging whether this variable contained a person's age, or was a date field (eg when a senior's discount program was to be activated).

I argued that if the main business rules of the organization worked correctly. if the GEEZERs missed their discounts, the organization would find out and make it up to them. In other words, fix on failure could work if most of the processing were done correctly.

My participation in this forum has been so sporadic that I have a feeling that others (Flint?) have made this argument too, and I happened to be absent. Certainly it's an argument being made now, and I suspect it will be central, when we do get a handle on the big mystery. But right now, as I say, we don't know and so our speculations will remain shaky.

-- Peter Errington (petere@ricochet.net), January 23, 2000

Answers

I think that most of the organizations that were liable for failure put money and effort into remediation. The places that did not remediate were not so reliant on computers.

Once they did put any effort into remediation, they were likely to use automated software to find dates. The automated software available now is very good at finding dates, finding hard coded dates, finding leap year calculations, and identifying whether GEEZER is a date or an age.

Nevertheless, there are still errors lurking out there that will show up. But you probably won't hear about them.

I think there were posts by PNG on this board that adequately explained why Japan is not having a lot of problems with their business software. They use the year of the emperor's reign as their date, and in general handle dates in a different fashion.

-- kermit (colourmegreen@hotmail.com), January 23, 2000.


Hi, Peter.

Did my response to you on Debunked ALSO annoy you? Did you stop back and notice the response I gave to another that perhaps all those countries DID do something, but we just didn't know about it? The "evidence" was purely anecdotal, but it sure led ME to wonder.

-- Anita (notgiving@anymore.com), January 23, 2000.


Could it be that most of the code is SO modularized that there are one or two modules or DLL's that actually process dates???

Or, in the alternative, perhaps the Citicorp solution of super enhanced rollover processing was used...spokesmen claimed a Thirty Fold increase in programmer productivity. That would just *BLOW* Capers Jones and that "Mythical Man Month" crap out on its ear.

Either way, Pollies WIN, as do Sincere Doomers not looking forward to TEOTWAWKI. Losers are those wishing to leverage Y2K into something else!



-- ~~~~ (Losing it @ Lost it .com), January 23, 2000.


No, Anita, I didn't see your answer and I'll go back and look. You were the only thing about that forum that didn't annoy the hell out of me, so I just decided to bag it.

-- Peter Errington (petere@ricochet.net), January 23, 2000.

To Losing it:

Please direct me, if you can, to where I can find out more about this Citicorp programming environment. That's the sort of thing I've been looking for.

BUT wasn't it Citicorp that spent so Godawful much on remediation, hundreds of millions of dollars?

-- Peter Errington (petere@ricochet.net), January 23, 2000.



Folks,

Nobody can really be sure yet how effective Japan's Imperial calendar has been in dodging Y2K, although many keep repeating themselves "of course!" without even knowing how it may affect or not affect IT processing. At any rate, rest assured that Russia, India, Italy, Indonesia, Germany, Brazil (just to name a few) didn't rely in no Imperial or Royal B.S. to approach the problem. And believe me that these (and other) countries hold the remaining 75% of code outside of the US. So it's not the case that they are not exposed to Y2K either.

Having said that, obviously enough, the one trillion dollar ($1,000,000,000,000) mystery remains Peter... unless, with time (be patient folks!) we all find out the real ethiology and epidemiology of Y2K, and THEN, and only then, would we be able to declare it dead.

We still have EOM, February 29, buffer fill-up, hard-coded dates, etc., etc., plus other non-IT practical consequences of Y2K such as oil/fuel shortage/scarcity which can kick the chessboard sky-high without even knowing what hit us all.

Take care.

-- George (jvilches@sminter.com.ar), January 23, 2000.


Could it be.... that the programmers actually worked hard and fixed most of it, what with having a real deadline and all?

-- Forrest Covington (theforrest@mindspring.com), January 23, 2000.

Losing it,

Thank you for appreciation of "sincere doomers" as winners along with the pollies. You see the only losers as those wishing to leverage Y2k in to something else, and I agree. I'd like to add to the losers list those pollies who were doing the opposite of the "doomers leveraging Y2k in to something else." They would be those pollies who never gained an appreciation for the interconnectedness of the world, and its implications.

-- Suhvivah (y2ksurvivor1@webtv.net), January 23, 2000.


What George said :o)

I'm amazed at the people who bailed already - if we don't have an 8- 10 immedaitely it's a done deal right?

Wrong!

-- Andy (2000EOD@prodigy.net), January 23, 2000.


'Tis Mystery, indeed...

* Local manufacturing plant that processes hazardous chemicals does not complete remediation...including not remediating the global values on their PC's, and putting orange "non-compliant" stickers on better than 50% of their automated process controls that were not assessed.

* They shut the plant down on Dec. 31st, and bring it on-line one piece at a time on Jan. 3rd, ready to fix on failure

* There are no failures whatsoever to date and they are running at 100% capacity

* Those responsible for what remediation was done do not know why there were no failures.

-- (RUOK@yesiam.com), January 23, 2000.



Andy. I don't know diddely about embeddeds and never have indicated otherwise. But for IT, I think it's over.

-- Peter Errington (petere@ricochet.net), January 23, 2000.

Peter,

i SUSPECT that it is at least 10 days too early to declare the IT portion done. So far, Payroll has run once (or may be twice). Bills either have or have NOT been sent out, and have or have NOT been paid, and THAT is something that we will not know about for SEVERAL months as the accounts age, or as the companies panic and sink. Receivable payments either have or have not been processed acurately, again something we won't know about until the accounts age.

I'd LOVE to chuck a shovel full of dirt on it, declare the oil problem an OPEC scam, and stand in the street DEMANDING that Gov Taft DO SOMETHING about gas prices, but I can't. NOT YET.

Check back in about 5 months.

Chuck

-- Chuck, a night driver (rienzoo@en.com), January 23, 2000.


Chuck, congratulations on Ohio's taste in governors, from an ex DoD.

-- Peter Errington (petere@ricochet.net), January 23, 2000.

Moderation questions? read the FAQ