Why Y2K fixes won't work and why there will be a reccession, according to a 12 year computer Information Technology Managergreenspun.com : LUSENET : TimeBomb 2000 (Y2000) : One Thread
This information was provided to me on 11/23/99 by a dependable friend who has been in computer industry for 12 years,started as a Test Engineer, then became a Project Manager. Now, Thurston (I will leave his full name and company name out for now even though he gave me permission to use it if necessary) is a Information Technology Manager at a manufacturing company in Silicon Valley.
" I do share a lot of same opinion in Y2k problem with you. From my past experience in the computer industry, I have two points to share with you and others.
1) Y2K fixes may not work. It cannot be tested for real until year 2000. I have been a Test Engineer for computer software and hardware for several years. Software testing in principle is impossible to do in a complete manner. In software methodology, a complete test is defined to be a test covering all possible inputs which are even more than an astronomical number. In today's software industry, most testing are done in a way that trys to simulate a tiny percentage of the possible inputs. In other words, many possibilities cannot be tested and therefore, many Y2K fixes may not work as expected. Let me use an example to illustrate this problem. Windows 95 had several releases. Why? The first version had problems. Microsoft did as much testing as possible on the first release but there are many problems still. So they have another version which is better. But there are still problems. Now they have patches to that version. But they still have problems. Now they put the fixes on their web sites so that you can get it yourselves. Another example is Windows NT 4.0 which has 5 major bug fix patches so far. Each patch fixes some problems and sometimes a patch will create more problems. It is well known that in software industry fixes can be a problem itself. Y2K compliant software is like Win95. How can you guarantee that a Y2K fix works? It can only be tested for real when the clock clicks to year 2000. It is naive to think that a Y2K fix will work without problem. I think a lot of Y2K fixes will still have problems. Since all software is going to be tested for real at the same time, you can expect a lot of failure will occur to Y2K compliant software due to incomplete testing.
2) Recession is inevitable due to manufacturing problems. In US and most industrial countries, the manufacturing segment is one of the most important segment of the economy. If there is a problem happened generally in the whole segment, it will turn into a nationwide economic problem. I share the same view with Ed Yardeni that there will be a recession for at least six months because of this problem. This is how it happens. JIT (Just In Time) scheduling is how many companies acquire materials for manufacturing so that a company can keep the stock of materials to a minimum. (This is how my company operates.) They only acquire the materials right before the production phase. In other words, companies do not have extra stock of components for manufacturing in case of shortage. It only takes a shortage in one components to stop the production of the finished goods. Any Y2K glitches in one of these suppliers will have a domino effect on the rest of the material supply chain. In other words, a few company's problem will become a nationwide problem. What kind of problem? The revenue for that company will certainly go down. The revenue of all companies depending on this company will also go down due to component shortage. I am not saying that it will not produce anything. I am saying it will go down. I am assuming human effort will correct some of these problems. But you cannot argue that there will be delays. Lets say 10% of shipment is delayed due to late arrival of components wihich is very possible. Late shipment will cause a drop in revenue. How does it look in the stock market? Bad and it will cause the stock value to go down. If this happens to many companies at the same time, it is going to be a recession and it will take a while to recover. To add to the woe, JIT is actually done by software applications. Many of them may have Y2K problems. If the programs make any Y2K mistake, it will further delay the company's operation and a further drop in revenue is inevitable. Since it will happen to many companies at the same time, a recession is in turn inevitable.
-- Watchman (firstname.lastname@example.org), November 29, 1999
If there is a reccession next year I'll still be living off my preps as here in madison every schmuck working at mcdonald's already has a b.s. or b.a. yet nobody does anything that could really be considered productive work,layoffs would be incredible.
-- zoobie (email@example.com), November 29, 1999.
"Windows NT 4.0 which has 5 major bug fix patches so far"
Make that 6. SP/6 is now up at Microsoft. And keep in mind that SP/3, SP/4 and SP/5 were all supposed to be the "last word" on Y2K compliance. One product. Complex yes, but still only one product.
Tick... Tock... <:00=
-- Sysman (firstname.lastname@example.org), November 29, 1999.
I concur with the Thurston' thesis. His views are my views on the difficulty of simulating end-to-end testing, etc. Testing must happen in real-time to ferret out the countless problems with code logic as well as the data passed from subroutine to subroutine and from system to system. Given the well documented poor track record of computer projects over the years, it is a daunting problem to bring any one of them them in on time - and bug free. The probabilities are truly scary when system interoperability and system interconnectedness are thrown into the equation.
While I do believe that not all unremediated systems will fail, I also am certain that not all remediated systems will perfom as intended. The longer it takes us to fix-on-failure, the worse it can get - and it can get worse by arithmetic proportion. To think that we can throw billions of dollars at the problem and come out smelling like the proverbial rose is, in my way of looking at it, asking too much of men and machines.
Though no one can accurately predict the eventual outcome of Y2K, my current assessment is that we will encounter a really bizarre set of worldwide failures that will tend to amaze even the most resolute. I personally don't believe these failures will accumulate at either extreme of the debate, Polly (0) vs. Doomer (10), but instead will fall somewhere in between.
However, given the questionable state of socio/economic/politico reasoning in the US right now, I'm afraid I can't be too optimistic about our ability to handle any disruptive problems, Y2K or otherwise. I subscribe to Murphy's Law, but I sometimes feel Murphy was too much of an optimist.
-- TruthSeeker (truthseeker@ seektruth.always), November 29, 1999.
This is and always has been a REAL TIME problem.Why wasn't 09/09/99a problem? Most large legacy shops spent 6-8 wks hunting down and eradicating end of file commands instead of remediating code properly.On a direct,personal level I know that the job was botched.Embeddeds--who knows---Cobol--40-50% in guaranteed failure mode. Good luck to us all
-- Get Real (email@example.com), November 29, 1999.
Most of the large companies and government agencies have only worked on the most critical of the "mission-critical" programs.
LAYOFFS are imminent for MANY people in many industries and businesses. I doubt that very few of these people even suspect anything is wrong yet. They are all asleep at the switch.
Entire corporate divisions in some cases are about to be jettisoned in an attempt to SAVE THE CORE BUSINESS STRUCTURE.
It's going to be really ugly.
-- snooze button (firstname.lastname@example.org), November 29, 1999.
Sysman: Don't forget the patch required to fixNT Service Pack 6. They broke some of the IP stuff. Heck, how important was that anyway? ;)
-- Colin MacDonald (email@example.com), November 29, 1999.