government to stop code fixing?greenspun.com : LUSENET : TimeBomb 2000 (Y2000) : One Thread
Talk of contingency planning here is the text from government computer news oct 19, also check out the DOD gearing up for solar storm in same :HTTP://www.gcn.com With roughly 14 months left to prepare systems for 2000, some agencies are beginning to think about what was once unthinkable: setting a cutoff for date code work, whether or not systems are fixed. The idea might seem nonsensical to agencies rushing to finish 2000 work by the March 1999 deadline set by the Office of Management and Budget, but some government officials said the push is coming from users concerned about the confusion that could accompany last-minute changes to systems. Users worry they might be left guessing about what works and what does not, agency officials said.It is possible the administration will impose a ban, said Kathleen Adams, the Social Security Administrations assistant deputy commissioner for systems. The Chief Information Officers Councils Year 2000 Committee has suggested the idea to John A. Koskinen, chairman of the Presidents Council on the Year 2000 Conversion, she said. One of the things weve talked to John about is a moratorium on any changes after a certain point, said Adams, chairwoman of the CIO Council committee. The issue is:At what point? Great Britain and Canada have imposed a moratorium on year 2000 code work after Sept. 30 of next year, said William A. Curtis, the Defense Departments year 2000 project manager. A handful of agencies are considering imposing their own bans on date code work sometime next year to make sure they have workable contingency plans in place come Jan. 1, 2000. The call for a shift in agencies year 2000 efforts is coming from computer users, not technical staffs, Curtis said. Frontline workers are less concerned about whether systems are completely fixed than they are about whether they will be able to do their jobs, he said. Although most systems will likely be ready, a few glitches could stall some programs, Curtis said. Halting code work would give agencies a chance to set contingency plans for doing their jobs, he said. Curtis noted that agencies must test and rehearse contingency plans just as they have had to test code fixes. If you dont know how to use them, theyre of no use, Curtis said. Once you test it and you certify your system, you have to lock them down, Adams said of contingency plans.
-- papa bear (email@example.com), October 23, 1998
I think this is more proof that the govt is clueless about y2k (if you want the REAL smoking gun, look at today's Yahoo article on Gary North's site where the IRS says all mission critical systems will be finished by 1 Jan 1999!!!)
I do govt work, and let me tell you, we code until the very LAST minute to meet a deadline. I have literally been in the engine room of a ship making code changes WHILE THE SHIP WAS UNDERWAY.
As I see it, this article illustrates that the response from govt management is beginning to resemble...PANIC!
-- a (firstname.lastname@example.org), October 23, 1998.
Once unthinkable? This is not uncomon; it's called a 'freeze.'
This is, in my opinion, a realistic approach.
We did it all the time (and yes, it was on government systems) as the time approached for operational testing approached. If it's done right the hard freeze is preceded by a short period of time in which fixes to major problems might get by a review board that would have to approve anything installed. This serves as a throttle on the number of changes being introduced and insures that only the big stuff is hacked at.
It makes sense for several reasons. First, everytime you make a change you run the risk of introducing unforseen problems. Second, as the 2000-01-01 date approaches everyone will be in death march mode, which is when the chance of error is greatest. Third, it gives the using agency 3 months to figure out what it is they'll have to live with..... and if they can live with it. Finally, they'll probably still patch right up until the bitter end, only with less chaos.
-- rocky (email@example.com), October 23, 1998.
Given that, how do they handle "patches" or fixes or repairs during the freeze?
This isn't a usual developemnt process where the single thing is released as a single program. That interface/interfrence seems uncontrollable if a frezze exists - some of the frozen programs will freeze with prtial fixes in place, some with bugs in lace that will cinflict with corrected programs.
-- Robert A. Cook, P.E. (firstname.lastname@example.org), October 23, 1998.
It sounds as though we will have the worst of both worlds. Contingiency plans take time. They requier the same steps as y2k correction:awareness,inventory,planning,etc including testing and a complete dedication to implementation with an actual date. We must either fix y2k or set up a plan;one or the other. I believe it is too late for either.Very discouraging.
-- Jan S C Czarnecki (email@example.com), October 23, 1998.
Rocky makes sense, even though I'm not a programmer. I would think it's then not really time for contingency planning, but establishing alternate procedures. (I see contingency planning as a "what if?" question rather than "what now?".} I have an ugly feeling the alternate methods will be less than adequate in many cases.
-- Mike (firstname.lastname@example.org), October 24, 1998.