The FAA, IBM 3083's and microcodegreenspun.com : LUSENET : TimeBomb 2000 (Y2000) : One Thread
Last week Gary North had a posting for a Washington post article in which Jane Garvey of the FAA is quoted as saying, "the [system] will transition the millenium in a routine manner." see http://www.garynorth.com/y2k/detail_.cfm/2097
The article states that the microcode in IBM 3083 machines does not keep track of two digit years, "instead, it stores the year as a two-digit number between one and 32." The fallacy, I believe, in the logic of the programmers hired by the FAA is this: The operating system may well retrieve the year from the microcode as a two-digit number between one and 32, the operating system will not however pass this value directly through to the application programs requesting the current date. The operating system will translate the year into the standard MM/DD/YY format we all know and loathe. The year 01 in the microcode will be represented to the application program as 75.
The programmers that wrote the FAA enroute system were not, I'm sure, prescient enough to suspect that the code would be running in 2000. Even if they were, IBM's operating system passed the date to the application in standard format. Standard format for the '70s, anyway.
I would be curious to see if I am the only one that thought of this (highly unlikely) or more probably, where the holes in my thinking are.
-- Dave Tulka (email@example.com), July 27, 1998
If the microcode will survive, you still have the problem of MM/DD/YY format. This is the big problem in the software. This must be fixed. Do date math and when negative numbers result from date calculations, you have big problems.
I think the bigger problem with anything the FAA says or does is their track record. They already spent a billion dollars to fix their host computer system a while ago. Guess what, they could not do it. A billion dollars of your money right down the drain. Now they expect us to believe they can do it in a much shorter time frame while fixing all their other systems as well.
Dave, I find no holes. Remember, in the private sector you are only as good as your last victory. The FAA has a lousy track record.
I wish the FAA well with their latest endevoure.
-- j (firstname.lastname@example.org), July 27, 1998.
Thanks for the reply. My pointed concern is with intra-operating system events (e.g. OLTP logs, System Management Facility records, JES queue, access method timestamps, etc.). I've been around MVS and MVS/ESA enough to know that these are pretty important pieces of the pie.
-- Dave Tulka (email@example.com), July 28, 1998.
I to share your concerns, however, I have never worked in an IBM shop. I work with HP's 3k and DEC VAX/ALPHA. I know that some HP's will not be supported(read that as not gonna work) and DEC is working on a major upgrade to their operation system. Given the Compac takeover I would be very nervious if I ran a DEC shop.
It is my understanding that the microcode not the operating system is the problem with the FAA's old IBM's. If IBM says the microcode is not gonna make it, it isn't gonna make it.
-- j (firstname.lastname@example.org), July 28, 1998.
Good point. It seems to me that there is a lot of confusion about just where the Y2k problem actually resides. I keep hearing stupid statements like "Apple computers don't have a problem" and "UNIX is immune". The statement from the FAA falls into the same category.
Y2k problems can reside in the hardware, the microcode, the operation system, commercial software packages and custom-written applications. The largest category, and the one most prone to having problems, is the custom-written stuff. The FAA looked at the microcode on just one model of computer (albeit a very important model) and anounced that they will not have a problem (with the microcode) until 2007. They still have to investigate and repair everything else - and I'm still not planning on flying anywhere early in 2000.
-- Ed Perrault (EdPerrault@Compuserve.com), July 28, 1998.
I work on HP3000s, too, and AS/400s. I'm moving a 960's worth of data (application won't survive Y2K) to the AS/400.
Does anybody know if this is the same system/application that lost track of all the planes in the Pacific when they tried the 2000 test a couple of months ago?
If so, then they shouldn't be afraid to publicly test it again?
-- l (email@example.com), July 29, 1998.
The FAA's HOST replacement project involves replacing the IBM 3083 processor & the 3082 control unit with the IBM G-3. What about the associated hardware? What is the condition of: 3814 switching system, 3274 control unit, 3380 disc drives, 3087 coolant control unit, 3089 power control unit, 3480 tape drive? If anyone out there can answer these questions, please e-mail me at firstname.lastname@example.org, also read http://home.natca.org/natca/publicsafety/faay2k.html
-- Dan Thorsen (email@example.com), July 30, 1998.