After reading as much as I can find about Y2K. What is the liklihood of seeing early warning signs for 1/1/2000. Say, 1/1/1999 where systems forcast a year ahead, or 7/1/1999 for companies whose fiscal years starts. I know from some of the articles I have read there are some ticklers showing up but would we see a dramatic increase for 1999? Thanks in advance for your input.

-- John Callon (jcallon@gate.net), August 13, 1998

Answers

We will see many early warning signs starting in 1/1/1999. Many programs contain functionality that calculates year end totals and they will begin to crash on that date. In the same vein, many companies and government organizations have their year ends at differnt times, so if their year end is February 1st for example, on February 1/1999 their systems will start running the Feb 1/1999 to Feb 1/2000 cycle, and it not compliant will either crash or provide incorrect data. I read recently that the Chinese Stock Exchange has an anniversary date of January 17th I believe, so in essence they have only slightly more than 5 months to complete their Y2K remediation and testing. Not a chance in hell to put it bluntly! As the awareness of these problems increases, prudent people will be withdrawing their cash, buying supplies etc. likely resulting in a run on the banks. My personal feeling is that this will become serious enough by June or July of next year that the feds will declare martial law. So these people that write foolish articles about how they plan to withdraw a few hundred dollars in December of 1999 to get them over the hump are either extremely naive or else smoking alternative herbal products.

-- Craig (craig@ccinet.ab.ca), August 13, 1998.

Not contributing an answer, but offering a related question that this forum might be able to resolve. I was talking with a COBOL programmer the other day who told me that the number "99" in COBOL is reserved to signify "end of procedure". Any COBOL programmers out there want to verify this, and if true, can we expect serious problems from this source Jan 1,1999? By the way, I REALLY appreciate the discussion in this forum. Thanks.

-- William K. Cheek (bcheek@onramp.net), August 13, 1998.

No - '99' is not reserved in COBOL for anything. But some programmers have used 99 to signal either end of file or invalid date. I know of one particular system that uses 99 as the standin for invalid or missing year. I know that they will have to fix their system in the next four months. But I don't know how widespread this is.

-- anonymous (cobolprogrammer@cobol.net), August 13, 1998.

Often all 9s are used in a key field to signify End of File (EOF), this is a throw back to sequential file design when files were processed/sorted usually in ascending sequence. The value of 999s (F9 in hex) is a higher value than most other byte representations (see IBM 'green card'). The key field would be whatever was designed in often something like Account Number, or some sort of unique idendifier for the record. The program would test for ALL 999s in the read routine, then assume EOF (ie no more records) then perform closure routines. You could have a real Account Number of all 999s. To fail on a date field, it would have to be the key field, with ALL 999s ie 999999, programs would accept 090999 or 010199 (MMDDYY) as valid. The use of 999s in a key field possibly extended to signifying CONTROL records (ie those with a special use eg containing sub-totals) also high values for the EOF record. Sometimes sequential files were converted to databases in their original format retaining the use of 999s where sometimes it was semi-redundant. I doubt whether there will be a 99 problem unless the program only checks say the YY bytes which is unlikely. You see I told you y2k was boring.

-- Richard Dale (rdale@figroup.co.uk), August 14, 1998.

Typo - you could NOT have a valid Account Number of all 999s in the above scenario.

-- Richard Dale (rdale@figroup.co.uk), August 14, 1998.

Apart from technical glitches leading up to Y2k, which may or may not see the light of day, I will be watching for an unusually large number of high-level corporate officers retiring prior to 2000 for "health reasons".

-- Bob Johnson (foo@bar.com), August 19, 1998.