On 23 Mar 98, David Gillett wrote:

><snip>
>3.  Errors due to non-Y2K-compliant systems can only be in terms of
>getting the century wrong.  Last year, a member of Congress claimed
>that systems like the DMV or SSI might mistake 25-year-olds for
>75-year-olds and vice versa (amongst a list of other improbable
>consequences...); this CAN'T happen.
<snip>

Actually errors of mistaking a 25-year-old for a 75-year-old could
happen.

In COBOL, calculation fields will not save the sign of an equation
unless the program specifies that one should be used.  Since we are
dealing with the assumption that the current year will be greater than
the date of birth year, the date fields would be expected to be created
without a sign.

If the computer date year field is defined as two digits and the input
date year field is defined as two digits then a programmer would have
calculated age as (computer date year) minus (input date year).

00 - 75 can equal 75  in COBOL.  So in the year 2000, some programs
might calculate the age of someone born in 1975 as 75.

Insurance companies and others have been able to avoid problems with
people over 100 years old by requiring that their primary insured is 18
or older.  I imagine that individuals over 117 need to be processed
manually.

-Alan Geist
[log in to unmask]
>
>