'Here's what I infer, the Jo Anne effect will be a general problem
starting in January 1999, that's 127 days from now. I don't know how
many systems exhibit this problem. In some cases, the company will not
notice the problem except that their accounting systems will lose 8% of
their annual totals every month. Odd, they'll say, why aren't we
earning any money?
______
'From:
kiyoinc@ibm.XOUT.net (cory hamasaki)
16:13
Subject:
Re: DC Y2K Weather Report 90 & "Jo Anne Effect"
Yes, it's a narrow range of problems but still an important discovery.
On Wed, 26 Aug 1998 07:25:54, "Tom Benjamin" <tbenjami@island.net> wrote:
> cory hamasaki wrote:
>
> >There had been speculation on what's the magic of Fiscal Year 1999,
> >Fiscal Year 2000. Jo Anne, in a classic c.s.y2k post, explained
> >that certain financial systems sort (in the sense of 'group' rather than
> >'collate') transactions into last year, this year, next year piles.
>
> Thank you for clarifying this for me. This is a fairly narrow range of
> failures, but I can see how many management information systems will be
> affected by this kind of thing. I understood the problem to be something far
> broader. What happens when a service purchased by someone flops over into
> the year 2000? Can a twelve month insurance policy be written after January
> 1st, 1999? Or will that choke some systems? Can something end before it
> begins?
>
> Tom
Prior to Jo Anne's posting, there was speculation on why Fiscal Year
2000 would be a problem and Fiscal Year 1997 would not. We intuitively
knew that at some point Y2K problems would hit fiscal year processing
but didn't think the logic through.
Jo Anne's expertise is in accounting. She pointed out that inherent in
the definition of a Fiscal year is two bounds, the date that marks the
previous year and the date that marks the next year.
For each transaction, the system must ask two questions: 1) is it before
the start date of this fiscal year? 2) is it after the end date of this
fiscal year?
The failure can occur either when the system generates the end date or
when it performs the comparison.
If the system generates the year of the end date as: 19100 or 1900, we
have one class of comparison problems. If it generates the end date as
00, we have a potential for another class of problems.
In some cases, the date will be generated correctly but the comparison
will still fail. For example, the date might be coded as 100001, year
100, day 001, which we recognize to be January 1, 2000, but the
comparison was coded to ignore the century indicator.
The exact details of the failures are the common Y2K got-cha's, we've
seen them before.... mismatches on the data, reversed comparisons, bad
date arithmetic, erroneous leap year calculations.
Jo Anne's contribution was explaining how these well known problems
will affect fiscal year processing and giving us a heads up.
Here's what I infer, the Jo Anne effect will be a general problem
starting in January 1999, that's 127 days from now. I don't know how
many systems exhibit this problem. In some cases, the company will not
notice the problem except that their accounting systems will lose 8% of
their annual totals every month. Odd, they'll say, why aren't we
earning any money?
The problem will take a jump in September 1999, when the other common
fiscal year switch takes place. This Y2K problem affects accounting
and management information systems, not the line systems that move fuel,
food, and other essentials.
Before the denial-heads (I'm thinking, moshe) jump in and deny the
importance of these systems, I'll tell you pre-emptively, if accounting
and management isn't important, lets shut them down rightnow. ...I'm
still reeling from the denial-spin on the ATC failure in New England, no
planes crashed, the ATC mainframes are not mission critical. I'm
watching the evacuation from the hurricane, they're interviewing people
who plan to 'ride out' the cat-3 hurricane... hey, it's -bks-!
The run-for-the-hills crowd should watch the two dates. Depending on
how the press plays it, the sheeple could start to bleat,
"baaaa-baaaa-it's true.", "baaaa-waaaa-I'm scared."
I'll be counting on -bks- and the other Y2K denial-heads to pool their
funds and run ads in the major newspapers, "Y2K is *not* a problem." and
"Nothing to fear except fear itself."
The answer to Tom's other question is, it depends. Those issues,
expiration dates and such are certainly related but are not what Jo
Anne's article covered. Some of those systems will exhibit problems,
some will not. That's the reason for all the testing.
cory hamasaki 11,823 hours. |
