Patrick O'Beirne wrote:
> At 15:11 03/04/1998 +0200, you wrote:
> >What research do I need? I turn the computer off, and on again and set
> >the date in the BIOS if need be. :-)-O (This is the short version of
> >it, of course :-)-O)
> Very short indeed. It omits the whole question of whether the _software_ works.
> I know you know this, but for the benefit of those less experienced than
> you, here is Y2K 101. The "you" in the following is the generic "dear reader".
> Using two digits for the year will cause software that worked for 97,98, and
> 99 to fail when those digits are compared or calculated with 00.
> This is a software problem. It is only a hardware problem where the hardware
> contains embedded software. A failure in date derived from hardware may or
> may not matter, it depends entirely on how it is being used. Automatic
> processes that act without giving the user an opportunity of intervention
> can be things like PIMs, backup schedulers, process control systems, and so on.
> It is completely predictable whether software will fail - you just need to
> have it tested for all the date functionality; e.g. date comparisons,
> database selections, age calculations, end of periods/years, etc.
> If you do not test, or rely on trustworthy guarantees where you cannot test,
> one of two things will happen to you between now and 2000. Either the
> software will not fail, in which case you have no problems - no thanks to
> you. Or it will fail, and you will have the utter certainty of knowing that
> the consequences are entirely due to your negligence, as you knew of the
> possibility and did not check.
> If you do validate the software, you have demonstrated due diligence and
> should proceed to help others less informed than yourself; especially if you
> depend on them. None of these problems may apply to you, but they may apply
> to someone that you know.
> You cannot make any assumptions. It is possible to write logic in any
> language that will fail to handle 00 dates. This includes user-developed
> applications such as databases and spreadsheets. In my spreadsheet
> workshops, I demonstrate one way of entering dates that is always safe and
> unambiguous; one that treats any two digits as always 19xx; and one that
> shows a person's age as 69 when the spreadsheet is opened in Excel 95 but
> -31 when opened in Excel 97. Most users will not have those problems, but
> that is no comfort to those who have.
> If you have common off-the-shelf packages, an upgrade may or may not be
> required - you have to check. If you have customised or once-off
> applications, you may have to have it repaired, re-engineered, or retired -
> you have to check, nobody else can.
> Best wishes,
> Patrick O'Beirne B.Sc. M.A. MICS. Year 2000 & euro Consultant
> PSP, TickIT, Year2000 assessment on PCs, euro(EMU) conversion
> http://www.iol.ie/sysmod Tel: +353 (0)55 22294 Fax: 22297
> Systems Modelling Ltd, Tara Hill, Gorey, Co. Wexford, IRELAND