Re: Year 2k

Dan Clemmensen (
Fri, 27 Mar 1998 21:12:29 -0500

Hal Finney wrote:

> Just because the field is dominated by nut cases and people who benefit
> from predictions of disaster doesn't mean that there's no problem, of
> course. To believe that is to fall into the ad hominen fallacy.

My company makes data communicatons switches. We've been in business for 12
years, and we acquired another company which has been in business even longer.
We now have three ongoing product lines and ten or more "mature" product lines
which are still being used by customers but which are no longer being marketed.
There are probably a thousand companies in a similar situation.

We are characterizing all our products' Y2K behaviour. For mature products, we
tell the customers what to expect. For ongoing products, we have already
verified that the latest releases are fully Y2K compliant. In general, our
products are in good shape, except for some products that are based
workstations or PCs. We use UNIX workstations or our network management system.
The oldest supported workstations are Apollos. The workstation manufacturer
doesn't commit to Y2K compliance in the OS.
Some of our switches are based on PCs. Because of the very high rate of
progress in the PC industry, we basically used a new motherboard and BIOS every
year. Guess what? the clock/calendar primitives for a PC are in the BIOS and
depend on clock chips on the motherboard. This means we've shipped product with
ten or so different motherboard configs, and we've had to characterize them
all. The new ones are Y2K compliant. The older ones are not, and will give
confused dates and times after midnight, 1 Jan, 2000. Most of them will work
just fine if they are re-booted after that time. Customers using the
non-compliant PCs must either upgrade or must be willing to take the equipment
out of service for a minute or two at the proper time.