Disasters and Recovery

From: Eric Smith <eric_at_brouhaha.com>
Date: Sun Jan 17 17:26:36 1999

I wrote:
> I wouldn't expect a hard drive to be workable after 100 years of storage,
> even in an inert gas atmosphere.

Doug <doug_at_blinkenlights.com> wrote:
> Why not? Which parts would degrade?

Most everything inside the HDA. Remember the Quantum and Seagate drives that
wouldn't work after only a few months of storage, because of problems with
the lubricant?

Not to mention that hard drives almost all have firmware in flash or
OTP EPROM, and some have calibration information in EEPROM.

IMNSHO, it would be much easier to preserve a state-of-the-art 1970
minicomputer (e.g., a PDP-8 or PDP-11) in operable condition for the next
hundred years than a current PC.

If the 1970 minicomputer breaks, there might still be some suitable parts
in 2100 to fix it. And if there aren't exact replacements, it should be
possible to fabricate replacements. All of the parts are simple and
well-documented.

When a chip in your 1999 PC goes bad, the odds of being able to find a
suitable replacement in 2025 are pretty slim. By 2100 it's unlikely that
you'll be able to get any parts. For instance, what if the 82371AB PIIX4
chip fails? Right now it is easy to get those. But if there weren't any
available, could you program an FPGA to replace it? Certainly it is
theoretically possible, but it's not bloody likely. Do you have the
logic diagrams or equations for the PIIX4?

And mechanical things like disk drives are similar. It would be much easier
to fabricate a replacement part for an RK05 drive than for a Quantum
Fireball 12G IDE drive.
Received on Sun Jan 17 1999 - 17:26:36 GMT