Print

Print


At 09:28 PM 3/19/2003 -0500, matt Sohn wrote:

>As archivists, we are primarily concerned with preserving audio material and
>artifacts for future generations, yes? What is the best way to do that?

Clearly, the way we did on Voyager. Inscribe the information in analogue on
a gold disc, accompany it with well-crafted graphics and fire it off into
space. Aim a few thousand or million in various directions within our local
galaxy - send a few off to the Magellanic Clouds just to be "certain".

Realistically, we have few data on the longevity of recordings on CD. There
are suggestions of vulnerability of pressed discs. Since all such
weaknesses are shared by recordable media, they cannot have longer life
under the same conditions. I have found no data on longevity of erasable
alloy. I have seen little more than anecdotal evidence of longevity of
write-once media and their dyes. Accelerated-life testing is necessary of
course, but the models do not credibly support the factors leveraging those
tests to lifetime in real storage.

Ultimately, we can ensure that the disc is recorded well. We can secure it
in facilities which we believe to be conducive to long life. And we can
hope that it survives until needed.

In fact, we can do a bit more. We can check samples periodically for
detrioration and as recoverable errors build up and transfer the discs as
needed. We can record two archival copies on different media stored in
different facilities. In short, we can provide protection to gain assurance
based on time and budget.

This thread deals with recording hardware, not with prognostication or even
prediction of data life. Needless to say, whether you write with a Plextor
or a FlyByNight, if you write a good disc, the drive which did it is
irrelevant.

Finally, when you appreciate the nature of "recoverable" errors, you will
undoubtedly understand why they are significant. It is more than but rather
like carrying canaries into a coal mine. The miners do not care about the
health of the canaries; their interest is limited to the phenomena for
which the birds are indicators.


Mike
[log in to unmask]
http://www.mrichter.com/