I feel a need to follow up on this topic for two reasons.
First, to correct a few points in the summary that
Robin Wiltshire posted (I don't know if these problems were in the
original article or Robin's paraphrasing, but I suspect the former.)

Describing the Domesday disc as 'unreadable' is overstating
the case somewhat. It isn't unreadable, merely far more difficult
to deal with than it ought to be. The reason for that is that
no one gave any thought to long-term preservation of the information
until too late in its life. A task that could have been simple, if
carried out at the right time, is now difficult and expensive.
Not impossible by any means - but much harder than it ought to have
been.

There are far more than handful of BBC micros left. We've got
at least 10 taking up too much room here at ULCC. Admittedly,
the Domesday disk requires a special attachment for the laserdisc
of which there were far fewer. But preserving the information requires
only one example, and it doesn't even need to work properly - all
that matters is that the ROM chip is readable. (I paraphrase here myself
a bit, so apologies to those at Camileon - the team working on
this - for over-simplifying a complex task.)

Are there many more examples out there, Robin asks ? Yes,
thousands, if not hundreds of thousands, if not millions.
The 'campaign' in question is the second reason I wanted to
follow up on this message. This article, and others like it,
was part of the publicity surrounding the launch of the Digital
Preservation Coalition, a UK alliance of organisations from all
sectors with an interest in digital preservation. We're trying
to raise awareness at every level of the problem which digital
preservation poses, and the lack of resources being devoted to
it at present. It's not a new problem - at NDAD (http://ndad.ulcc.ac.uk/)
we're preserving data created on computer over 40 years ago. Amongst
other things, our work should ensure that the 1961 census, when it's
released in 2062, can be made accessible far more cheaply than the
1901 census cost to digitise.

Approaches which eschew digital preservation for digital objects
are, in my view, nearly always misguided. Yes, using nickel substrates
does give you a high degree of assurance that information will
survive for a long time. For certain applications - recording
information on nuclear waste repositories is the most obvious - it
is a sensible strategy to back up other, more useful, ways
of recording the information digitally.

But analogue information takes up a lot
of space and is difficult to process or analyse. I can search
a database of millions of documents in seconds to find what I want
if they are kept in digital form. I don't want to wade through
one million nickel slabs (or granite slabs!) to do the same.
If every other record of the information has gone, and if I
really, really care about finding that document perhaps I will go
through that effort. But otherwise, I'll simply decide it isn't
worth it. If the information is preserved digitally, the cost
of access (to me) falls dramatically, and therefore it can
be worth using or preserving even for relatively trivial later
uses.

There's more about the Digital Preservation Coalition at
http://www.jisc.ac.uk/dner/preservation/prescoalition.html

For most archivists and records managers, this isn't someone
else's problem and it isn't tomorrow's problem - it's yours, and
it's here now. It isn't really that difficult to deal with
as long as you do something early, rather than leave it too late.

Kevin Ashley
Digital Archives Group Manager
ULCC