23 Oct
2002
23 Oct
'02
3:57 p.m.
On Thu, 17 Oct 2002, John Lawson wrote:
As I have mentioned here before, based on (much)
research I did while at
MGM on this very subject (35,000 titles on video tape, from 2" Ampex
4-head thru 1" Sony to 3/4" and 1/2" and also some on 1/2 helical-scan
reel to reel and into beta, digi-beta, DLT, CDROM, DVD....all needing a
uniform archival and modern retrieval scheme..) that binary data written
in many paralell micro-film-scale tracks on well-processed virgin Mylar
movie film - 35 or 70 mm - has a longevity >250 years with minimal storage
care and can be read back with simple optical devices that are easily
made. [Lamp - lens - transport - lens/aperture - photocells]
Well, now that's a truly interesting approach to the problem. I'm sure
it integrated very well with their existing film archiving system. What
was your storage density like? How many MB or KB/ft of 35 or 70mm film?
And what about the encoding scheme for the data. Was it a raw dump of
the binary version of the data or did you process it through an encoder
to add error detection/correction?
And the storage of source code, operating systems,
and translation
sortware was included in the scheme.
Computer hardware too?
One of the projects on my "todo" list of the future here at work is a
project to bring online and then rearchive several hundreds of terabytes
of archive data written to Sony DTF-1 magtape.
Even though it's only a few years old, most of the data is worthless
because it's all strung together in such a way that much of it needs to
be online for any of it to be useful. The other catch is that some of
it is dependent on software packages that are no longer readily
available. And even if we could get the binaries, there's no way we
could get current licenses to actually run the software.
When I try to explain what is probably reasonable and necessary to
preserve the most important parts of our data, the accounting folks'
eyes start to glaze over.
-brian.