I asked:
>> Are CDs or DVDs at least as reliable as tape for storing executable and
>> data files for 1 year? 10 years? [...] Is there any error correction
>> in the drives themselves (my impression is "no"), or does driver or
>> application software have to do it?
Darren Dunham <> kindly replied:
> The drives certainly have error correction. All CD/DVD data is written
> redundantly.
That's good to know. Although I knew the data was written redundantly on
the media, I wasn't sure whether that was done by the drive itself or the
program writing the data.
>> Suggestions? Recommendations? Things to avoid? Other options I should
>> consider?
> Your question made me search and come up with this document:
> http://www.itl.nist.gov/div895/caref...dlingGuide.pdf
Thank you! I hadn't seen that before.
[Short summary of 50 page PDF file: optical storage *can* last longer than
tape under ideal conditions, but can also degrade quickly in the presence
of water / high humidity, or poor handling. In temperature-controlled
conditions of low 70s(F) and mid-to-low humidity with careful handling,
good original media, and no exposure to UV/sunlight, 5-10 years is entirely
possible. The bigger problem is technological obsolescence, so one should
plan to copy archival stuff to whatever the new, "current" digital media
format is every 5 years or so.]
> They suggest that long term storage of such media should be possible
> under ideal conditions. Also, don't wait too long before using blank
> media.
That, too.
> With tapes, the big worry is apparently that the substrate itself breaks
> down or degrades somehow (I've heard tales of 8-track tapes that have
> basically glued themselves together so that the tape couldn't be unwound
> without damage).
There's that, and, IIRC, there's also the opposite case of adhesive
failing.
> I've also been told that when a tape is archived for a long time, the
> magnetic domains become more permanently magentized.
Actually, I think the effect in question is magnetic "bleed through", where
the magnetization of rolled tape layer N modifies layers N+1 and N-1 and
vice versa. It's kind of like taking the original data and adding an
"echo" of what came K seconds ago and what will come K seconds from now
(where K is 1 rotation of the tape reel).
> Other links I found...:
> http://www2.verbatim-europe.com/inde...d=1&cat_id=347
>
> http://www.manifest-tech.com/media_d...patibility.htm
>
> Details from German magazine test
> http://cdrlabs.com/phpBB/viewtopic.php?p=142231#142231
Thanks!
On a separate, but related subject, Seagate just announced a new, record
recording density in the lab: ~250k tracks/inch with ~1.5M bits/inch per
track (IIRC). Seagate said when this technology reaches the market in
2009(?), 2.5 inch drives will hold 500GB, and 3.5 inch drives will hold
2.5TB. Assuming Blu-Ray disks hold 40GB, that's 63 Blu-Ray DVDs worth of
hard disk space!
-WBE