On 2024-05-02 Adam Thompson wrote:
> I wholeheartedly applaud this example of My Tax Dollars At Work!
>
This is actually a very good article with lots of juicy info.
However, unless I'm missing something, they don't show the data to
support their longevity conclusions in Table 2?
I started optical backups with my current automated method in 2003. A
few years ago (so ~ 20 year old discs) I took out the first few hundred
discs I burned (CD-R's at the time) and ran them through my checker.
Only one had an issue and only on one ~80MB archive file. I have a db
that has md5sums for every file so I can easily tell when a file isn't
100% perfect.
That's my empirical result: a 99.99[999...]% success rate on CD-R
longevity. I think will go through the oldest ones again, as I was
meaning to finish it and write a report... When I get free time, lol.
I'm intrigued that their chart puts the most common CD-Rs at double
the life of common DVD-Rs! I hope that's not the case, as most of my
checks have been of my oldest CD-Rs, because I didn't start DVD-Rs until
much later when the price/GB crossed over. But it is something to take
into account when I set my "re-burn after" tuneable.
I also have a very hard time believing stamped RO DVD (i.e. movies you
buy at the store) are 10-20 years. I have a ton of DVDs that are more
than 20 years old and have never seen a failure. Well, kids scratching
them to heck does do them in, but never just bit rot.
Interesting that they put stamped CDs at 50-100, compared to DVD's
10-20... I wonder if the general longer life of CD* over DVD* has a lot
to do with density and bpi (if there is such a thing in optical): less
dense (more dye area per bit) being more forgiving?
I also have a vaaaast CD collection and follow things on discogs quite
closely and can report that a miniscule number of (stamped) (music) CDs
suffer bit rot, which jives with the 50-100 years in Table 2, because
the oldest of these are around 40 years (yikes!). The notorious bit
rot ones result in practically everyone's copy becoming an error-filled
mess in roughly the same 5-10ish year window, usually after around 10-15
years. And they are almost certainly due to horrifically bad QA at the
factory for a specific run, not a general failure of the CD technology
itself.
And I'm finally exonerated!! People always thought I was completely
nuts for only labelling my optical media in the inner circle miles away
from any data-containing layer. (Now they'll just think I'm nuts for
other reasons!)
As usual, if any member has old DVD-Rs doing nothing and want to donate
to the Trevor Cordes Data Backup Project, bring them to a meeting!
_______________________________________________
Roundtable mailing list