CDRs have been around for awhile, any studies on CDR data integrity over time?

[astro]

CDRs have been around for awhile and I would imagine that at this point enough time has passed passed that real world test have been done on CDR data integrity over time, but I can’t really find any data on this.

Any studies or links available on this question?

[astro]

Apparently my google-fu was failing. Here is some data. More delicate than i thought!

At present, stated CD-R lifetimes are estimates based on accelerated aging tests, as the technology has not been in existence long enough to verify the upper range, 20 years. With proper care it is thought that CD-Rs should be readable one thousand times or more and have a shelf life of three to five years. Common handling practices can reduce shelf life to only one or two years.

Real-life (not accelerated aging) tests have revealed that some CD-Rs degrade quickly even if stored normally.[3][4] The quality of a CD-R disc has a large and direct influence on longevity – cheap discs shouldn’t be expected to last very long. Unfortunately, branding isn’t a terribly good guide to quality, because many brands (major as well as no name) do not actually manufacture their own discs. Instead they are sourced from different manufacturers of varying quality. For best results, verify the actual manufacturer and material components of each batch of discs.

An example of a CD-R burned in 2000 showing dye degradation. Part of the data on it has been lost.Burned CD-Rs suffer from material degradation, just like most writable media. CD-R media have an internal layer of dye used to store data. In a CD-RW disc, the recording layer is made of an alloy of silver and other metals — indium, antimony, and tellurium.[5] In CD-R media, the dye itself can degrade causing data to become unreadable.

As well as degradation of the dye, failure of a CD-R can be due to the reflective surface. While silver is less expensive and more widely used, it is more prone to oxidation resulting in a non-reflecting surface. Gold on the other hand, although more expensive and no longer widely used, is an inactive material and so, gold-based CD-Rs do not suffer from this problem.

Paper paste-on labels for CD-Rs have been linked to degradation of the recording surface, although the issue is not without controversy [6]. Permanent markers are commonly used to mark the label side of CD-Rs and DVDs. This practice has been discouraged because it is believed xylene and toluene, common substances in permanent marker ink, can cause surface deterioration.[citation needed] Additionally, volatile organic compounds may be released which will remain inside the enclosed atmosphere of a CD-R’s storage box, causing harm.[citation

[Khadaji]

Interesting info, Astro. Any idea what a more permanent medium might be?

[Bijou_Drains]

I guess memory cards are better but they are not as cheap. Maybe down the road they will sell 300 Gb flash drives that are $10 or less.

[ScatteredFrog]

Heh…I’ve had CD-Rs for years – including a music CD-R someone burned for me in 1995 – and all still work like a charm…except for two, which fizzled out after only a few months – serves me right for getting the cheap-o things in the Best Buy impulse thingy at the cash register!

[beowulff]

Yeah, I have some of the first CD’s I ever wrote - back when drives were $1,000. They’re still fine. I keep them in a binder on a shelf.

[Dewey_Finn]

FYI, David Pogue, computer columnist for the New York Times, reported on “data rot” for CBS News Sunday Morning earlier this month. He defined data rot as when the hardware is no longer available to read older data storage media, or when the software is no longer available to read the file formats. He advises moving digital files to whatever the current hardware is, about every ten years. Personally, I wouldn’t wait that long, but instead would make a habit of moving it every two or three years.

The text of the story is here. And the article includes a link to a Library of Congress webpage on how to preserve digital data.

[Sapo]

Part of the beauty of this practice being that normally changing your data to a newer format also reduces it in physical size by a good chunk. A ton of tapes fit on a diskette. A ton of diskettes fit on a CD, and so on. Think of it as decluttering as much as future proofing.

[Deeg]

I don’t think music CDs are a good test of data integrity. Don’t players have software to smooth over a few data blips caused by scratches?

[Una_Persson]

It’s not statistical, but an amazingly large number of data discs (maybe 25%?) I burned in the early 1990s are no longer readable by any CD or DVD drive, even though they’ve been stored in a protected, humidity and temperature-controlled environment.

[sharding]

I’ve had some fail as well. Not a huge percentage (certainly nowhere near 25%), but maybe around 1-3%. Most of the ones that have failed have been truly cheap generics. I’ve also had a few low-end name brands (e.g. Memorex, TDK) fail. Never had a high end disc (Mitsui/MAM-A, Taiyo Yuden) fail, and those actually make up a larger portion of my collection than anything else. I’m sure it’s possible, though. I have one batch of no-name blanks from Fry’s, purchased somewhere around 1997 or 1998, of which every one I still have has failed (become unreadable).