Extinct Data
Comments: 1
An article by Simson Garfinkle on technologyreview.com called The Myth of Doomed Data has the Cassandra in me all riled up. The story is audaciously subtitled “The handwringing about obsolete formats is misguided. The digital files we create today will be around for a very, very long time.”
Some will be around for a very long time — and some most certainly will not. I’d guess we’ll average about a 50/50 survival rate over the long haul judging by how I’m personally doing here (being neither cavalier about nor obsessively protective of my data).
Garfinkle has no excuses whatsoever not to be much, much smarter than me about matters technological, but if this article makes you feel better about your data’s long-term chances of survival, you’ve got some unpleasant thinking to do. Talking about the BBC’s failed Domesday Project, where videodiscs of archival data became obsolete and had to be converted at great expense to currently operational formats, he writes
To be sure, this has all been an expensive and time-consuming process. But it has been done, proving that the process is possible. Not all digital material is worth preserving—most, in fact, is not. But Domesday was worth preserving and, as a result, it has been.
Every sentence except the first here is problematic, so let’s just dive right in.
But it has been done, proving that the process is possible.
Possible if you have the resources — the time, the money, the technical expertise. This is true whether you’re the BBC or grandma dealing with e-mail from her family. Keep in mind the potentially data-swallowing void that lies between “possible” and “feasible.”
Not all digital material is worth preserving—most, in fact, is not.
This is true — but who decides what’s worth keeping? I do. You do. Everybody does … so the possibility of preservation has to be available to all of us and must furthermore accommodate a staggering volume and variety of data.
The other problem here is kind of obvious: how do I decide what gets preserved if I can’t even read it in the first place? In the absence of perfect documentation (perfect enough to enable live-or-die data decisions — but then, all your documentation is perfect enough, isn’t it?), preservation choices are predicated on being able to access the data to decide whether it should make it onto the next generation of storage media or not. This access may not even be possible.
I have boxes and boxes of Zip and floppy discs and I can no longer conveniently read either format. I blame myself for the Zips, but floppies? They were standard … inescapable. A lot of work that’s not even 10 years old is almost out of reach. I could transfer everything to another format (CD? DVD?) but I can’t afford the time. I could have someone do it, but I can’t afford the money. You can see where this is leading. Pretty soon, unless you’re the BBC, you’re just as inclined to throw up your hands and shove those boxes of floppies back in the closet as do anything about them.
But Domesday was worth preserving and, as a result, it has been.
This sentence was placed in the article to sound like a kind of proof, but it’s meaningless.
It does not follow that because something is worth keeping it will be kept: “worth” does not lead us to “can.” Those boxes in my closet contain much that’s worth preserving but it may well be that I come up against my resource limits and am ultimately unable to do anything about it. Oh well, the data can’t have been worth it.
To be fair to Garfinkle, I think this article really suffers from sloppy editing (being a blogger, I can get away with this). How else to explain this getting through:
Indeed, for every Domesday Project that has lost its data to proprietary equipment and file formats, it is easy to point to another project for which information created decades ago is still available.
50/50 … just like I said.
He then begins to talk about physical media and talks about ATA drives. How many of us have our critical data from years back stored on those? Anyway, skirting the horrors of removable media, he reassures us about those ATA drives:
If the disk spins, you can frequently get back the data.
Frequently…!? Feeling better about your data now? Go back and read the subtitle of this article again: The handwringing about obsolete formats is misguided. The digital files we create today will be around for a very, very long time.
Maybe that should be “the digital files we create today will frequently be around for a very, very long time.”
The thing to remember is that how frequently “frequently” is may not be entirely up to you to decide.
•••
Posted to General Rants • 2003.12.04 (Thu) • 22:50
Comments
Posted by Fazal Majid 2003.12.05, 03:22
There is a pretty good book called “Digital Dark Ages” that inspired me to take action. With data, as many other things, “the watched pot doesn’t boil over”, data that is actively tended seldom disappears, at least if you follow basic practices like backing up regularly. Offline data stored on tapes orCDs follows the “out of sight, out of mind” principle, and is most at risk. All the commonly cited examples, like NASA’s mission data or the BBC’s tapes, fall in that category.
The real challenge is when responsibility for the data is passed around, for instance with a change of employees or when a family member passes away. For individual users, the answer probably means making an addendum to a will. explaining how to recover all those digital photos and so on.
My personal answer to the potential risk of format obsolescence (acute for me as I use Canon’s proprietary RAW format for my digital photos) is to shoot a significant proportion of my photos in archival black & white.
Post a comment:
Send This Story to an Enemy
• • •
