Back to index of
personal essaysI discovered this the hard way. I had a lot of digital photos carefully organized, written to recordable CD's (known as "CD-R" disks), stored in a binder, and kept on a shelf for posterity. Not too long ago, I went to get one of my old photos off one of those disks, and I discovered to my dismay that my computer couldn't read the disk. All that data I'd thought was safely preserved forever was gone.
In fact, it turns out that of my entire collection of CD-R's--the oldest of which was almost ten years old--any disk older than about five years was almost certain to give some trouble when reading it. Some were much worse than others. Most of the disks newer than that were fine, but I did even have trouble reading some of my disks that weren't yet two years old.
Let me back up a bit and put all this in context. I'm a complete computer nerd. I've been programming on computers in one way or another since I was at least ten years old. In all that time, I've never willfully thrown away anything I've written; and yet today I have almost none of the work I did as a young boy or as a teenager, or for that matter, even the work I did in college. This is because it's all been lost through one reason or another over the years: floppy disk decay in '82; a hard disk failure in '86; a lost backup tape in '94; a simple "oops" in '98--you get the idea. My mother might well still have samples of my kindergarten crayon artwork stored in a box in her attic, but of my digital past, there is almost nothing. So I'm well familiar with the reality of data impermanence.
This is why I was so excited by recordable CD's. Finally, a non-magnetic data medium! And not only that, but it's a CD! A blissfully permanent CD! By then, of course, everyone was familiar with the CD's indestructibility. Short of cooking one in the microwave, it's nearly impossible to kill a standard audio CD.
I eagerly bought a CD-Recorder drive for my computer when they first became affordable, sometime in 1998. I also bought a film scanner, an expensive gadget capable of scanning 35mm photographic negatives, and converting them into high-quality digital images.
My idea, you see, was to scan in our entire photo collection, so that we would be able to keep our photos in pristine condition, forever. In spite of the problems of impermanent media, digital has the potential to endure for much, much longer than traditional formats. This is because once something is digital, it is no longer tied to the frailties of its medium. It can be copied and recopied to new media, without any loss of quality, as long as each new copy is made before the old copy is lost.
This is the key promise of the digital revolution: immortality. We live in an impermanent world. Everything has a beginning and an end. Physical photos are nice to hold in your hand, and to hang on the wall; but they will eventually fade. Your great-grandchildren may not ever get a chance to admire that wedding photo you've displayed proudly on your mantel. A scrapbook on the coffee table is tactile, real, and satisfying, and there is nothing about it that seems fragile--but it is by its nature one-of-a-kind, and should it burn in a fire, or drown in a flood, it will be gone forever.
Did your great-great-grandfather ever pose for a photograph? If he did, it's almost certainly faded to the point of near-transparency by now. An oil painting would have lasted longer, but still not forever: the paintings of Rembrandt's time, for instance, have by now darkened substantially with age. From the artists of ancient Greece, no paintings have survived at all; 21st century people are likely to believe, incorrectly, that ancient Greek art consisted entirely of sculpture and architecture.
We can't bear to imagine the gradual decay of our cherished images, even though that has always been their eventual fate. But now, for the first time in human history, this is no longer true. Digital photographs do not fade. Fifty years from now, or a hundred years, our digital photos will exhibit exactly the same vivid colors and clarity they have today--and they will continue to do so, for as long as they exist at all.
Assuming we don't screw it up, of course, and lose them completely through an "oops" of some kind. And there's the great irony of the digital revolution: though it promises tremendous potential for complete preservation, it also comes with the risk of complete loss.
In 1998, I decided to take that double-or-nothing gamble--without quite recognizing the full stakes of the gamble--and I began to digitize our entire photo collection, with an eye to preserving it for the indefinite future.
To do that properly, I first had to make several difficult guesses. First, the choice of medium was obviously important: it must be something that would still be readable on the computers in (to pick a number) twenty years. It couldn't be some medium that would be quickly obsoleted by something new. This is part of why I was so happy with the new recordable CD format: I was confident that the CD had achieved a level of market penetration that guaranteed that computers would have CD-Rom drives for many years to come. This is something that couldn't be said for floppy disks, or proprietary formats like Zip disks. Sure, you can still buy a Zip drive today, if you hunt around--but good luck buying a new one in ten years, when the one you have today stops working.
Also, the medium should be relatively impervious to the effects of aging. Again, the CD scored highly here: its reputation for longevity was unsurpassed in the digital world. It turns out that this reputation was not completely deserved, but I didn't know that at the time. There were rumors that CD-R's were expected to last for maybe a hundred years in ideal storage conditions; somewhat less in suboptimal conditions. I didn't mind the "somewhat less"; I figured I should get at least a couple of decades out of my CD's, and so every ten years or so I'd just copy the old CD's to new ones.
There's still more I had to consider, though. The actual data had to be written in a format that the software of twenty years later would still be able to understand. As with the choice of physical medium, this selection of image file format was a bit of a shot in the dark, since it's hard to predict where technology is going. Back in 1998, web pages almost universally used the GIF image file format to display pictures. JPEG was starting to catch on, but it was a relatively new idea, relying on "lossy" compression, which slightly corrupted your image in exchange for making it a much smaller file. In the meantime, Kodak had entered the digital scene with a bang, by introducing its Photo CD format, which stored your pictures in a so-called "visually lossless" image file format that Kodak had invented.
The Photo CD was a tempting choice--here was a disk that was designed specifically for the purpose of archiving photos for the long haul. You'd send your film to Kodak for processing, and they'd develop it in the normal way, but also scan it for you and store the digital images on a Photo CD, which they would mail back to you along with your photos.
Kodak was offering this Photo CD service long before things like film scanners and CD-Recorders were generally available, and for a while it was the very best way you could scan your negatives to digital. I experimented with Photo CD some in 1997, and was very pleased with the results. The only downside was that you really needed to scan the negatives at the time of processing; once the negatives were cut into strips, it was much more expensive to run them through the scanner. So all of our old negatives weren't really an option for Kodak's service.
But now that I had a film scanner and a CD-Recorder, I could produce my own Photo CD's. I considered scanning the images and storing them in Kodak's proprietary Photo CD format, since that was a format that was designed for archival storage, and for that reason alone, chances were good that it would still be recognized in a couple of decades' time. But I decided to go with JPEG instead, after I did a few experiments to convince myself that in its highest-quality mode, it was also "visually lossless".
It turned out to be the right choice. Now that every digital camera in the world can store pictures in JPEG format, JPEG is pretty much the de facto image file format. It will clearly be around for many years to come. On the other hand, the Photo CD has almost completely disappeared from the scene. Some professional photographers probably still use it, so there are certainly some tools to be found that can decode the Photo CD format, but not many, and they're getting fewer and harder to find.
After I had made these decisions, I began to scan in our thousands of photo negatives, convert them to JPEG format, and write them to CD's, which I kept carefully organized by date. I did everything right. I made two copies of each disk, keeping one at my office, and one at home--the office backup was in case there was some catastrophe like a fire at the house. I discovered that occasionally, after writing a CD, the CD was actually not correctly written (even though the computer thought it was), and was partially unreadable. So I wrote a script to write each CD, then immediately reload it and compare it with what I had just written. If it didn't match, I threw the CD away and did it again.
I filled dozens of CD's this way. I bought a CD notebook, the kind that stores hundreds of disks, to keep the CD's safe on the shelf. In fact, I bought several.
In 2002, I finally bought a digital camera. This made the whole process much easier--I no longer had to scan each negative--but I still continued to file the digital images by date and record them to CD in the same way. The CD's continued to accumulate.
Recently, I started adding an additional security step: I would run a program called par2 on the image directories before writing them to CD. This is a brilliant little program that generates additional files in the directory, redundant with the data in the files already there. If some fraction of the original files are lost or damaged, these new files can be used to recompute the lost data. It's almost magical how well it works. Unfortunately, I hadn't been using this program since the beginning, so my oldest CD's, which needed it most, didn't have this extra layer of protection.
With the double backups, I wasn't too alarmed at first when I tried to load a photo from one of the old CD's and failed. No problem, I thought; I'll just bring home the backup CD's from my office, and get the photo from there. It was when I had the same trouble reading the backup CD's that I started to sweat.
Suddenly, I was forced to contemplate the loss of our irreplaceable photos. These included cherished baby photos, the subjects of which will soon be starting college; smiles and hugs from beloved relatives who are no longer with us; picnics my wife and I took together when we were first dating; even the first fuzzy baby ducks of springtime, at the pond where we no longer live. Could these treasures, which I had thought to keep forever, really have been so casually destroyed?
This story does have a happy ending. Thankfully, I was able to recover all of our pictures--every single one. Mostly, this was due to the lucky fact that we didn't buy a digital camera until five years ago, which means that all of our pictures more than five years old came from our 35mm camera, so we still had the negatives. Because of this, I was able to re-scan the negatives for the photos I couldn't recover from our oldest CD's.
If the disks newer than five years old had given me much more difficulty, we would truly have lost photos, since the pictures from the digital camera have no such real-world existence. A handful of the younger disks did give me trouble, but I was able to use the magic of par2 to save me from that.
Interestingly, the very oldest disks that I had--the Photo CD's that I had purchased from Kodak back in 1997--worked flawlessly. These, it turns out, are more than just ordinary CD-R disks; they are special "archival quality" disks with a thin layer of 24K gold instead of the cheaper materials that are used in conventional CD-R's (they have a gold tint, instead of blue or green or whatever you get in other CD-R's). They are also specifically designed with care, to be long-lasting, as opposed to being manufactured as cheaply as possible.
There is just one company in the U.S. still manufacturing these archival quality disks, a company called MAM-A. The cost is roughly one dollar per disk, as opposed to the conventional CD-R's, which nowadays can be bought for about fifteen cents a disk. But even a dollar a disk is cheap. I ordered a stack of one hundred gold archival disks, and filled all of them up with our photos. Now I've ordered a hundred more. Though I'll still use the bargain disks for anything that doesn't need a lifetime measured in years, I won't use anything other than the gold disks to store data that I expect to keep for any real length of time. At the suggestions of some of my friends, I'll also buy a portable hard drive and keep another copy of all of my pictures on that. Hard drives can fail too, but having copies on multiple different media types makes it more likely that at least one copy will survive.
(For those of you who might be wondering at this point, DVD-R uses basically the same technology as CD-R, and so it is likely to suffer from the same longevity issues. But it is also possible to purchase gold archival quality DVD-R disks.)
I realize all this is a lot of effort to put into preserving the past, for the benefit of an unknown future. It's hard to say whether it's worth it--will anyone even care about these photos in fifty years? Well, if my wife and I are still around then, I feel safe in saying there will be at least two people who will.
For now, I'll continue to look to the future, and treasure the past, but I won't forget to live in the present. After all, someone has to take all those pictures!
Back to index of
personal essays
Return to scrapbook index