(When we last left our hero, he had a mere 45 GB of photographs to sort through. How did he cope?! We now return to the story in progress...)
There were a whole bunch of obvious duplicates I could weed out. Tee's work habit was to stock iPhoto libraries with all the original camera files in subfolders named "Originals." Those that interested her got copied into the parent folder of "Originals." I used Bridge light table windows (screenshot below) to compare the contents of each dated folder with its Originals sub-folder. Why Bridge? Because it happened to come with my copy of Photoshop. Really, any image viewing program that lets you display large numbers of photographs on one screen with associated metadata, lets you sort them by different criteria, and lets you select ones to move or delete ought to work just as well.
This is where it occasionally got tricky. Most of the time files of the same name in those two windows would be the same photograph, but they'd have slightly different file sizes. Tee's workflow involved re-saving the photographs rather than merely duplicating them. Doing some careful, direct pixel-by-pixel comparisons in Photoshop, I found only the most trivial differences, so I kept the larger version. Occasionally I would find files of the same name where the "copy" was markedly larger. This was a guarantee that Tee had made some sort of alteration. I always preserved both versions of those photos.
This sounds a lot worse than it really was. It was entirely visual work; I'd queue up some jazz, put on some headphones, and zone out to the music while I mowed down duplicates. After a week or so, I'd made my way through all the library folders, and Tee's hard drive had shrunk by another 8–9 GB. Now came the hard work.
Half of the photos resided various project and works-in-progress folders; they usually duplicated or were derived from the iPhoto files. Fortunately Tee often carried the camera's image number through the naming process. I could use Mac OS's Spotlight utility to search on a string like "0194" and scan the list to match photographs across folders. Some projects derived from a couple of specific days' photographs. Once I'd paired up two folders this way, I'd usually find they had a whole bunch of photographs in common. (I also wrote down these folder connections for future scholars.) Others derived from years of photographs and I had to do wild-card searches on frame numbers for most of the individual photographs to find their sources.
Derived photographs could be very subtly different. Tee reworked images in stages, and I needed to preserve all the intermediates. I wrote a time-saving Action in Photoshop to compare photographs. It pasted one photograph as a Difference layer over a second photograph and added a curves adjustment layer that increased the contrast four-fold (below). If there were any differences, they jumped right out.
Once, there were two photographs that had different version letters, but I honestly couldn't see any difference between them. The Action showed that Tee had removed four birthmarks from the subject—not exactly obvious!
I'm pretty well convinced I didn't miss more than 1% of the duplicates and didn't mistakenly delete even a handful of unique images. Now the hard drive was a svelte 30 GB. After I wrote "readme" RTF's for the files in the new archive, explaining to future scholars what I had done and how I organize all the files, I wrote two copies of the archives back onto the hard drive that had been provided. Then I burnt four sets of DVD's on Kodak Preservation DVD blanks. Nothing is certain, but these are as good as you can get. One set stayed with me, one with the executor of the state, a backup copy went to my friend Laurie Edison's home, and the "official" copy to the University.
A few files got munched when I burnt my DVD's, probably 1 in 10,000. Bless Photoshop's Contact Sheet II function for catching them for me. I created an Errata folder, containing fresh copies of the munched photos for inclusion in a catalog CD. I also copied my "readme's" and my folder cross-reference notes to that CD. The bulk of the CD was a set of about high-resolution JPEG files—the full set of of contact sheets generated from the DVD's of Tee's work. 12 images across, 12 images down, and about 130 sheets in all, as illustrated last time. The sum total of an artist's digital life. In two handfuls of gold disks. A strange thing to contemplate.
And equally strange—viewing all those photographs in chronological sequence, over and over, was a little like viewing a stop-motion encapsulation of someone's life. By no means the sum total of those years of living. Still, an important subset of them, sitting on my hard drive, presenting itself on my screen as I scrolled through page after page of images, finally arriving at the very last folder, shown above. Dated 7/27/06, it contained mostly photos of Tee. More self portraits? I have no idea. Either way, they're the last photos in her files.
Tee's life ended exactly one month later.