For one thing, there's an enormous difference between seeing something backlit or through a film of plastic, and seeing ink on a sheet of paper. Aside from the unnatural shine or glow of the former, there's the problem of color accuracy, because the physics of projected light is different than the physics of reflected light. A computer monitor renders colors by sending out light combined from red, green, and blue pixels. Digital cameras capture images using the same colors: RGB. But, as all schoolchildren learn, the true primary colors of the world — which, away from an electronic screen, is always seen via reflected light— are blue, red, and yellow. For the purposes of printing, this is shifted a bit, to cyan, magenta, and yellow, plus black, a color space known as CMYK. So digital color information has to be translated from RGB to CMYK before it can be printed, a process which is by no means automatic, and in even in the best cases, is imperfect.
Moreover, there's a drastic difference in the way individual monitors render colors, and while some of them can be calibrated, most people don't realize as much, and don't bother to go through the process even if they do. (If you want to try this at home, go into your monitor's settings and change the color temperature from 6500k to 9300k, or play with the gamma a little bit, and see what a difference it makes.) For most purposes, it doesn't matter if what shows up as crimson on my screen appears a little pinkish on yours—though if that blue shirt you bought online turned out to be more sky than navy, you can probably blame your uncalibrated monitor.
It matters even more for art. A decade or two ago, students learning about, say, Stephen Shore, would head to the library and find a book, or perhaps see some slides projected in a classroom. Now, they sit at a computer and call up some images, which is much worse; because what they're looking at is only a very rough approximation of what the photographer actually intended. The colors are wrong, the details are missing, the subtleties have vanished. In the absence of those qualities, one tends to focus on content—on what the picture is of or about— effectively rendering all photography a species of photojournalism. The Ansel Adams you see on a computer monitor didn't make lush, delicately toned nature prints; he took pictures of trees and rocks and stuff.
Perhaps photography itself will change, adapting itself to the new technology the same way it adapted to hand-held cameras and high speed film. But it hasn't happened yet. In fact, photo books are experiencing a boom of sorts. Once, they were treated as documentation after the fact, catalogs of shows in which the actual prints were the important thing. But over the past few decades, the book itself has become the point, a way of collecting and showing images that the artist has no intention of exhibiting or selling as individual prints. In effect, they've become large-edition artworks for entry-level collectors.
At the same time, histories of the photo book are starting to appear, a canon is forming, important titles are being reprinted, and suddenly there's a substantial market. These days, it costs at least $1,800 to get hold of a first edition of William Klein's book of New York photographs; $950 for a hardcover first of Garry Winogrand's Women Are Beautiful; $1,000 for the original printing of William Eggleston's Guide; $1750 for Ed Ruscha's Every Building on the Sunset Strip. And these are relatively recent books. The other day I went looking online for a copy of Facile, a 28-page chapbook published in 1935 by Paul Eluard and Man Ray — perhaps the greatest collaboration between a writer and a photographer ever published. There were two available. A marked up copy was $5,000, but when I checked back again, a few days later, it was gone, leaving only a single clean copy, which was selling for $10,000. I could buy 20 iPads for that kind of money. But I'd rather have the book.