In a recent comment thread, I started to discuss with Gary Frost how one might archive things for a long time—in essence, make sure that some bit of information we have now might survive for some time and be available to our descendents. (Here I want to append, "If there are any." Is that too cynical?)
I'm hoping that our little misunderstanding turned into an understanding. My main point was that digitised information can, if you maintain it, more cheaply survive local catastrophe than most more traditional forms.
That's not to say that's any sort of panacea. You have to do more work, more frequently, to make that happen. If someone happens to drop a book in a pit it wouldn't be a tragedy; someone else would take it out, more or less intact, somewhat later. (You would be surprised at how many centuries a book can survive in a pit. Possibly as long as a plastic bag.) But putting a floppy disk in your closet can be the kiss of death if, five years later, you have no living computer able to read it.
But, as I said, if you take advantage of the "copyability" of digital stuff, you can keep it for a long, long time, quite cheaply. The key point is that you have a limited amount of time during which you can copy it. In other words: get your stuff into the new format now, because it won't take long for it to become nearly (or entirely) irretreivable.
I'm not sure if this is really the tagline that digital archivists want, but it's the truth: sharks only live as long as they continue to swim.