Being able to make copies easily and store them in different location are probably the best protection for any document. And now days copies are virtually free

Being able to make copies easily and store them in different location are probably the best protection for any document. And now days copies are virtually free and moving them is extremely easy. I'm wondering if someone as done studies on the last technical revolution of this kind, aka the printing press. To which extent the printing press helped prevent the definitive loss of documents? There are probably a tons of factors to take into account, like history, age, kind of documents, etc, but it could be quite interesting. Lastly, as a though experiment, I'm wondering how an archeologist would react in front of a disk drive in a thousand years. Let say the disk drive is in a good condition, but the archeologist has to discover everything: 1) First he has to understand that the disk contains binary data spread in a certain way, if he has equivalent technology or better than today, this should work. (if he realizes that he should be extremely cautious when opening the disk drive, at least for spinning rust). 2) Then he could have to understand the logical layout of data on the medium (aka the FS), this step is kind of optional has you can recover fragment easily, but it would help a lot. And if data was stored in a DB, it's a second layer equivalent to the FS. 3) Then you have to understand the file format, if it's directly a simple TXT file, it should be doable, and it's probable someone can deduce the encoding used, ASCII is easy, UTF-8 is a bit more complex but doable. But, and it's a big but, if the data is compressed in anyway, and a lot of formats are these days, I skeptical that the archeologist (and his technical assistance) would be able to deduce it's a compressed format and not some random garbage, and then re-discover the algorithm used. And if it's a more complex document, like images (even if BMP should be doable), sound and video, it starts to get really tricky. 4) And then, you need to decipher the language and/or alphabet used. Basically it feels a lot like Champollion needing to decipher 4 layers of hieroglyphs before getting the content. Deciphering one can already be hard enough. That being said, as source code is often recorded as simple ASCII files which should be decipherable, he could get lucky and get the code parsing the layers, or at least get a specification of it.