> HTML, XML, and ZIP files aren't going anywhere (or there'll be a collapse of society so thorough that print books are unlikely to survive, either)
I very strongly disagree. The amount of technical overhead required to render a simple zipped XML from some storage medium is immense. Even if you store this on tape, you need a tape machine to read the bits. If you manage to pull this off, you won't get anything useful, because the XML is zipped. Deciphering a ZIP and translating it into bits which only make sense if you know that they represent binary numbers and must be read according to an obscure translation table (UTF-8) which is stored also digitally, someplace else, is more difficult than deciphering the enigma code, the hyroglyphs, and the Inca knot language combined. And even if you manage to pull off this unbelievable feat, you still require a parser, a renderer, a display or a printer to even get close to the possibility for a human to read it. For all of that, you not only need electricity, but the devices required are usually easily broken, and manufacturing them requires knowledge and a technical overhead (and also an amount of electricity) orders of magnitude larger than simply using them. Not to mention all the raw materials required to produce a display, a CPU, etc, which may simply be unavailable to future generations.
Now imagine humankind after a complete societal collapse. Tribe A finds, in some basement that survived the apocalypse, a collection of basic medical books. Tribe B finds, in another basement, an unlabeled hard disk containing the entire Wikipedia, as a zipped XML file. Which tribe is more likely to survive?