I think it's worth thinking really hard why we're in this state, especially since computing pioneers were actually very optimistic that data and computing would be way more personally malleable than it is now(I've been working on a small comic on this theme myself[2]). For example, check out this short demo[3] of Smalltalk where Alan Kay hooks up a single frame from an animation of a bouncing ball to a painting program, to modify that one frame while also monitoring the loop. Smarter than paper, but way more flexible.

I'm about 98.2331% certain that copyright has a major share of the blame, as an effective system for dealing with digitised documents would all but certainly have to involve duplicating and reverse-engineering them in multiple regards. My solution to the HTML / PDF / ePub, etc., document formats is to recompose them as some minimally sufficient document format (often Markdown, occasionally more advanced formats, with LaTeX being nearly always sufficient). This has resulted in a significant detour through questions concerning typography and just what a document is, though in a huge fraction of cases, there's little reason to go beyond paragraphs, the occasional italic/bold emphasis, and section or chapter markings.