[hypertele-Xii]

Unbelievable, I've never copied text from a PDF without some error. It's baffling how broken that system is compared to say, text selection in web browsers.

If you start doing fancy CSS tricks you can break that too, of course.

[db48x]

In a PDF, every single glyph could be independently positioned. Most aren’t quite that pathological, but frequently every text run is independent.

[PaulHoule]

A PDF can contain metadata that correctly sequences the characters which is great for cut-n-paste, accessibility and reflowing the text. That metadata is frequently absent and the mechanism for incorporating it is by no means simple.