Out of curiosity, would it be possible to implement a combination of pdf.js and annotatorjs for annotating pdfs within hyperstudio without having to convert them into plain text? Hypothesis, who also uses a modified version of the Annotator backend seems to have gotten it working here. PeerLibrary seems to have done something similar as well. If there's interest in that, I'd be willing to hack around on it. Also, has anyone thought of combining this with a reference management engine to make a more collaborative, web-based version of Mendeley or similar research suites?