Academics and other people that rely on scientific publications. Most of the world's knowledge in science is locked into PDFs and screenshots (or even pictures)

Academics and other people that rely on scientific publications. Most of the world's knowledge in science is locked into PDFs and screenshots (or even pictures) of manufacturer's (often proprietary) software... So extracting it in a more structured way would be a win (so HTML may not be best). On a related note, I've seen people using Okular to convert PDF tables to a usable form (to be honest its table extraction tool is one of the best i've seen despite being pretty manual).