In the article Refurbishing the Camelot of Scholarship: How to Improve the Digital Contribution of the PDF Research Article, Willinsky, Garnett and Wong points to two things about the current use of PDF (Portable Document Format).
One thing is that all-electronic publications, like OA journals, still use layout, design and typography from the days of the printed journals. Margins are scarce, to avoid long lines text are set in columns, giving a non-optimal reading experience on-screen, too little space between lines etc., all creating a less than optimal reading experience – both on screen and on paper. The current design was created to save space and paper, enabling journals to cram a maximum of content into a minimum of pages to hold distribution costs down. Today, distribution is free, and only interested readers create paper copy – copies that should be as readable as possible, not save paper. And typography and design should acknowledge to need to create documents that are easy to read on a screen.
The other one is to point out the possibilities of embedding structured metadata in the PDF, using existing PDF elements and additional software. And, of course, the need to open PDFs to annotation before publishing them. Embedding metadata will enable harvesters and services to extract necessary information in a much better manner than the current state of affairs – guesswork. And enabling annotation will make PDFs much more useful to end-users.