I agree with most of your points, but disagree strenuously with this one: It's 2016, and these dead, flat, static PDFs are functionally identical to the paper

I agree with most of your points, but disagree strenuously with this one: It's 2016, and these dead, flat, static PDFs are functionally identical to the paper they replaced! There's a lot of value in static PDFs and how we format documents for publication. $CURRENT_YEAR doesn't invalidate good ideas from ($CURRENT_YEAR-n). Just in terms of technology, static PDFs are: - Easily distributed. - Easily archived. - Readable forever, even as platforms change incompatibly. - Readable on any device. 20 years from now, do you really want to have to run an old browser, or debug shoddy JS/HTML/whatever, just to read a paper from $CURRENT_YEAR-20?