Hacker News Reader

PDF conversion is a difficult problem. To my rudimentary understanding of the format, PDF documents aren't really plain text, they have no semantic information, they are just individual glyphs and figures mapped to absolute coordinates on a fixed-size page. This format is really meant to be as a portable printer-friendly document format, not ereader-friendly dynamically resizable format.

I can't comment on the RFC part of your comment.

View on HN

PDF started as a pure (100% ASCII!) page description language, but it later was transformed to a binary format, and even later, accessibility support made it easier to retrieve the text from a PDF document (before that, all you would have were the glyphs. For English text, these often mapped 1:1 to characters, but they did not need to) and some support for reflowing documents was bolted on (http://www.adobe.com/uk/epaper/tips/acr5reflow/).

Both accessibility and this reflowing do require participation by the creator of the file, though.

Reply on news.ycombinator.com