What’s so hard about PDF text extraction? | Hacker News