HTML Reference Docs

This html version of the C standard is cool. It's converted from a pdf by a script, which is probably a reasonably useful technique. There are also html references for x86 and aarch64 instruction which are both nicer to use than their official counterparts (check out this vs this for example). All of these have saved me time and mental anguish.

I've done some of this myself using pdftotext. You can't really parse the output, but there will be consistent patterns, so you can write something like "lines 300-325 is a list" and get html that way. It's hard to completely automate it, but you can reduce data entry to manageable amounts this way, while still having fine-grained control for weird parts.

I've got a couple of projects converting unstructured data to html, and I'll write a longer guide on how to do it once I'm finished.