Hi All, Thank you for the suggestions.
I think the most elegant solution is to have a hook on GitHub that compiles the PDF on a remote server. But it takes a lot more work, because I don't necessarily have *everything* in Git - my local 'library.bib' usually isn't included, nor my custom emacs config, latexmkrc, etc. I'd just like the compiled PDF easily readable by anyone, but I don't want 100s of historical copies. There are a few solutions. 1) Maintain a branch with the 'no-history' files. When they need to be updated, commit and amend, then force-push. See https://stackoverflow.com/questions/22824922/git-commit-and-push-a-binary-file-but-dont-keep-history 2) Add the 'no-history' files to their own commit in the main branch. When they need to be updated, make a new commit and rebase/fixup from the previous commit. See https://stackoverflow.com/questions/12964145/how-to-config-git-to-overwrite-non-text-file-instead-of-version-controlled-it -k.
