2024-08-31
On July 25, 2024, I defended my PhD thesis. On August 14, it was published on the University of Waterloo archives, and I had officially completed my PhD. There are plenty of things that, with an infinite amount of time, I would have improved in my thesis—be it unresolved questions in the research, better nuance in descriptions of how various elements work, or just improved formatting—but at some point, one must call it complete, and so it is.
I knew long before I had finished it that I wanted to create a web-accessible version. HTML/CSS have much better accessibility properties than PDF files, are easier to index and search, and are generally more useful to more people. (Though I realize the reality bares more resemblance to the classic PhD joke: The only people who will actually read your thesis are you, and, if you're lucky, your committee.) I knew that tools for compiling LaTeX to HTML existed, and figured it would be relatively straightforward, since it was simply converting one markup notation to another. It turns out this is incorrect.
I won't go into details for how difficult this process actually is, since it turns out Damien Desfontaines already wrote about this exact problem, and I wouldn't have much to add. My solution was ultimately very similar to his: make4ht, but with the original LaTeX edited for better compatibility, and some post-build shell scripting to fix the broken bits. (Also, running a later version of texlive than Debian ships, and a single line patch buried in a corner of tex4ht.)
The result is a single web page representing my entire thesis. There are still things I would like to improve (e.g., formulas are currently SVG paths, and should probably be MathJax or KaTeX for better accessibility), but I'm overall pretty happy with the result, and glad it at least exists. I might make some improvements over time, but the PDF on the university archives will remain the official copy, and the goal of any edits would be to better reflect that version. Any new research, content corrections, etc., will exist as future publications or blog posts.
If you do happen to read any part of the thesis, and have questions or ideas you'd like to discuss, feel free to reach out. :)