geealbers
commented
3 years ago My first idea was to surround the words with a span element use CSS to assign them a font that included the character: <span lang="egypt">*wꜣs*-scepter</span>. However, there were two-dozen instances of it and so that would have led to a lot of excess markup in the files that would have hindered editing. I also didn't like the font options I was finding.
My second idea was to used an inline SVG: *ws*-scepter. I liked that the alt text then could be the character (keeping it reasonably human and screenreader readable), and I had complete control over how the character looked, however it would have again led to a lot of excess markup in the files.
My last idea was to add the specialty font on as a fallback for the the body of the page: .quire-page__content p { font-family: "Noto Serif", "Junicode", serif !important; }. I need to test it in other browsers, but this seems to be working nicely. The only real issue then is the quality of the specific fallback font, and the characters don’t really go well with our base font.
Long term, I'm interested in creating a custom font with these characters, and maybe even contributing them back to the open source version.
Essay 4 uses the ꜣ character, which is a small letter Egyptological alef (U+A723). It's Unicode and part of Latin-Extended-D, and generally not represented in many fonts. In the browser, the browser itself will find a fallback for it, but in our PDF output, PrinceXML has no such built-in fallback system. So for the PDF we need to explicitly provide a font or other rendering option that covers this character.