Closed pbelmans closed 6 years ago
Without a good way of stripping all TeX this won't be possible: it's the same reason as why one needs \texorpdfstring
after all.
Just a remark, simply rendering TeX code as plain text might be good enough to provide proper <title>
s for most tags, even if it's of course not perfect. A title like Tag 1234: Points of the small \'etale site---Stacks Project
might still be better than the plain Tag 1234---Stacks Project
.
To gauge how bad this ASCII fallback would be in case of the Stacks Project, I just did a quick grep over all sections and subsections. It turns out that we use TeX commands in section and subsection titles for only one purpose, namely accents like in \'etale topology
, {\v C}ech complex
, K\"unneth formula
. If desired, we could even hack together support for these particular uses.
For example, the page https://stacks.math.columbia.edu/tag/02KH could be titled Tag 02KH: Flat base change---The Stacks project.
One needs to be careful about mathematics and other TeX in the title.
Right now Google calls the search result for
flat base change
(Flat base change), which looks a bit silly.