With my assistant, I have made a big push to implement my "LaTeX No LaTeX"
project from two years ago. It never really got past a proof-of-concept. The
detailed work was really laborious and exhausting, even if each small step was
satisfying. I caught my assistant during the 3-day window that the Feds allowed
Anthropic's Fable 5 model to run free in the wind, and let him loose.
TL;DR:
* We can make PDF/UA-1 (accessible) PDFs from PreTeXt source.
* We do not use LaTeX at all, MathJax does all the mathematics.
* Tool chain is
- PreTeXt -> XSL:FO (XML spec for 'Formatting Objects")
- XSL:FO -> PDF/UA-1 (Apache FOP processor)
* Development loop is supervised by
- PreTeXt schema
- XSL:FO schema
- FOP processor fails if accessibility is violated
- veraPDF accessibulity checker
Notes:
* 118 commits. ;-) But maybe just 10 hours of work.
* There is a "roadmap" file in /doc if you want more info.
* Lots more work to do - I know about many rough edges. But we have good
coverage of PreTeXt elements and print-oriented publisher switches.
* There are some Debian dependencies, so you may have to fiddle if you want to
try this at home.
* Perfect now maybe for that one-file PDF accessible syllabus your
administration thinks you can easily produce? (If somebody sends me a PreTeXt
syllabus that is not too involved I might be able to build the PDF. Tonight.)
* Presaged on "pretext-a11y" a couple days ago. ;-)
Website Links at
https://pretextbook.org/examples.html
Development test article:
see new "Accessible PDF via XSL Formatting Objects" section
Sample article:
see last link in this group for new PDF
Rob