More on the infrastructure end of #citum, the core repo now has ~700 automated tests. But they're not super transparent to humans. So I've integrated a solution for that into the Github CI:
https://docs.citum.org/behavior-report.html
It's not complete, but I'll iteratively add to it so it documents the logic of the entire engine crate.
Another feature I added to #citum a bit ago is #typst integration.
https://docs.citum.org/examples.html#typst-pdf-publishing
It includes a feature-gated compilation option to allow you to use the citum engine as a complete end-to-end djot -> PDF processor (at the expense of binary size). But you can do the same thing with simple shell scripts or pipes.
I'm about done with new feature work; now it's just about polishing and hopefully much more extensive user testing.
Another new feature I've just added to #citum is compound citations that I guess are common in chemistry.
https://docs.citum.org/examples.html#compound-numeric-sets
This is feature requested for CSL awhile back that we had no idea how to implement, and in retrospect would have been impossible given the dependence on processor behavior.
https://github.com/citation-style-language/schema/issues/437
Looking for a clean solution, this is another case where I drew inspiration from #biblatex.
One of the nice things about this new project is the opportunity to look afresh at questions I've sometimes thought about.
This week: annotated bibliographies.
The solution I settled on is to treat annotations as linked to reference styling and data, but also distinct.
Bruce D'Arcus