I can't comment on the popularity outside math & science research, but [PGF/TikZ] is a well-established and very capable graphics / diagramming language. It's a text-based, LaTeX-based, declarative* language with a lot of extensions for figures ranging from complex mathematics to Gantt charts. It is mostly intended for manual / programmer-specified layout (though with reasonable enough defaults happening automatically) and subjectively I'd say it sets the bar for flexibility and expressiveness in that context. I would definitely not use it to automatically layout 100s of nodes in an ERD, but it would totally be suitable for making the "non-automatable" example from the D2 roadmap.
(Err, "declarative with an asterisk" because the "declarative" macros do eventually get converted into imperative instructions, but that's invisible to the author/programmer unless they want to dig deeply.)
I can't comment on the popularity outside math & science research, but [PGF/TikZ] is a well-established and very capable graphics / diagramming language. It's a text-based, LaTeX-based, declarative* language with a lot of extensions for figures ranging from complex mathematics to Gantt charts. It is mostly intended for manual / programmer-specified layout (though with reasonable enough defaults happening automatically) and subjectively I'd say it sets the bar for flexibility and expressiveness in that context. I would definitely not use it to automatically layout 100s of nodes in an ERD, but it would totally be suitable for making the "non-automatable" example from the D2 roadmap.
Anyway, just a thought.