TeX has one, and one only, significant problem. And that is its lack of a context-free grammar. You cannot parse it without executing it. That means you have no

TeX has one, and one only, significant problem. And that is its lack of a context-free grammar. You cannot parse it without executing it. That means you have no linting, translation, etc. Does the author really lament that scientists do not use structured representations for publication? Did they ever try that in practice?