The names you suggest are generic terms that already exist in the respective jargons with different, more generic (and in fact more intuitive) meanings. It's like noticing that the name "Jonathan" has no mnemonic function in the term "Jonathan apple", then suggesting the term "fruit" as a better replacement. This would not solve, but exacerbate the issues raised in article, in that this would make all of learning, teaching and expert communication harder (compare your suggestions with the ham sandwich theorem, a snappy name which serves as an exceptionally good mnemonic device for both the theorem and its usual proof, despite not mentioning Lebesgue measure, Euclidean space, exact division, or really any specific objects involved in the statement of the theorem).
Non-Otto four-stroke cycles abound (cf. Diesel cycle); new industrial methods for ammonia synthesis come into existence every few years and there's not much point in singling one out and calling it _the_ ammonia synthesis loop. The case of shortest-path algorithms has been thoroughly discussed upthread. Your suggestion for Calabi-Yau and Girard's paradox are not coherent with the meanings of these terms.
No big deal: giving a purer meaning to the words of the tribe is not easy, and requires deep domain expertise. And those with deep domain expertise do think about this task, have some strong incentives to think about it (teaching evaluations factor into promotion and tenure) and generally do fairly well at it.
(Startup idea: your-jargon-reform-is-bad as a service, backed by GPT-3?)