Thanks. It's true that the eglot code is hard to understand (and based on your links I suspecter hard_er_). I had been provisionally assuming that that was because the author is a much more advanced lisp programmer than me (which is true; I'm fairly sure he's one of those emacs developers that is also an experienced common lisp developer).

In any case, I would prefer that neither become part of GNU Emacs since then the development process and code review will become completely opaque (I'm not sure code review is really a thing once it's in Emacs. Just maintainers with push rights to some repo on savannah or something)