I like this concept of “Arnesonian negative space” in contrast to “Gygaxian rules codification.” I suspect that, for most people who question druid metal armor restrictions, it’s not a question of realism, but an objection to a perceived arbitrary reduction in the Arnesonian space. The objection I typically see is not “it’s unrealistic for a druid not to be able to wear metal armor,” it’s “I want to know what happens if my druid character does wear metal armor.” If the answer is that they lose all Druid levels and have to start over as a 1st level character, well, ok. That may be a disappointing answer to a player who wants to play a druid and wear metal armor, but at least it’s an answer; the Arnesonian space is restricted, but at least it’s restricted because a rule says what happens if you do the thing instead of just “you can’t because you can’t.”
I think this may be more frustrating to an audience more familiar with WotC’s D&D. As you noted, in TSR’s D&D there were many such arbitrary restrictions - the only answer to why a magic user can’t use a longsword is also “you can’t because you can’t.” Whereas, in WotC’s D&D, a wizard absolutely can use a longsword. They just wouldn’t want to in most circumstances because they don’t gain whatever benefits the edition being played grants for having proficiency with a weapon. In this context, a rule that just says “Druids won’t wear metal armor” seems jarring, because such restrictions are not as commonplace. A player accustomed to such D&D expects a consequence for doing the thing their class isn’t supposed to do, rather than outright denial of the option to even attempt it.