No. The part of the spec you quoted says that if the local part of the email address is in the format "atext+ (\. atext+)*" where atext is "Any character except controls, [spaces], and specials", then the quoted-string form shouldn’t be used. In other words, don’t use quotes when you don’t need them. This has nothing to do with how to "interpret" dots; they are interpreted like any other char except they can’t occur everywhere in the local part (e.g. "[email protected]" isn’t valid).
That's pretty clearly stating (to me) that the dot-atom is locally interpreted. It doesn't say anything about how to interpret a dot-atom. Just to use it in preference to the quoted string if the rules you mention apply.
It just so happens that Google decides to treat them as the same for incoming mail.
Apple is under no obligation to treat them as the same. Neither is any other web service.
If you expect the web services you use to treat them as the same, then I foresee major disappointment in your future.
For security, I would prefer the more stringent interpretation over those that are more forgiving.
Not really. It's left open to interpretation from my reading of the relevant RFC. The spec says the dot-atom form should be used but does not say in what way it should be used. Google collapses x.y.z@gmail and [email protected] to the same thing which is fine they're welcome to do so. However the Apple IDs [email protected] and [email protected] are entirely different entities. So we have a namespace collision in one space but not the other because although an Apple ID looks like an e-mail address it's not. I'm just saying that can create a problem.