Are we reading the same posts? I'd say they clearly advocate for the feature always working, even in the Ravenloft example
This isn't true. I've advocated for no such thing. In a previous reply I believe was in response to one of your posts, I stated I don't know what you mean by "the feature always working" because that sounds like the criminal would be constantly sending messages, one after another, 24/7/365, and said that would be weird, which I believe you glossed over in your response, if you had one.
It's also unclear what you think the part of my post you quoted has to do with supporting your false claim. How exactly does my refutation of the idea, repeatedly and incorrectly put forward by
@Oofta, that some of the features have an unstated prerequisite that it must be established at the table independently of the feature that the PC knows or is known by a specific person or people before the feature can be used by the player lead you to the erroneous conclusion that I "advocate for the feature always working"? My position on this is simply the language which it has been claimed the rules contain isn't there and that this claim has been made to mischaracterize the background features in an effort to disparage them. If you think I'm wrong, it should be a fairly easy task to show me where it says this.
I've argued about the Ravenloft example because I think it's a crap example, as I pretty much stated as soon as it was brought up by
@Oofta, because it posits a situation in which the PCs know nothing, which puts it too far outside the bounds of usual D&D play to be useful as an example for discussion and of which I'm highly skeptical. This is why I've asked for actual play examples of the problems with the background features that have been claimed to exist, but no one seems to have any.
What I have tried to advocate for is the features working when, and only when, the players use them by making action declarations that invoke their features, as long as the described actions fall within the game's genre considerations and their PCs have the fictional positioning to take the described actions. If both conditions aren't met, something has gone wrong with the table's consensus on the established fiction and a discussion needs to take place to get everyone imagining pretty much the same thing. Hypothetical examples of dysfunctional play where this step is not being taken don't show there's anything wrong with these features.