John Leslie writes: > Michael Thomas <mat@xxxxxxxxx> wrote: > > John Leslie writes: > >> Paul Vixie <vixie@xxxxxxx> wrote: > >>> > >>> the principle i've always followed is that > >>> "all communications must be by mutual consent" > >>> ... > >> > >> Excellent principle, Paul. I'd like to put it at the head of the > >> list. > > > > Ok, I'm dense. How do I meaningfully consent to > > somebody for which I have no a priori information > > about their consentworthiness? > > Much the same as you do with the telephone: some people just pick up, > expecting to complain to the telephone company if it's an obscene call; > others check caller-ID, and let an answering machine take any calls > they don't recognize; still others hire a sectretary to screen their > calls... > > > I mean, I can blackhole them after the fact, but until I have some > > information to inform my consent, I'm not sure what this principle > > buys you. > > It doesn't necessarily buy you anything: it's a way to look at what > we're trying to engineer. Well, I don't understand because it sure seems to me that the principle requires omniscience in isolation which is, well, IRTF territory at the very least. Or is this just a covert way of saying that we need an e-Yentl? Note that I'm not against e-Yentl per se. I just question what this principle actually serves from an engineering/design perspective. It would be a lot clearer if the intent is to say that third party introductions are a necessary possibility, that it come out and say that instead of leaving the possibility of oracles explicitly open. Mike