Re: the names that aren't DNS names problem, was Last Call: <draft-ietf-dnsop-onion-tld-00.txt>

[Ted Lemon <[e-mail]>]

On Jul 24, 2015, at 6:31 PM, John Levine <[e-mail]> wrote:
> Well, OK.  If the plan isn't that we get to look at every name in the
> next round, what names do we get to look at?

If something like this were to have any hope of working, which I am not convinced it would anymore, it would have to be a list of names that came through every so often, and got last called.   Any individual name for which someone raises an objection would have to be discussed.   No global objections.   Any name that doesn’t get an objection raised can proceed through the paid process.   Names that do get objections get discussed, and if there is a valid reason for objecting, we say no.   But I have no idea what that would be.   I think that we would obviously have objected to names like .home and .onion on the grounds that they are in use in a way that makes sense (at least to some of us!).   It’s possible that a process like this would turn up reasons not to allocate names, but it’s hard to imagine it scaling, and I can’t come up with any objective criterion.   I just would prefer that these names be allocated because they _could_ be useful in the future.

Short of that, which I think is impossible, I think 6761 actually accomplishes what we need it to accomplish.   If anything, I would suggest that we tweak the criteria a bit: a name that’s in wide use already qualifies.   A name that is proposed to be used in an IETF protocol that is being actively worked on qualifies, until it’s no longer being actively worked on.   A name that is proposed to be used by a protocol being actively developed by another SDO or an open source organization qualifies.   The term should be generic, not a trademark, e.g., I would say that “.gnu” fails here because it’s too close to being a vanity TLD, even if the use that’s proposed for it is otherwise legitimate.   But we’d have to debate that.

I think that one of the reasons people are resistant to this is that in fact it really does suck that it’s either the IETF or ICANN that has to do this, and it could indeed be a DoS attack on the IETF in theory.   But I think that most of the DoS thus far has been arguments over points that we could probably reach consensus on, and needn’t discuss every time.   And I don’t think we can get out of this responsibility—it’s not ICANN’s business, I think, to judge whether a _protocol_ use for a special-use TLD makes sense.   I don’t mean to impugn the qualifications of ICANN participants; I just mean that it’s the wrong hat for them to be wearing.