Starting shortly.
", "time": "2023-03-28T08:02:26Z"}, {"author": "Daniel Gillmor", "text": "i can use any topic?
", "time": "2023-03-28T08:03:37Z"}, {"author": "Notification Bot", "text": "Daniel Gillmor has marked this topic as resolved.
", "time": "2023-03-28T08:05:49Z"}, {"author": "David Lawrence", "text": "dbound has resolved? Sweet.
", "time": "2023-03-28T08:08:41Z"}, {"author": "Murray Kucherawy", "text": "Stole my line.
", "time": "2023-03-28T08:10:37Z"}, {"author": "Tim Wicinski", "text": "\"Complement not supplement\"
", "time": "2023-03-28T08:12:53Z"}, {"author": "Pete Resnick", "text": "Or both
", "time": "2023-03-28T08:13:05Z"}, {"author": "Tim Wicinski", "text": "Also I'll be taking notes in the notes tool;
", "time": "2023-03-28T08:13:08Z"}, {"author": "Yoshiro Yoneya", "text": "I respect PSL, but relying on it is a kind of single point of failure.
", "time": "2023-03-28T08:15:11Z"}, {"author": "Tim Wicinski", "text": "This is where I was thinking as well
", "time": "2023-03-28T08:15:43Z"}, {"author": "Andrew Sullivan", "text": "I am not sure I think there is a difference between, \"Here's an alternative way to submit to the PSL,\" and, \"Here's a way to amend the PSL.\"
", "time": "2023-03-28T08:16:31Z"}, {"author": "St\u00e9phane Bortzmeyer", "text": "Also, the PSL is not maintained by the domain holders. That's an authority problem.
", "time": "2023-03-28T08:16:33Z"}, {"author": "Murray Kucherawy", "text": "Doesn't PSL + \"way to amend the PSL\" mean now we have two problems?
", "time": "2023-03-28T08:17:08Z"}, {"author": "Steven Valdez", "text": "It sounds like here's something else that is in parallel to the PSL that entities can choose to consume and combine, but not actually affecting the PSL itself at all.
", "time": "2023-03-28T08:17:08Z"}, {"author": "Geoff Huston", "text": "With respect Paul H, I beg to differ with your assertion!
", "time": "2023-03-28T08:17:23Z"}, {"author": "Andrew Sullivan", "text": "I think \"is this PSL or not?\" is a rathole that doesn't affect whether this is a useful thing to do.
", "time": "2023-03-28T08:17:43Z"}, {"author": "Andrew Sullivan", "text": "Or maybe I should write that \"The PSL\"
", "time": "2023-03-28T08:18:04Z"}, {"author": "Pete Resnick", "text": "@Geoff Huston With what do you differ?
", "time": "2023-03-28T08:18:08Z"}, {"author": "Geoff Huston", "text": "Paraphrasing, that the PSL works well enough for those who use it (directly or indirectly)
", "time": "2023-03-28T08:18:52Z"}, {"author": "Tim Wicinski", "text": "I do feel as we delve into discussing \"The PSL\" we get tied in a Gordian Knot
", "time": "2023-03-28T08:21:29Z"}, {"author": "Andrew Sullivan", "text": "I just want to say that the \"significantly different scaling issues\" is not in fact true. It was one of my main concerns when I started work on this many years ago. It was just that nobody else in the WG thought that mattered.
", "time": "2023-03-28T08:22:56Z"}, {"author": "Andrew Sullivan", "text": "This slide 5 is exactly right.
", "time": "2023-03-28T08:24:03Z"}, {"author": "Andrew Sullivan", "text": "Well, not quite exactly. The additional thing (and on this slide 7 too) is that the DNS was the solution to get rid of a giant hosts.txt file, and any PSL-style list is basically hosts.txt all over again.
", "time": "2023-03-28T08:25:51Z"}, {"author": "Tim Wicinski", "text": "I feel There is also data quality question with the PSL
", "time": "2023-03-28T08:31:55Z"}, {"author": "Daniel Gillmor", "text": "size of the file isn't the only question. the other questions are volume of churn, and latency of change
", "time": "2023-03-28T08:32:00Z"}, {"author": "St\u00e9phane Bortzmeyer", "text": "\"data quality\" = another way of saying that the PSL has an authority problem
", "time": "2023-03-28T08:33:12Z"}, {"author": "Tim Wicinski", "text": "agreed!
", "time": "2023-03-28T08:33:23Z"}, {"author": "Murray Kucherawy", "text": "Non-determinism of what version I have vs. the one you have.
", "time": "2023-03-28T08:33:27Z"}, {"author": "Daniel Gillmor", "text": "that is, how much changes over the course of a month, and how long does it take from when a domain owner wants to assert something to when consumers get to see it
", "time": "2023-03-28T08:33:34Z"}, {"author": "Tim Wicinski", "text": "I think \"Authority Problem\" is a great way to describe it
", "time": "2023-03-28T08:33:54Z"}, {"author": "Daniel Gillmor", "text": "the internet is an \"authority problem\"
", "time": "2023-03-28T08:34:15Z"}, {"author": "Tim Wicinski", "text": "Thanks Ian and Paul for speaking
", "time": "2023-03-28T08:38:09Z"}, {"author": "Murray Kucherawy", "text": "Martin Durst at the mimc
", "time": "2023-03-28T08:39:15Z"}, {"author": "Murray Kucherawy", "text": "mic*
", "time": "2023-03-28T08:39:17Z"}, {"author": "Daniel Gillmor", "text": "@Murray Kucherawy you can edit your texts here in zulip within ~10 minutes if you notice a typo. it's a brave new world
", "time": "2023-03-28T08:40:06Z"}, {"author": "Murray Kucherawy", "text": "Neat!
", "time": "2023-03-28T08:40:20Z"}, {"author": "Murray Kucherawy", "text": "Does that work in the Meetecho client?
", "time": "2023-03-28T08:40:29Z"}, {"author": "Daniel Gillmor", "text": "i dunno, i'm jut using https://zulip.ietf.org/ -- is that the meetecho client?
", "time": "2023-03-28T08:40:55Z"}, {"author": "Lorenzo Miniero", "text": "Murray: no, that's not there yet
", "time": "2023-03-28T08:40:56Z"}, {"author": "Daniel Gillmor", "text": "@Lorenzo Miniero does the meetecho view reflect the edits when they happen?
", "time": "2023-03-28T08:41:52Z"}, {"author": "Lorenzo Miniero", "text": "Currently not
", "time": "2023-03-28T08:42:41Z"}, {"author": "Daniel Gillmor", "text": "gotcha. thanks for working on this :+1:
", "time": "2023-03-28T08:43:33Z"}, {"author": "Tim Wicinski", "text": "\"DNS Industry\"
", "time": "2023-03-28T08:44:00Z"}, {"author": "Pete Resnick", "text": "@Eric Rescorla Do you have concerns about the tractability of the problem or mostly just \"not worth investing the time\" (or something else)?
", "time": "2023-03-28T08:44:47Z"}, {"author": "Warren Kumari", "text": "Thats a really good point!
", "time": "2023-03-28T08:46:24Z"}, {"author": "Benjamin Schwartz", "text": "FYI: https://github.com/publicsuffix/list/wiki/Guidelines#rfc8553-dns-authentication
\n\"LEAVE THESE _psl IN PLACE WITHIN YOUR ZONES POST-VALIDATION IN ORDER TO ANNOUNCE CONTINUED INCLUSION IS DESIRED\".
+1 Andrew; self assertion vs. assertion about child by parent are two different issues, not necessarily compatible and not necessarily soluable within the DNS
", "time": "2023-03-28T08:50:38Z"}, {"author": "Benjamin Schwartz", "text": "So in theory we can already recapitulate the PSL directly out of the DNS just by querying for _psl.www.example.com, _psl.example.com, and _psl.com.
", "time": "2023-03-28T08:50:49Z"}, {"author": "Tim Wicinski", "text": "Great Idea
", "time": "2023-03-28T08:55:08Z"}, {"author": "Ian Williams", "text": "Benjamin Schwartz said:
\n\n\nSo in theory we can already recapitulate the PSL directly out of the DNS just by querying for _psl.www.example.com, _psl.example.com, and _psl.com.
\n
IIRC https://publicsuffix.zone does this, I believe.
", "time": "2023-03-28T08:56:36Z"}, {"author": "Ian Williams", "text": "I don't know if they use _psl leaf records or if they process the PSL directly, though.
Ian: nopes. publicsuffix.zone is a transformation of the actual PSL onto a DNS data structure. I.e. the input is the PSL file and the output is a DNS zone. You can then query any domain's public suffix by querying for PTR at <domain>.query.publicsuffix.zone.
\n(I can tell because I run this :-) )
I think the other problem you have is that you don't know where in the DNS to stop querying, because _any_ leaf could have such an entry.
", "time": "2023-03-28T08:58:24Z"}, {"author": "Benjamin Schwartz", "text": "Alas, most names in the PSL do not currently appear to have these records. The wiki says \"At some point in the future, automation will be used to remove stale entries from the PSL, and missing _PSL records will be a flag used in determination of domains to remove from the PSL.\". Evidently that time is far off.
", "time": "2023-03-28T08:58:26Z"}, {"author": "Ian Williams", "text": "(hi again, Peter! :wave:)
", "time": "2023-03-28T08:58:27Z"}, {"author": "Peter Thomassen", "text": ":wave:
", "time": "2023-03-28T08:58:39Z"}, {"author": "Ian Williams", "text": "@Benjamin Schwartz you broke up in the first bit there
", "time": "2023-03-28T09:00:13Z"}, {"author": "Pete Resnick", "text": "Ben, maybe re-type in the room. Missed the first few words of your comment.
", "time": "2023-03-28T09:00:21Z"}, {"author": "Peter Thomassen", "text": "Ben: _psl records are certainly not needed for ICANN public suffixes, only for private section. So I would not expect _psl.com to ever show up. (ICANN public suffixes are on the list without request by the suffix owner)
", "time": "2023-03-28T09:00:29Z"}, {"author": "Tim Wicinski", "text": "Thank everyone
", "time": "2023-03-28T09:01:15Z"}, {"author": "Andrew Sullivan", "text": "I just don't understand why any of this flows through the PSL at all. The PSL is a way to do something, and dbound could be a way to do something depending on what \"something\" is. I think it would be great if some dbound arrangement became a useful way for PSL maintenance, but that's about all I see as important in such a discussion.
", "time": "2023-03-28T09:01:17Z"}]