[{"author": "Tommy Pauly", "text": "Sounds fine for remote
", "time": "2022-03-23T12:01:35Z"}, {"author": "Colin Perkins", "text": "Can hear you well on the remote
", "time": "2022-03-23T12:01:37Z"}, {"author": "Theresa Enghardt", "text": "Meetecho, can you turn up the audio in the room a bit, please?
", "time": "2022-03-23T12:02:00Z"}, {"author": "Meetecho", "text": "Theresa: checking
", "time": "2022-03-23T12:02:26Z"}, {"author": "Meetecho", "text": "Theresa: is it a bit better now, or still too low?
", "time": "2022-03-23T12:04:00Z"}, {"author": "Theresa Enghardt", "text": "I think it's fine now
", "time": "2022-03-23T12:04:09Z"}, {"author": "Theresa Enghardt", "text": "Thank you!
", "time": "2022-03-23T12:04:12Z"}, {"author": "Meetecho", "text": "Ack, thanks for the heads up!
", "time": "2022-03-23T12:04:25Z"}, {"author": "Thorben Kr\u00fcger", "text": "By when should the review be completed?
", "time": "2022-03-23T12:10:58Z"}, {"author": "Thorben Kr\u00fcger", "text": "Can't commit as yet, let's take it to ML
", "time": "2022-03-23T12:12:24Z"}, {"author": "Tommy Pauly", "text": "Schedule looks good!
", "time": "2022-03-23T12:14:25Z"}, {"author": "Martin Duke", "text": "HTTP strikes me as one where the mapping is pretty unclear
", "time": "2022-03-23T12:16:40Z"}, {"author": "Brian Trammell", "text": "i mildly object to the \"the\" in that sentence...
", "time": "2022-03-23T12:17:28Z"}, {"author": "Brian Trammell", "text": "http is weird (is it a pseudotransport, or is it a session layer protocol, or is it an application on top of TAPS, or all three?)
", "time": "2022-03-23T12:18:06Z"}, {"author": "Jake Holland", "text": "are you using CONNECT or POST?
", "time": "2022-03-23T12:18:39Z"}, {"author": "lpardue", "text": "its a transfer protocol
", "time": "2022-03-23T12:18:42Z"}, {"author": "Martin Duke", "text": "It could certainly be over TAPS, but I was under the impression that having TAPS sit on top of HTTP was a goal
", "time": "2022-03-23T12:19:02Z"}, {"author": "Jake Holland", "text": "are you doing long polling?
", "time": "2022-03-23T12:19:08Z"}, {"author": "lpardue", "text": "I do long trolling Jake
", "time": "2022-03-23T12:19:22Z"}, {"author": "Jake Holland", "text": "GrumpyOldTroll was a troll first, sorry. nyah.
", "time": "2022-03-23T12:20:07Z"}, {"author": "Thorben Kr\u00fcger", "text": "are the minor audio glitches of tommy on my end or has anyone else been experiencing those?
", "time": "2022-03-23T12:20:10Z"}, {"author": "Alan Frindell", "text": "ooh that's tiny writing
", "time": "2022-03-23T12:20:28Z"}, {"author": "Theresa Enghardt", "text": "Yes it's glitching a bit
", "time": "2022-03-23T12:20:41Z"}, {"author": "Hang Shi", "text": "what is the relationship between TAPS and webtransport?
", "time": "2022-03-23T12:21:05Z"}, {"author": "Hang Shi", "text": "Webtransport is TAPS for web?
", "time": "2022-03-23T12:21:26Z"}, {"author": "Michael Welzl", "text": "WebTransport doesn't (currently? intend to?) do TAPS and runs in a browser
", "time": "2022-03-23T12:22:24Z"}, {"author": "Theresa Enghardt", "text": "I think WebTransport has specific documents for HTTP/2 and HTTP/3, so unlike TAPS, it's not a common abstraction on top of different protocols
", "time": "2022-03-23T12:23:55Z"}, {"author": "lpardue", "text": "the IETF is only working on WebTransport protocol. For API, see https://www.w3.org/TR/webtransport/
", "time": "2022-03-23T12:24:32Z"}, {"author": "lpardue", "text": "that target of that API is the web platform, which is a subset of possible endpoints that might want to use webtransport
", "time": "2022-03-23T12:25:42Z"}, {"author": "Martin Duke", "text": "I would think that Webtrans is another transport, and if it chose to expose a TAPS interface an application could use it if it met the requirements
", "time": "2022-03-23T12:25:50Z"}, {"author": "Eric Kinnear", "text": "WT over H2 and H3 are mappings of the same web API/\"thing\" so from that perspective it is trying to be a common abstraction, no?
", "time": "2022-03-23T12:26:08Z"}, {"author": "lpardue", "text": "yeh
", "time": "2022-03-23T12:26:55Z"}, {"author": "Jake Holland", "text": "it would be pretty neat to see a try at a taps api on top of webtransport. in javascript would be cool but anything going to wasm would still be cool.
", "time": "2022-03-23T12:30:00Z"}, {"author": "Alan Frindell", "text": "Is there a way to ask for a stream (Connection) on a particular connection (ConnectionGroup)?
", "time": "2022-03-23T12:31:00Z"}, {"author": "Jake Holland", "text": "(a server side library of some kind would also be cool)
", "time": "2022-03-23T12:31:01Z"}, {"author": "Alan Frindell", "text": "To Lucas' point - Abort can RST_STREAM and/or STOP_SENDING depending on current stream state
", "time": "2022-03-23T12:35:18Z"}, {"author": "lpardue", "text": "to use an example: HTTP will fin the request stream after reuest headers. If I want to cancel a large download response, I need STOP_SENDING
", "time": "2022-03-23T12:35:46Z"}, {"author": "Alan Frindell", "text": "For H3 - is there a way to specify stream priority?
", "time": "2022-03-23T12:35:54Z"}, {"author": "Tommy Pauly", "text": "Yes, Abort should be RESET_STREAM or STOP_SENDING
", "time": "2022-03-23T12:36:41Z"}, {"author": "Theresa Enghardt", "text": "Alan, we have a \"connection priority\" parameter which should work for this
", "time": "2022-03-23T12:36:52Z"}, {"author": "Alan Frindell", "text": "or both
", "time": "2022-03-23T12:36:53Z"}, {"author": "Eric Kinnear", "text": "Lucas: Seems like this one of those places where TCP isn't as expressive, and it would be worthwhile finding where the general limits are there rather than just historically what's existed
", "time": "2022-03-23T12:37:11Z"}, {"author": "lpardue", "text": "to answer the main question: this work seems worthwhile to do in TAPS
", "time": "2022-03-23T12:39:27Z"}, {"author": "lpardue", "text": "it's listed in the QUIC-WG related docs, you need to fix your regex :D
", "time": "2022-03-23T12:42:56Z"}, {"author": "Jonathan Lennox", "text": "It's because it's called draft-taps-quic-mapping, the regex's think it's a draft for QUIC by Mx. Taps.
", "time": "2022-03-23T12:44:22Z"}, {"author": "Tommy Pauly", "text": "Lol that's not a great failure mode
", "time": "2022-03-23T12:44:44Z"}, {"author": "Michael Welzl", "text": "@Alan: is there a way to ask for a stream on a particular connection: yes, you clone from it (from the first \"Connection\" that you get to the destination). Stream priority: yes.
", "time": "2022-03-23T12:45:34Z"}, {"author": "Jonathan Lennox", "text": "Is there a way to say \"even if this is going to an existing destination, it should use a fresh connection\"? (Presumably useful in a failure recovery case.)
", "time": "2022-03-23T12:46:40Z"}, {"author": "Michael Welzl", "text": "@Jonathan: Yes :-) by not cloning but just doing Initiate to the same destination.
", "time": "2022-03-23T12:47:20Z"}, {"author": "Jonathan Lennox", "text": "Ah - the impression I got was that initiate would create a new stream on the existing connection if possible
", "time": "2022-03-23T12:47:39Z"}, {"author": "Tommy Pauly", "text": "Like TAPS water!
", "time": "2022-03-23T12:47:57Z"}, {"author": "Alan Frindell", "text": "Thanks Michael. That seems a little awkward - what if I have an open QUIC connection to a destination but no active streams on it?
", "time": "2022-03-23T12:47:58Z"}, {"author": "Michael Welzl", "text": "Nah, it's the other way round - Clone does that. If Clone doesn't see a mux'ing transport, it becomes like Inititate.
", "time": "2022-03-23T12:48:23Z"}, {"author": "Tommy Pauly", "text": "Alan - you'd be able to hold onto the connection group for this
", "time": "2022-03-23T12:48:30Z"}, {"author": "Alan Frindell", "text": "ok - clearly I'm a TAPS noob, thanks for setting me straight.
", "time": "2022-03-23T12:48:53Z"}, {"author": "Theresa Enghardt", "text": "No worries, thank you for asking questions :)
", "time": "2022-03-23T12:49:44Z"}, {"author": "Jonathan Lennox", "text": "I have one IETF process quibble about this document, which is related to an issue I was in the rough on for Arch - this draft is listed as being Standards Track, which personally I think is correct, but Arch (and thus the other protocols' mappings) is Informational.
", "time": "2022-03-23T12:50:15Z"}, {"author": "Jonathan Lennox", "text": "Sorry Impl
", "time": "2022-03-23T12:51:06Z"}, {"author": "Aaron Falk", "text": "Good question, Jonathan. Needs some thought.
", "time": "2022-03-23T12:53:16Z"}, {"author": "Aaron Falk", "text": "fwiw, there's no normative language other than a MUST in the security considerations
", "time": "2022-03-23T12:55:05Z"}, {"author": "Jonathan Lennox", "text": "There's no RFC 2119 language but I would say that e.g. \"Calling Initiate on a QUIC stream Connection causes the implementation to prepare a new QUIC stream to the Remote Endpoint.\" is normative
", "time": "2022-03-23T12:56:11Z"}, {"author": "Michael Welzl", "text": "That sounds like language that would fit well in the QUIC mapping document.
", "time": "2022-03-23T13:01:37Z"}, {"author": "Jonathan Lennox", "text": "That's quoted from the document, my point is that it's a normative statement
", "time": "2022-03-23T13:02:02Z"}, {"author": "Michael Welzl", "text": "Ah sorry
", "time": "2022-03-23T13:02:23Z"}, {"author": "Jonathan Lennox", "text": "NP
", "time": "2022-03-23T13:02:34Z"}, {"author": "Mirja K\u00fchlewind", "text": "thx!
", "time": "2022-03-23T13:02:42Z"}, {"author": "Zaheduzzaman Sarker", "text": "I like it.. lets call all the taps doc as tapping doc :-)
", "time": "2022-03-23T13:02:45Z"}, {"author": "Zaheduzzaman Sarker", "text": "thanks for the session..
", "time": "2022-03-23T13:02:53Z"}]