Wednesday, March 23, 2022< ^ >
Room Configuration
Room Occupants

[11:25:09] Meetecho joins the room
[11:45:39] Harald Alvestrand_web_872 joins the room
[11:45:45] Harald Alvestrand_web_872 leaves the room
[11:45:49] Harald Alvestrand_web_249 joins the room
[11:49:21] Paolo Saviano_web_352 joins the room
[11:51:23] Paolo Saviano_web_352 leaves the room
[11:52:44] Martin Dürst_web_136 joins the room
[11:54:05] WU Hsing-fang_web_231 joins the room
[11:55:38] Chris Ullrich_web_149 joins the room
[11:57:02] Robert Stepanek_web_255 joins the room
[11:57:03] John C Klensin joins the room
[11:57:27] Phillip Hallam-Baker_web_494 joins the room
[11:57:30] John Klensin_web_383 joins the room
[11:57:44] Yeshwant Muthusamy_web_506 joins the room
[11:58:36] Darrel Miller_web_117 joins the room
[12:00:00] Francesca Palombini_web_243 joins the room
[12:01:02] Renan Krishna_web_754 joins the room
[12:01:06] John Preuß Mattsson_web_940 joins the room
[12:02:09] Takahiro Nemoto_web_227 joins the room
[12:02:33] Robert Stepanek_web_255 leaves the room
[12:02:37] Robert Stepanek_web_274 joins the room
[12:02:41] Manu Sporny_web_659 joins the room
[12:03:02] Murray Kucherawy_web_407 joins the room
[12:03:46] Robert Stepanek_web_274 leaves the room
[12:03:50] Robert Stepanek_web_378 joins the room
[12:05:53] <Harald Alvestrand_web_249> checking that chat works
[12:06:01] <Francesca Palombini_web_243> ack
[12:06:46] Francesca Palombini_web_243 leaves the room
[12:06:50] Francesca Palombini_web_226 joins the room
[12:12:53] Alexey Melnikov_web_818 joins the room
[12:13:01] <Murray Kucherawy_web_407> I think I agree with Standards Track.
[12:13:27] <Murray Kucherawy_web_407> and this seems like a good list as well
[12:14:39] Cullen Jennings_web_726 joins the room
[12:16:26] <Phillip Hallam-Baker_web_494> I believe we should apply the principle of permissionless innovation to registration. That is, nobody should ever have to ask permission to deploy an Internet application.
So the question to as is whether registration of a toplevel type is necessary to deploy an application. The answer is no.
Toplevel types are an issue of organization of the registry and not access. So standards track is appropriate.
[12:16:30] Pete Resnick_web_827 joins the room
[12:19:34] <Pete Resnick_web_827> Standards Track gets Expert Review too doesn't it?
[12:20:00] Behcet Sarikaya_web_158 joins the room
[12:20:58] <Darrel Miller_web_117> Having a document to register new Top-Level types would provide an opportunity for further description of the top level type.  For example, "example/*" says "The example media type is used for examples".   This feels like an opportunity for further elaboration.
[12:21:46] BEHCET SARIKAYA joins the room
[12:22:01] <Francesca Palombini_web_226> don't think so Pete
[12:22:34] <Francesca Palombini_web_226> there is not even experts assigned for Standard Action policy (ex:
[12:22:57] Ali Begen_web_898 joins the room
[12:23:36] <Francesca Palombini_web_226> specification required requires DE review
[12:25:15] <Darrel Miller_web_117> A better example is "model/*".  It has no description of its purpose. How would I discover what it means to register within that top-level?
[12:26:06] <Phillip Hallam-Baker_web_494> @jfk just made me think of something that might be dispositive here: The expert assignments might well be different for different toplevel types. That is an administration function IESG/AD needs to have control over.
[12:26:15] <Pete Resnick_web_827> Ah, "specification required" not "standards track". Right. Thanks FP.
[12:26:16] <Murray Kucherawy_web_407> @Pete: We could certainly say this is Standards Track, and the IESG is strongly encouraged to ensure media type reviewer input.
[12:28:07] <Pete Resnick_web_827> @Murray: :thumbsup:
[12:29:08] <Harald Alvestrand_web_249> note - we're running approx 10 min behind schedule. keep brief.
[12:30:33] <Murray Kucherawy_web_407> The draft registering "haptics" as a top-level type can be done in parallel with the one describing top-level types in general.  We could take them both together, the former being a first example of the latter.
[12:30:48] <Murray Kucherawy_web_407> (just a suggestion)
[12:31:40] <Harald Alvestrand_web_249> yes, that's what we are doing.
[12:31:49] <Murray Kucherawy_web_407> (y)
[12:31:56] <Harald Alvestrand_web_249> haptics can't go *before* toplevel, but I hope we can send them to IESG together.
[12:32:11] <John C Klensin> @Murray: That is sort of what I'm feeling for, but I've been thinking about it as closer to "Specification Required but with IETF Last Call as input to the Expert Reviews"  For this purpose, they should consider themselves a team and reach a conclusion jointly and only after input from Last Call".
[12:32:12] <Murray Kucherawy_web_407> :thumbsup:
[12:32:53] Ching-Heng Ku_web_589 joins the room
[12:32:58] <John C Klensin> Actually, the controversy around "haptics" that drove this WG is a perfect example of the problem ... and the need for what was described as "wanting cover".
[12:32:58] <Murray Kucherawy_web_407> @John: Can we issue a last call on something that isn't a document?  Never tried that.
[12:33:08] <Murray Kucherawy_web_407> I guess it would be a generic IESG Action?
[12:35:28] <Francesca Palombini_web_226> on a less important note, I think that whatever the WG decides on as for what the correct process is for registration, that should be described not only in the document but also summarized in an IANA Note under "Registration Procedures" because otherwise it could be missed.
[12:35:56] <Murray Kucherawy_web_407> "Specification Required but with IETF Last Call" to me is almost indistinguishable from Standards Track.  If the specification is external but still needs a Last Call, an AD may as well sponsor it as an I-D.
[12:37:13] <John C Klensin> "Specification required" does require a document, so not too precedent-setting.    In some ways, the only difference from Standards Track is that the final decision would be made by the type reviewer team.  And, if we wanted to make as little apparent change as possible, almost exactly the same thing would occur if we stuck with Standards Track but required not only Expert Team input to the IESG but that the IESG pay attention to it.
[12:37:36] <Murray Kucherawy_web_407> I think the latter is easier to describe, yes.
[12:43:41] Adrian Sterca_web_505 joins the room
[12:46:24] <Darrel Miller_web_117> Doesn't the Content-Encoding header handle gzip?
[12:46:58] Adrian Sterca_web_505 leaves the room
[12:47:53] <John C Klensin> @Darrel:  Yep.   And that ... well, I guess I need to get in the queue when the speaker stops.
[12:49:44] <Murray Kucherawy_web_407> I agree that trying to get Ned to review and comment would be quite valuable.
[12:50:08] John Preuß Mattsson_web_940 leaves the room
[12:50:21] <Murray Kucherawy_web_407> If you're worried about the security aspects in particular, the chair can request an early SECDIR review.
[12:53:29] <Manu Sporny_web_659> Ok, I can add a specific example for gzip bomb
[12:55:23] <Murray Kucherawy_web_407> @Manu: Happy to help review whatever.  I have to review whatever you do anyway. :)
[12:56:06] <Manu Sporny_web_659> Yes, the "combination of suffixes" has already come up... +json vs. +ld+json (which one can you pick) -- the spec says something about this... but perhaps we should be more clear there.
[12:56:33] <Manu Sporny_web_659> re: reviews-- thanks Murray, will definitely take you up on the offer! :)
[12:56:35] <Darrel Miller_web_117> @manu is the primary goal to allow user agents to process the response as the more primitive media type, or is to allow distinguishing between two different base formats for the same media type semantics?
[12:57:18] <Manu Sporny_web_659> @Darrel - at least that first one... the second one, I don't quite understand?
[12:57:37] <John C Klensin> @Manu: Yes, if the spec can dig us even partly out of the mess, it would be A Good Thing.
[12:57:48] <Darrel Miller_web_117> e.g. application/did+ld+json and application/did+ld+yaml
[12:57:59] <Manu Sporny_web_659> Yes, excellent question... text/plain+gzip --> is that validated after this draft (I think, probably not? but great question)
[12:58:11] <John C Klensin> @Harald,  how about text/plain+UTF16, which is another part of this path, unless carefully excluded.
[12:58:34] <Manu Sporny_web_659> @John -- good to know, would appreciate guidance if you feel the current spec is digging a bigger hole (which we don't want)
[12:59:12] <Murray Kucherawy_web_407> I agree, Martin's document updates 6838.
[13:00:16] <Manu Sporny_web_659> @Darrel -- re: application/did+ld+json and application/did+ld+yaml -- yes, we do want to be able to distinguish between two different base formats for the same media type semantics (if I truly understand your question at depth)
[13:00:28] <Darrel Miller_web_117> I have been largely supportive of suffixes in the past, but the use of +gzip is very concerning. How that interops with Content-Encoding: gzip seems like a source of future pain.
[13:00:53] <Murray Kucherawy_web_407> We could also amend the form to just say "If you don't know what this is, leave it blank."
[13:00:59] <Manu Sporny_web_659> @Darrel -- yes, agree, there is a non-trivial possibility of creating a problem here.
[13:01:16] <John C Klensin> As that other co-author of 6838 (and nothing that Tony isn't here either), "gone shortly after OS10" is consistent with what I know, but Ned should definitely be involved in any further discussons of this.
[13:01:50] Darrel Miller_web_117 leaves the room
[13:01:51] Francesca Palombini_web_226 leaves the room
[13:01:52] Pete Resnick_web_827 leaves the room
[13:01:53] Alexey Melnikov_web_818 leaves the room
[13:01:54] Murray Kucherawy_web_407 leaves the room
[13:01:55] Francesca Palombini_web_420 joins the room
[13:02:13] Manu Sporny_web_659 leaves the room
[13:02:14] Cullen Jennings_web_726 leaves the room
[13:02:20] WU Hsing-fang_web_231 leaves the room
[13:02:20] Yeshwant Muthusamy_web_506 leaves the room
[13:02:26] Behcet Sarikaya_web_158 leaves the room
[13:02:29] Renan Krishna_web_754 leaves the room
[13:02:30] Harald Alvestrand_web_249 leaves the room
[13:02:30] Chris Ullrich_web_149 leaves the room
[13:02:30] Phillip Hallam-Baker_web_494 leaves the room
[13:02:30] John Klensin_web_383 leaves the room
[13:02:30] Takahiro Nemoto_web_227 leaves the room
[13:02:30] Ali Begen_web_898 leaves the room
[13:02:30] Ching-Heng Ku_web_589 leaves the room
[13:02:30] Martin Dürst_web_136 leaves the room
[13:02:30] Francesca Palombini_web_420 leaves the room
[13:02:30] Robert Stepanek_web_378 leaves the room
[13:07:19] Meetecho leaves the room
[13:10:41] BEHCET SARIKAYA leaves the room
[13:28:40] John C Klensin leaves the room