Internet Engineering Task Force Interim Meetings, 26 Jan 2022

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: IETF-CBOR-20220126-1500

Description

CBOR meeting session at IETF
2022/01/26 1500

https://datatracker.ietf.org/meeting//proceedings/

A

They are like 12 seconds old, so you don't need to apologize, but I think we we quickly should discuss the intended status of the document, because right now it's marked as a bcp, and I find it weird for bcp to actually define what is essentially a new.

B

Protocol all right, so uh we should maybe not uh which I guess that should go on the agenda rather than just please a moment.

A

So we have a working group meeting where everybody has switched on the microphone. That's uh no christian doesn't no. He does.

C

Yeah and so and and john just came by so but yes.

D

And I just find it. I like.

C

The way how this is getting started to just flow along anyway.

D

No, I just find that leaving the microphone enabled and using the local mute works much better, because there's no delay when I unmute that way.

B

Yeah, that's a really good plan. Actually should do that. um It's so annoying that delay.

D

Whether you have a proper local mute well, no there's in the lower right corner of the meet echo screen. There's a there's, a mute that that mutes it it mutes your audio. Without disconnecting your audio stream.

A

Oh wow, I didn't know that that's a great.

D

Feature yes, and that that that makes it so much nicer, but so anybody who's talking frequently can just leave the microphone enabled on the left and mute on the right. Yep.

A

But it makes it harder for the chairs to recognize when somebody has forgotten anywhere themselves exactly.

D

So but yeah I mean for people like, I said, for people who expect to pipe up a lot. It's a it's a big help, so it looks like yeah. We got. Sean turner came on we're still just the four of us and.

C

Sean yeah, I'm not sure, we'll be too many more francesca mentioned that she's. Another meeting um marco often drops by.

D

Yeah marco is usually on, and ian did not send regrets this time, so I was expecting him, but you know we don't really have an agenda this time so.

A

Yeah there also wasn't a reminder: mail.

D

A

That's my rely on those.

D

I know I I meant to send one and I and I neglected.

D

This wintering in florida stuff is throwing my whole schedule off.

D

Speaking of which michael how's, the weather up your way.

B

Well, it's very canadian of you to ask about the weather. um You know what else do we talk about here.

E

It is apparently.

B

Minus 42 today this morning with the wind chill- oh my god, uh not it's in the negative, 30s plus a wind, chill. uh I very happy to work in the basement under the heating vent. um But my family has gone out to uh work in school and uh we'll see if they ever return or if I have to dig them out of a snow bank in april or something so.

E

Meanwhile, in washington dc, where I am, there might be a couple of inches of snow and we're preparing for the worst yeah well,.

B

You know it's, okay, it's it's! Okay, that's the right way! It's okay! uh The president from toronto is it is you can call the army when you get an inch of snow, dig people out, yeah, yeah.

F

So people always.

B

Making fun of toronto for that, but the truth is that they that you know they'll get like you know two inches of snow in the morning and it melts by the late afternoon, and so they really don't actually own much snow, removing equipment.

B

So that's why that happens, whereas in ottawa it actually snows every year, uh despite what the city staff think um and we actually own a significant amount of snow removing equipment. So it's a difference of like preparedness right.

B

All right with the with the weather discussion of part of the meeting over, we should put that on the agenda properly.

D

Yeah there you go so I think we're uh it's we're five minutes in. So let's go ahead and talk about whatever we can talk about.

E

D

uh Move on from there.

C

The file magic intended status was one of the the open points from the original. From from the earlier chat.

A

My christian, the pr is not recent. Oh.

A

The comments are recent.

B

Yeah, I'm looking at your comments, um karsten and I'm I'm not actually sure what to do about them at this state. I guess I need to let them sink in a little bit. um Are we talking about this? Now?

B

Yes, yes, go ahead, yep! Okay, so um I think let's the the big, the big. The big point is, I think that carson thinks this should not be a bcp and, as the document started to me, it's now a year ago that the reason why it was a bcp is that uh it was a practice and it was not in any way. You could do anything you like at all, but here's a suggestion on one way of doing it.

B

um Nobody needs a nobody following this needs to uh do anything specific, because the ayana actions that you need to follow are all uh first come first serve essentially, which you don't need to do anything about, um and so that's why I was like you know what this is just. This is what I did if it works for you, then you should do the same thing too.

B

For the same reasons, and that's why um it you know, you might have reasons to do something else: it's not a standard that anyone should uh put in an rfp or something like that.

B

um But we've now evolved a little bit and uh we say a lot more things, and so, if someone felt that it needed a bigger amount of.

B

Acknowledgement or whatever I don't have any problem with that, but it wasn't something I would have pushed for a year ago.

A

Yeah, I think the fact that we're using numbers that happen to be in the fcfs space is not really deciding this issue. I mean it might be an additional piece of information one would use, but um I think the question really is: what is the level of normative intent here and, um of course, the the? The idea is that if you use those tags then they mean what is written in this document. That's why there's a pointer from the registry to to this document.

C

Yeah, originally that document wouldn't have registered any text would just have phrased how you given guidance on how you registered text.

C

Thing is even even if the um even if the, um if the tags are registered, um doesn't mean that this needs just because the tags are registered, doesn't have any impact on the documents status either, because it's first come first serve. It could be a scan, cheat sheet of paper that someone scribbled notes on that would suffice as a reference um so yeah.

C

It's probably not not a push for the documents. That's another way. Another direction.

A

Yeah, I think the the main decider for me uh recently has become. Do we expect to have this as a normative reference in a different document, and if that is the case, then I'd rather have the the document being normative in the first place, instead of doing the downref procedure afterwards. So this is really the interesting question. Do we expect to use this as a normative reference in in some other document?.

B

Right, I think that's the major thing is: there's someone going to going to reference it.

F

Oh, this is ira. Sorry for cutting in um what document are you referring to? I'm sorry for being late.

B

The file magic document.

A

And by the way, eric you're, very quiet.

F

That was because my microphone was nowhere is near my mouth. Sorry about that much better! Thank you. I I would suggest that file magic might well get normative references.

A

Yes, I think that that leaves bcp and uh senate strike and bcp is kind of weird I mean we can. We can just do that and and see whether the isg has a different opinion, um but uh yeah really from from the systematics.

A

um I think this. This is a standard strike document as as boring as it may be. In the end, it's still something that we expect to be specified in other places.

D

Yeah, I certainly don't think that.

D

Yeah I agree, standards track is the right answer.

B

Standards track is the right answer, just repeating what I think you said.

B

Yes, okay, I won't argue with that.

B

um So let me bring up another uh comment that carson had, um which is relating to. I said the I said in my rewrite of the introduction. I said.

B

The third method is also proposed by which seabor format prepended tag is used to identify non-sebor files um and carson, says, there's actually a seaport tag for non-seaport data, and I'm just wondering if uh actually what I would like to do is is have some more comments as to how to explain that best and succinctly, um and I think just someone else looking over the making another suggestion would be great.

A

Yeah I was trying to make a suggestion and then someone couldn't remember how to do this. Oh.

B

There's a you, you there's a plus button, the plus in the front, becomes a button.

D

B

Suggestion and then you click, the plus minus insert a suggestion.

B

Would you like me to share with my screen no.

A

um I figured it out. I just had very little time and wanted to to put the stuff in before I had to go to somewhere else. uh So I I do that.

A

um Essentially the the I was just reacting to the sentence saying it is not really about sibo and well yeah. No, actually, it is about zebra yeah yeah. I get it so there are several comments here for me that that were really just about the the sentences on their own saying something which wasn't right, except when you actually add the context and and in a standard document you're trying to make sure that that there is no single sentence, that people read and say that's not true.

A

So that's why I reacted to you at one point: files were all record orientated and I definitely opened record-oriented files with a block-oriented access mode in the 70s that was standard procedure. So it was not the operating system.

B

You tell me I'll. All I was trying to say is that at some point this was not a problem, because the os knew everything about the files and then we went to stream of bytes.

A

Yeah and that's actually not true, but okay,.

B

Well, you know what I I was four years old at the time, so uh I will. uh I will.

D

Defer to someone with a longer memory.

B

um But but the point is, I actually think it was. I think that I think that tcpip could not have evolved without us going to a stream of bites thing, and so it's not like.

D

B

That this is a terrible thing that we went that way and lost the tag, um and you know so. That's the that's the point we could. I could delete the sentence if, if you feel it's, uh if, if it's impossible to make it accurate, it's just uh trying to ease people into how we got here right and I'm not trying to make a yeah.

A

I think the the point.

B

A

It used to be a situation where the operating system manages for you and unix was kind of radical. I remember when, when I learned this, I was kind of surprised.

A

uh It was kind of radical in not having uh file types, and so you you didn't know with which access method, which really was the the thing you were looking at in the 70s. uh You would need to open this file, and now the the byte stream representation of text is kind of orthogonal to the question whether you actually have uh file type information, so mac os has had file type information for a long time before they gave it up to be like unix and um so yeah again.

A

This is, I think, it's easier to to not uh try to make this point based on the somewhat limited uh review of history.

B

Good, uh so you think I should just delete the uh delete the sentence, probably it's better.

A

That will probably help. Yes, I think we can find a sentence which actually correctly says what you're trying to say, uh but I don't think we have to invest the effort.

A

And I actually agree it was an interesting conference of of tcp getting mainstream in 83 and and unix getting mainstream as well, um that really anchored the the byte stream or our network tiller type.

B

A

B

A

Is that we don't have it.

B

We don't have a a sub type of tcp. That tells you what uh uh what is the record format right? No, but ftp is that yeah, but only for telling you it's binary or something yes, anyway, all right, so I'm just going to delete those two sentences.

A

I think the the sense sentence is okay,.

B

The second sentence is okay, but it starts with the.

A

So it has, since very early, is okay, the the one about all fights were record oriented, that's not okay and then the next one that changed.

B

B

It would be better if I shared my screen. Yeah.

A

So I was not thinking we were turning this into an editing meeting here.

A

Christian, I recommend the the intro destiny uh internet history, mailing list. uh They have some really interesting views into the 70s history. I I I wasn't on the internet in the 70s, so I only got there a couple of years later so.

C

On the context of turning this into an editing meeting and sorry for earlier, when I typed and didn't notice that my global mute button went to another device, maybe let's cut this short here if we, if, if there's still drive to use the remainder of the hour later to progress to make progress here, we can still do that if, um once we've covered everything else, because there's one more item that I'd like um get done on the agenda um before this goes into a collaborative editing meeting which it can be, but at which point some people might want to leave already.

C

uh That is. Are there other things? You definitely want to have on the ipf 113 agenda, um and if so, please let us know or just write them down here, uh say something now so that we can plan accordingly.

A

Yeah, we still have one technical item on on the five magic document. The sea ball is a cbox uh thing. We we could discuss. Oh.

C

Yes, we sorry, we, I mixed up the items, and I thought we were all about that already and yeah anyway. um Yes, right, let's continue there and then.

C

Yeah doing the notes here.

A

Yeah, so I made an argument with which is more on the matter of principles of protocol design and and less absolutely. This cannot be done this way. I I die on this hill, otherwise uh point uh so we last at the last meeting. We said we're going to discuss it on the mailing list and we didn't uh so. I think we at some point. We should close this, because it's not that consequential.

A

So we just had this this nice discussion about this. The fact that the new rfc xml format has two ways where you put the document name and, of course there are hundreds of documents out there, where those two places don't agree and different tools have different at different times, looked at diff at different places for this information, so it creates a ton of of confusion and that's exactly the the same principle of protocol design that was violated there. You really want to normalize your information in such a way that it only occurs in one place.

A

So you you don't get different implementations that look at different places and react differently.

B

I agree with you, but what was the context that you're specifically talking about, in which doctor does the sibo versus cbox thing? Oh, yes, okay! That part all right, um so I I agreed with you that, if uh not a hill, I would willing to fall on die on. If you want to remain, keep it as cybor, then I'm happy to keep it as cbor. I feel that it's a little bit weird, but I don't.

A

Mind: okay, any other opinions on this.

F

This is ira, I withdraw my suggestion of a c box or something though I actually would have preferred it, but um I understand your rationale for why we should just say: sibor.

A

Yeah, I'm feeling bad because it really was a good suggestion. I like the term cbox, but it has a problem.

F

Everything has a problem in computing.

F

Okay, good we've got, we've got consensus michael, you have your marching orders.

A

Okay, so the authors will will see to it that we get a a new draft that that has these uh comments in it, I'm only going to get to it to next week.

A

C

So on the topic of um itf 1413 agenda, because I don't think we have any other items left for today, um michael just suggested chat, whether whether we actually need a full meeting so yeah that will depend on on what we want to do in vienna and how much uh how many things we hope to make progress on.

C

So yes, that's one reason to to collect that list. So kind of last plan was to just ask for one meeting uh slot again, because uh experience says that we make good use of it. But um yeah. Let's see what comes in in terms of agenda and if that's far less than expected, we can still reconsider that.

B

Based on the state of our drafts and things like this, I don't think we have much more than a status update, and I think that we could easily do that with another group, and I would rather, we didn't ask for a slot at all. To be honest,.

B

Rather than waiting.

A

Yeah, I think we we need to coordinate with the date at some point on on the time tag document, but I don't know when that point will be and whether it will happen to be at the end of march.

A

So I'm not going to suggest taking this that as an agenda item and of course we have to work on sibo packed, but I think both things are really easier done in in terms.

C

How about the the statues with cddl2? Is that something we need time to talk about.

A

We could again that depends a lot on on whether we we expect spectators to come in people who have committed their their week to the itf and and come to related meetings.

A

um If we have a reasonable number of those, it makes a lot of sense to to discuss this at the itf, uh but that's hard to predict.

B

I would say if we were having a thousand-person in-person meeting, that we would get those tourists, um um but I think that, given the remote that many people will be remote again, they could just as easily join our virtual interims, and um I think that the conflicts will mean that we won't get people anyway um to to.

B

I think that I think that a presentation about cddl2 would be useful, but would be useful in a wider audience, even as an ieb tech uh talk um and I'd rather see us do that than have a meeting about it.

A

Well, a meeting might be a presentation plus a discussion.

A

Yeah, we also could do an interim and just make a wider announcement. So people know at that particular meeting. We are going to discuss city day too and I'm sure the the rats, people and and the suit people and so on, uh will will come and discuss that. So that may be another way to draw in additional people.

C

From everything gathered gathered here, I take it the cdtl2 can. If, if we were to have a meeting, then cddl2 would be the main attraction and probably also the main point of it.

C

C

Are these these tech talk enough talks, announced and scheduled? They are probably they probably have need scheduling in ahead of the of the meetings right, I have been involved yet.

A

I think the the whole issue of reviving the tech talks after the lull we have had, I mean they stopped taking place before the pandemic, so the benedict was not the reason why they they stopped, but they the the pandemic, may be. The reason why they haven't been uh started again, because nobody knows how to do a a proper plenary in in the current hybrid uh situation.

D

So carsten are you, which are you talking about the deep dives or the um technical talk at the regular plenary meeting? I'm talking about the plenary meeting?

D

Okay, that that was always set up by the iab and the iab picked a subject that they think had brought enough uh interest across the community uh that that they wanted uh to put that in their portion of the plenary and they're.

B

Always looking for for for stuff right, we have had them during the the during our online meetings. We have had a couple of those uh where we haven't it's, because there hasn't been a topic they found or speaking right. So.

D

If you're, if you're looking for that, then talk the the talking to the iab is the right answer. If you're looking for the the other, the deep dive kind of talks that was set up by the iesg.

D

uh But that's in a separate session from the plenary.

A

So if I were the iab, I actually would be thinking about having a more general topic of data modeling, because we have more data. Modeling technologies than just cddl in the iitf and in particular yang has made some significant progress in in the last two or three years, and it's probably worth presenting that alongside uh what we've been doing in cddl and maybe an sdf as well so.

D

A

A really a worthy plenary topic.

C

Like like a shared shared presentation of of yang, cddl and and whatsoever, not um illustrating.

D

C

The state of the art and sdf illustrating what the state of the art in terms of modeling is.

A

Well, that that would be a dream, but uh I think more realistically, we we would have separate, represent presentations from the different camps that are coordinated a little bit.

D

So um then, then I think uh carsten you should approach the iab about that.

D

You can just send a note to iab iab.org.

D

A

D

Could approach a particular particular iab member who you thought would be interested in taking it forward either way.

A

Yeah, that's certainly not something that will happen at 113.

D

A

Yeah, so it would be the the new iab we would be looking at.

D

I still think um that, uh probably if you were gonna go with an individual member, maybe tommy, pauly or david skenazi would be the ones to approach or, as I say, just approach the iab as a whole.

A

Yeah, maybe I should send the heads up to to a couple of individuals in the iab so when, when we have a slightly more well-formed proposal,.

A

And bring it to the iab, there will be some people who are seated.

D

Yeah, that's a good plan.

A

D

And so what's what's uh our decision on 113 are we? I I've already made the request for a session which I can withdraw.

A

Yeah, my summary of the discussion would be withdrawn.

D

Yeah, that's what I'm thinking uh a christian. Do you read it the same? Yes, okay. I will withdraw the request. Yeah.

A

Well, one of the problems with the current setup is that you don't really get to supply a reason neither with won't meet nor with withdrawing uh requests. So I think we we always have to make sure to send a follow-up message that that explains why we are doing this, because the message just on its own looks like we are shutting down the working group or something.

C

Yeah, that's that's! That's a good thing to write. I can I can do that. um I just need to check which uh who whom this should be addressed to, whether that's only sibo or one, that's more recipients of the withdrawal request.

D

All right I I will. I will change the request to indicate that we're not meeting and christian I'll leave it to you to send that note. Okay,.

A

Yeah, I would just reply to the message that goes to the zebra mailing list.

A

um Read the chat, uh actually, I thought my summary of the uh 112 separate plenary feedback was that it didn't work, uh but maybe I'm I'm just misremembering.

D

Well, it was not universal. There were certainly some people who didn't like it and some people who did so that's why I say I don't really know what the iasg has decided not being on the isg anymore. So.

D

A

D

If they've decided yet because scheduling hasn't been done yet so.

A

I think it was discussed on some mailing list. I just don't know where to uh where it was and how to find it.

D

Well, it was either on the main ietf list or on gen dispatch. I don't I don't remember which either but.

A

Yeah, when it really should be on many couches.

D

Well, it's who knows.

D

All we talk about on many couches right now. Is uh uh this andrei petrescu's uh concerns about um having anybody within 20 feet of him.

A

Yeah, I'm not following that mailing list.

B

Fine to comment on the basis that this is recorded.

B

D

D

Anyway, I have done the uh cancellation, so uh this should be enough. There should be a note gone to the seabor list. Yes got it: okay,.

A

D

Now, whatever else is.

C

I left if nobody has any other business to bring up. That would conclude this meeting.

A

B

A

This side, the ups a working group side meeting on top of the regular seaboard slot very mad at them.

A

So on february 9th there will be a meeting on what is essentially doing asset management in gang, and that is of course well kind of related to what we are doing.

A

But I don't think that that should make us do anything. It's just annoying.

D

Where is that announced on the.

A

Op say working group.

D

Permitting list this? Okay, because it's not listed in the um in the upcoming meetings.

A

It's a side meeting so yeah.

D

Yeah, we don't.

A

Have a good way to handle these.

D

D

All right, so I guess thanks everybody for coming.

D

And we will be adjourned. Yes, yes,.

C

Thank you see you in two weeks.

D

See you in two weeks and I'll try to remember to do a reminder next time.

A

Thank you, yeah.

D