Internet Engineering Task Force Interim Meetings, 12 Feb 2020

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: CBOR WG Interim Meeting, 2020-02-12

Description

CBOR WG Interim Meeting, 2020-02-12

A

Yes, so, let's start the meeting, I have and the etherpad those that again in the chat.

A

um From last interim, we can look at I know, I need those to remember. What's.

B

A

Yes, so maybe we can start with the mail I sent today, so you might have seen Caston I'm I reviewed the pull request. It's all fine I had a couple come like a couple of responses to your response, but that's you can go ahead and write that yeah.

C

So what my plan is to go through your response and turn anything that I think still needs to be done into another issue and merge the pull request that we have, unless there is that's very quick to fix.

A

Two points I think one is about the you know the apps area working group contact and you were wondering if we should just change to iesg and what would be the contact then so either I I think is. Geez is good as a change controller and the contact could either be the Seaboard working group or the art at ATF. Org.

C

So so that's actually what I am planning to put in there see what that ITF, not all auto died after dog. That.

A

Sounds good because I think will work anyway.

C

I'm not sure that that arch dot ITF that always going to be a very long, lasting name but I think again.

A

I register wrote art, IDs, org or whatever.

A

What did they say like whatever follow-up Bora I? Don't know? What did you call it derivative of it like that, but that doesn't really.

C

When that gets replaced, there will be some forwarding, so I think I.

A

Think there was still like.

A

When I checked today, there was still.

A

Let's see not that important, but yeah there is still like plus Jason. The contact is still up. There. Working group was discussed at the 80s, the Doric.

C

Really, okay, but absence is forwarding to art right.

C

I I have made the comment repeatedly that we need to maintain some some best practices for media type registrations, because each time I'm right there in the draft with a media type reservation in it, I'm running into issues of this kind again and again and three. What the right ways of doing things. Another.

A

Thing I could do, is ask our beloved air director and see what LexA says.

C

Let me make the change and then we can ask alexei whether that's the right.

A

Yeah and what I was thinking about east Jason text sequence that says application, real-time error, discussion, art at IDF, dot, org or any is G designated successor, I mean.

B

A

Is going there and looking for a contact point and I see that in the Seaboard sequence you put in the people working remaining list, so you can just do that. I! Think! That's! That's! Fair! Okay!.

C

A

The other was very minor, but I mean all of these are minors. All these are editorials.

A

The other one was where I was asking about an example of what quote very small amount of code could be, and all I meant is that you talk about, for example, in class, one constraint notes as defined in 72 28, and it's just glass one constraint nodes with around a hundred kilobytes code size as defining, so that those are the numbers that you find in in 72 to 98, but just I had to go in and and look for it.

A

It's just like to remove sir or when we say about it, like very smaller code, is so general that it could be when you have an example. Next to it, it looks much better yeah.

A

C

Reason, I'm a little bit shy of doing this kind of thing is that, of course, the specific value? What is small in this space depends a bit on what you are trying to do. So, if you facing.

B

C

A hundred kilobytes is a small amount. If you are not interfacing, you just need to process SIBO in some form, then maybe 100 kilobytes, maybe a lot.

A

This was just like for the example anyway and, for example, parentheses, but I'm fine in a way keep it or leave it like I'm, good.

C

Wanted to point out why why I'm hesitant there's generally, it helps not having to follow links, but sometimes it's also very confusing to extract one specific aspect from a link to a document. That really is is opening a much larger can of worms and so I'm, always very hesitant of one document, essentially trying to explain what what another document is doing, because we have had a badge in the normative space, not not in the example space, but in a normative space.

C

We have had bad examples of one document trying to explain what another document is doing and remaining that explanation sufficiently, that there were actually in turbidity problems with that, so that that's always a bit hesitant.

A

Yeah this was the other one. Other comment that I said: I had something else to say, but yeah I'm, fine, you can go ahead and merge it and and yeah close that.

A

Okay, did you hear from Jeffrey no, no okay, five I think.

C

I congratulated him to the W pick. It's like this.

C

Which may be keeping him busy at this one.

A

To get he's okay before before emerging that, even if it takes a bit longer.

C

Yeah, so my my intention was to look at the other fuse that that have floating-point aspects to them and work them into this pull request so I'm, not that unhappy that he hasn't reviewed it yet because the pull request may actually evolve a little bit.

A

You were planning on starting with floating floating point things, that's not done yet. You said.

C

Right there are other floating point, related issues out there and I think the text addressing those can go into the same for requests, because this really needs to be aligned with the changes that already I 165.

A

Yes, I can see there is these two new issues or not sorry, one new issue since last meeting.

A

Which one are the issues that you're talking about now? The 150.

A

C

Yes, one-sixty is one of those and 153 is the other one. Okay,.

A

C

The one per request I haven't done the wooden address, 156, seventh and 63.

A

A

Sorry, the one request you haven't done: yes,.

C

Then I promise to the last homer and haven't done yet.

A

A

What numbers did you say? 163 the.

C

156 inch 63.

C

Three at 63 I'm, sorry.

A

Okay, that I had already put it: okay,.

C

I'm sorry 156 63.

A

Fine, it's fine. We already noted so that's.

C

I'm not looking at the at what you're typing here. There should be.

C

A

Right, this is okay, because on the.

A

Minutes from last time, you were planning on starting with the floating-point stuff first and then, including the fix me remember what this fix me was, but I.

C

Think there was still affects me in the takes so well.

A

Updates for this meeting, basically.

A

It's it's fine I'm, just the check.

C

Getting dangerously close to the rough deadline so.

A

Remember it's good that I bring up two minutes from last meeting, because I also forget a lot in in two weeks, but we were thinking. We were saying that we're hoping to start. The working group last call the 19th so next week and we should decide today if run it for one or two weeks, because if we run it for one week, then you might have time to do one more update after the working of the school ends before summation cutoff.

A

Otherwise it's um like it ends like he gives you a couple of days before summation cutoff and you probably won't be able to include yeah whatever comments you get.

A

C

Not too ambitious, so I think we should go for two week. Working last call and you should do it next week. Okay,.

A

So working, let's call, let's say two weeks and then for starting day, I guess when whenever we have the next submission like including this poor request for map validity, including my my review and hopefully Jeffrey. If we can get him to approve of the 165.

A

And yeah Laurence, you also had the one in people with you: I.

D

Have not that was.

D

138 once it's yes exactly, and that was close issue, 138 I'll.

E

Do that in the next area to I, have it open and I'll? Do it? The next day at you, I, think.

A

Like restore your heart, I think since so we're still waiting on on Jeffrey and and Kirsten has these updates. But when you have time, then Garson knows that he can go ahead and merge that as well. Okay,.

A

A

You're under 59.

A

Notify Carsten and mainly list so I also see that because I I will have to good shepherding of this. So good I see.

C

Yeah so 159 was already merged. Okay,.

B

C

See would love to hear from Lawrence whether we did the right thing.

A

Okay, I'm reviewing the texts that we have discussed. Okay, so working group last call.

A

Their next submission or yeah now now it's hard to be a date right, because it depends on new custom when, yes,.

C

The start of next week, so the 19th to commerce. Yes, yes,.

A

Yes, I understand and then we don't have a yeah. We don't have a hurry to be before so we should cut off. So we can give it two weeks.

A

Okay, yes 153 and 60, which are those that you wanted to include in Jefferies PR before reaching out the implementers.

A

That's still in the plan, oh I guess.

A

So I guess you want to wait for next submission before reaching out so should we reach out to implement this at the same time as working group last goal, or do you want to do that before well,.

C

Probably we can do this a little bit before, but I think it would be good. Have that that many remaining pull requests that are unmerged before we talk to them.

A

C

So, of course, some of them like this.

C

Edwards guy- and we can talk to earlier because that's their issue, obviously that that we are working on the. Why not reach out.

C

Edge are slightly ahead of time of the.

B

A

A

A

And I think that's.

A

This is basically, this is a little AP list to go through to make sure that everything is it's moving along.

A

Let's see so yeah these nine issues, anything we need to discuss the new issue. 167.

A

C

Yeah so I think the the interesting question behind this issue is when you have an initial initialized or undefined value for a time.

C

Is it actually helpful to know that this is the time or is it sufficient to use something like null or undefined as a general mechanism? So people are sometimes so stuck in the the type system they are using in the implementation language that they don't see the obvious way of offending something by simply using only find another yeah.

C

C

Haven't that this into today, but that was really the issue that I took from this. Yes,.

E

So you're designing a protocol and somewhere in there protocol is a field. That is a time and you want to be able to say I, don't know what the time is.

C

A

Is time but the time yeah one.

E

Option is to not put the data item in, but that doesn't work in some scenario that works in maps and like in you know, CWT claims, but it won't work on a lot of other protocols. So the question is how you specify I mean I, think we we had some discussion about this and kargh I think about unspecified or unknown like how do you, particularly like it with expiration? How do you say no exploration.

E

I'm not remembering very clearly, but what were we ended up with that? But but in.

C

Extra remainder, the those elements are mandatory. People have made up a convention of using the nine nine nine nine to indicate no expiration I.

E

C

That's not the outcome we want to.

C

So what? What, when can we add words notices? Is that not a numbers I fear loud there, because that's a floating forward value.

E

C

That's one way of doing this: I think that's pretty weird, so normally I think just putting another in of what would have been a time is the right way to do this, and if you have a protocol, that'll not putting your time somewhere, but still because you put two bits on there. Something like if you have an array and the third element is the time, then you have to put something there then use now or maybe even undefined. If, if that's really what you, what you're trying to do.

E

You know that seems seems pretty close. What I would ever choose.

E

The net that convention could be actually meant a pretty broad for all kinds of things. You know like integer values and other other things too. Other types of fields.

E

Can you write that in CD DL.

C

But something like time or not, yes, you can say that yeah.

C

Yeah, maybe I should write a response for this issue first and see what Montgomery it was has to say.

E

C

Formally, as a working group, we don't really accept contributions from people who don't identify themselves, but I think we can ignore that here.

B

A

I think what you say makes sense: Kirsten.

A

Okay and then the other issues are, we discussed. I think there is no really.

A

Nothing that I can see stuff.

C

Changes is this a piece of work that needs to be done? Yes,.

A

I think you mind registration text once more I like like the title.

A

Rid of the artwork yeah.

A

Yeah, that was one of those glowing Jeffress.

A

This 53 yeah, we converted all of those so I, think just needs to done and, and there is a typo that was noticed, great yeah.

C

I think it's even fixed in in either you or Jeffries or request, but I'm going to merge that that per request, because it's just nice that somebody noticed yes.

A

Yes, great so I think that that's it done. It sounds like a plan yeah just.

C

New tricks, if you don't yes,.

A

Exactly I, don't I don't have a PS on me, but you have a couple and Lauren's and hopefully Jeffrey as well as can take some time to do the PR and then we can move on from that and then my AP will be to call the working group last call. So let me let me know when you think the the version is I mean I will see it. I.

B

C

Submit a new internet draft and and and say this is the one.

C

A

Know it might happen a new submit like two or three version at the same time, so.

C

The minutes you maybe should mention that we quickly talked about the si o te ika phone next week and that's interesting, CDL and see what what is going on there so just think it's racial know that they talked about that.

C

A

Can say that again, so it gets into the recording as well. Yes,.

C

So next week that there is a joint hackathon of the three working groups, rats, cheap and suit and in particular suit, but about lots of reds- is using C bar and and CD DL, so that there is some related implementation work going on, in particular with respect to generate code from CD DL.

C

So it will be interesting to see what specific issues that come up the end. We can feed them back into the.

A

E

I'm gonna drop off here I got a good.

C

So I think we don't yes.

A

We're done very, very short, I mean not super short actually but shorter than usual, and yes, I talk to you in two weeks. Then.

C

Did we get those some interim meeting requests out this? Is they actually once canoed.

A

C

I would have to find them in my mail.

A

So we have three just to everybody, so we have one today and one in two weeks, the twenty-sixth of separate and then the last one is the 11th of March. Yes,.

C

A

Might actually be problem for me because I have something at work the whole day and I'm not sure that I can miss it. But let's see we'll talk about it in next interim and see.

B

A

Thanks Kirsten I, never.

C

Nice thing bye-bye.