Internet Engineering Task Force Interim Meetings, 30 Sep 2020

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: CBOR WG Interim Meeting, 2020-09-30

Description

CBOR WG Interim Meeting, 2020-09-30

A

So here here it is, it's recording so welcome to seabor working group interim meeting number 16.. uh This is an itf meeting. The itf note well applies, and this meeting is recorded and the presence is logged in the um kodi md link that I posted in the webex chat, and it is also in the agenda.

A

uh Okay. So for today we have a itf 109 update. So it's just from me real quick.

A

What's the status there and then run through the working group document status capstan, if you have some updates there and then any other business at the end of the meeting so for the itf 109, I requested a session as we decided last interim and it's going to be a 60 minutes session and I made sure to include those um those working groups or that we talked about last time that were to add.

A

But if there is anything else, please go ahead and and check that if you want me to add anything else, because I can still modify it until this friday, the um yeah the the session request so yeah, but otherwise that's that's done um a working group document status.

A

If you want to do you want to start with the oi tags document, carsten.

B

Yes, so two interns ago, we had a discussion about the the tag factoring or the the recursive application of tags to avoid takes to arrays and maps, and I wrote this up and I'm quite happy with what I wrote up.

B

So unless there is a pushback now, I would like to publish this as dasher 1 for for the working group draft.

A

Okay and it's uh in the github right now, it's.

B

A

Yeah, so I guess, if anybody has any comments has had time to see it otherwise go ahead. We get the feedback from the working.

A

A

Okay, uh then we had the new proposal dictionaries. It was the second topic.

B

B

I I didn't update the the pack draft for for quite a while now, because we hadn't quite come up with with an obvious direction to go here. My objective here is to keep nicely separated the the actual packing mechanics, uh which would uh be extended by a suffix packing. We currently only have shared items and prefix packing. So that's one part of the proposal and keep keep that separate from the question.

B

How do the tables that go into that uh actually get constructed? And I think that's where the dictionaries come in, because in the end the dictionaries have to form part of the table. That is then actually used uh for dereferencing the various um reference to shared items and and prefix and suffix uh packing.

B

So that is still my not my objective, uh but uh I haven't managed to come up with a clean way of doing this brandon's idea of uh essentially using the the end of what's in the pack document.

B

As the point where the dictionaries are then added looks pretty good, it does have the disadvantage that we don't get absolutely optimal uh compression of of the reference. So that's something where you maybe want to generate a few examples and um actually do numbers before we make a decision, so that's the status there. I think we just need a few proposals that go a little bit beyond just the ideas and are fleshed out enough, that we can derive these numbers and, of course we. We also need a few examples.

B

uh Example, data items that we actually put into this packing to find out how well that works out.

A

Sounds good, thank you for the update um yeah. I think we can move on to the next next uh topic, which is 70 49 49.

B

Bits yeah the status is that we had a couple of.

B

Editorial improvements we generated dash 15 based on on the comments and discusses, and then we got a little bit more feedback, and there is one final pull request: number two 22.

B

I think which I would like paul to actually cross check, then, when he merges that we can submit dash 16 and that should be a basis for barry to decide whether he can press the button.

A

A

That's good great, um nothing that we need to discuss then, in this uh request.

B

No, unless anything comes up, I mean there's always something that may come up at any point in time and better now than as errata.

A

A

Course so maybe, let's give a moment uh to open this pull request, and let me just copy paste the link in a minute, since we have the time.

B

It's somewhere in the large number of types I have opened.

A

A

B

So barry had a pretty nice restructuring of the first paragraph, okay and yeah I'm. I would assume that the rfc editor would have beat us up on that paragraph anyway, but I think it's better to have something now than to wait for the rfc editor to do it.

B

And, as far as I can see, that's the request.

A

A

This is great, thank you carsten. I read it through and looks good, but I'm not a native speaker so.

A

Okay, um any other business.

C

So the best part of the rewrite is that it removes most of the principal part, the two parentheticals in the row yeah.

B

Yeah, I noticed that people don't like parentheses.

B

C

um It's really, I think it's really hard. It represents a diversion in your thought um and I think it's really hard for non-native speakers to to follow. um That's my feeling, but I don't know I'd always always imagine I imagine someone speaking it on stage and they would you know kind of whisper it as an aside.

A

C

A

It as a non-native speaker, because I know that in theory you can just skip them and you're supposed.

D

A

The sentence, but if you want more context or details, then you go and read them.

B

So essentially, the correct way to read a paragraph with parenthesis is to read it twice: read it without the parenthesis first and then read it again within parenthesis.

C

So so I totally agree with you and and that's right and the other reason. If we could do footnotes in in our in rfcs, then many of them would become footnotes, because that actually really makes that more clear. um But um but I I anyway, I don't think it's a good way in it. I think that the principles are not are they're, not yeah. I mean anyway, uh it's a good rewrite.

A

Great thank you for reading it through um just note that.

A

Okay, um any other business, any other topic that we want to discuss.

D

A

D

What's the deadline to get uh input in on the best document.

A

D

I mean, I know it's, it's yeah I mean, but if you, if, if additional uh input is submitted, how fast.

A

Does it need to be.

D

A

A

Yeah, I I don't, don't really have a uh an answer to that, but.

A

What do you think when paul will have time to merge this.

B

uh I would expect once, if he's woken up, so I would expect this to happen within the next few hours.

C

A

B

And if there is a bug we have to fix that bug, so there is no deadline for bugs.

D

Understood, of course,.

A

A

uh Okay, anything else.

B

So, what's our plan for the next interim.

A

Yes, the next interim is in two weeks october, 14th.

A

Will we have any updates on the uh packed or proposed dictionaries or oid tags document.

B

So for for the packed thing which for me includes the dictionary discussion, um I think we should have a few proposals by then, so we can discuss which direction we take, um so that that would be something that I would actually uh prefer to have a uh time for, like on on friday october, the 9th we want to see proposals which we then discuss on the mailing list and then in the meeting on the 14th.

B

um So this is something where we really have to do work. It's not done yet. On the oil side, I think we are pretty much done. We could decide that we that what we said uh for weeks ago was not the right uh direction, so we would change again, but if it is the right direction, then I think we are nearing completion on the old document.

A

Okay, so it needs reviews now, at this point.

B

C

uh What I'm so, I think there was some emails about the direction change that I answered about seabor signed versus. um I don't want to call it d-e-r-signed, but recovery is that the thread you're talking about.

B

No, no, the odd thing is simply the question whether we allow any tag factoring and if yes, to what extent- and we had a good discussion four weeks ago- and uh I think we have now hit the right level of complexity.

C

Yeah: okay, yeah. I agree.

A

Okay, so submit the next version and then uh hopefully we get can get. Some feedback got some reviews.

A

A

um What else the cddl.

A

uh Control document.

B

Yeah, I still hope that I have an implementation at some point, so we can all play with it and and get a better impression of whether that's what we actually want.

A

Okay, so we're not gonna, add it to the agenda for next interim.

B

Well, at least, uh let's wait until I have an implementation.

A

Yes, okay, uh if it's only this and hopefully we ha we will have a celebration for 1749 this next meeting um or do you think there's anything else? I would like to add.

A

So for now it would be then the pac document, with possible proposal, multiple.

D

A

So we don't know and the uh oids document we see if we can get reviews before then. So we can talk about the reviews during the meeting right.

A

B

I don't think so.

A

A

Okay, um the last last chance to add anything.

B

A

And if there is any nothing else, then thank you for calling in today very short meeting and talk to you in two weeks.

A