Internet Engineering Task Force Interim Meetings, 12 Dec 2022

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: IETF-RATS-20221212-1500

Description

RATS meeting session at IETF
2022/12/12 1500

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

A

A

There's a um a refresh icon, which is like an arrow, Circle or half circle. Arrow foreign hit, the refresh I, think that will work for you.

A

So we're looking for volunteers who can take notes.

A

He's a couple of people to do that so I'm going to get jump on that here in the next few minutes.

B

I'm logged in to take notes: okay,.

A

Great thanks, Carl one other person.

C

This is Hank I'm, not sure if my audience working but I can.

A

Help with notes.

C

A

Time, thank you very much.

A

All right, I'm in uh Nancy, was having technical difficulties, she's gonna she's, trying to rejoin so I'm, going to kind of wait to see if we can get her to to uh resolve that.

A

I can start with some boilerplate uh chair stuff, so the the uh this is the nope. That's the wrong one, I'm, sorry uh that is okay. This isn't the hf115 reusing those slides, but this is the uh the virtual interim in December of 2022., uh just a refresher on the um note. Well so everybody I think everybody's familiar with the IP policies for the ITF, so just take note of what that is and anything that you say or contribute will be will fit under that IP policy.

A

We are recording the session and uh just a few tips here for the meeting make sure that uh your your audio and video are off unless you're you know chairing or presenting, and uh we recommend use of a headset go to conduct. uh You know be nice, be courteous, um please contribute and uh here's our agenda.

A

It's really one thing which is to continue our focus on the eat, blockers and issues, and so um hopefully we can get through all of those today and and actually uh make this a uh something we can wrap up pretty quickly. So with that I'm going to check and see if Nancy is here still don't see her.

A

Okay still having technical difficulties, so we will move on without her I hope she can join. So uh with that. That's our agenda, and so, if someone wants to make a change or suggest something different, please do so.

A

Otherwise, I think we are on track to begin with the eats update and blockers. So uh Lawrence is that, can you share that or do you want me to share it or.

D

Oh there we go okay, how.

E

Do I share that now, I haven't, haven't sorry, I haven't done this before.

E

D

A

All right there's the screen icon.

A

If you click that you can share.

D

uh Okay over here, okay.

D

A

Are saying the um the meet Echo screen? Okay,.

D

In the right window here sorry.

A

Click on the second list in the in the in the meetings.

E

Now I've got the I just got a lot of windows open here. Yeah.

A

If you want I can share it too, that's not a problem.

E

Yeah just give me a minute I'm. Sorry, I was organized for this.

E

C

This is saying the slides seem to be have accepted by uh through the data tracker materials. So you in theory you can just use them here from etecho right. You don't have.

A

Social yeah, just click on just click on the link and it'll come up.

E

Is that uh does that look right.

A

E

Okay, there we go sorry for the so okay um and.

E

E

Okay, uh here just very quickly here's the changes we made in draft 18, which was published since London, um so secbook was renamed to uh owing Boot and it's described as OEM authorized boot.

E

um uh The um the term attestations um is no longer used in the intentity claim, I think it says now um uh evidence and were attestation results, um and then there was some clarification in the privacy consideration section about freshness and replay protection, and then there was also a number of typographic and other really minor changes. Oh, oh so yeah. So those were the three changes in draft 18 and then um you know honest the review and he found a bunch of little minor things, wording, changes and missing words and wrong capitalization.

E

So we've merged I, think 10 of 10 of his PR PR's into um in GitHub, and those will will show up when draft 19 is published.

E

uh Any comments or questions on that, but before I go on.

D

E

um So these are the remaining issues.

A

With your voice breaking up I, don't know if others are having the same issue.

D

A

Are you in a headset.

C

um To me, it's like stuttering, that's from time to time, seems.

A

To be like pixelated pixelated, a little bit, I, don't know and then maybe get closer to the mic or.

E

Yeah I will get closer to the mic and speak more slowly, um I generally I'm, not using a headset, but generally my Mac works really well. For this um I suspect the problem is I'm on a poor internet connection.

E

And nothing I can do about that.

E

A

This speaking, hey we'll see how it goes.

D

E

um I I'm, actually I.

D

E

D

E

Have not installed starlink yet, um uh so this is issue number ten, um very old. The issue requests guidance on how to create Yang objects for claims in our response here from the authors. Is that we're just focusing on seabor and Jason in in this document and that they Cannon Yang would be a step too far.

E

E

Comments or questions on that.

C

Yeah, so this is saying um just quickly, um so it's about the semantics here ISU, because you can of course um encode Yang in C boy Jason, but that's not the point. I assume that we are not um trying to map um named words and yang those statements, basically uh and and try and make them make them values here. Is that correct.

E

Well, I I, don't know very much about yang, um so the this.

F

Identified such terms.

F

Have you identified any terms that are being reused from Yang in the eat doctor? Oh.

C

No, no, no, no! No! No I'm! Just asking I I I'm, not even familiar with the issue. I'm! Just hearing that we are declaring.

C

um You know the responses overreach for the eat document, which focuses on zebra and Jason, which basically says. Yes, we are doing gang and I I. Don't if there's anything you wanted to say that I think you want to exclude young here, uh but but the response opens it for yeah yeah. So that is my confusion.

F

Okay, yeah go ahead; Lauren, sorry for interrupting.

E

We don't want to change anything in the document. The document does not mention yang and we don't want to add anything about Yang to the document.

B

A

Yeah and I think that was the that was Lawrence's request. If somebody had an objection, I think hey Hank was just clarifying that uh that he was reading the response correctly, which I think is that he said my interpretation was that that his response was in line with the direction which is to not include Yang in the uh in the internet. Draft.

C

Exactly yeah, so if that doesn't the response does not reflect that. But if you chose the issue it's gone so that's also good.

C

E

This was a request to reorder the sections in the eat document.

E

um The request put some of the claim definitions put put their request, put the claim definitions later in the document.

E

This issue is from over a year ago, the document has been reordered in some other ways, since also the document. The order of the the document matches the order of JWT and cwt.

E

So again, no uh I mean I would like to close this issue.

A

D

You appreciate.

A

That again, what was the other document just for the purpose of the notes, takers.

E

uh So the ordering of sections in eat matches the ordering of sections in JWT and cwt. Okay.

D

E

Yeah, both of them have the claims. Definitions like the third or you know, it's the first intersection. The claim definitions is the first major section.

G

With no change for this one and in the original uh thing that's filed, uh Michael says I, wonder if, as opposed to please move it, here's the problem, whatever it said, I wonder if so it wasn't strong to begin with and so I think it's fine to just close this one.

D

E

E

So this issue number two and I won- had two parts.

E

One is that the issue asks how s-u-e-I-d labels are assigned.

E

The the text in the document and since uh draft 13.

E

explains that their assignment is intentionally left open.

E

So basically the document already says the labels are not, as uh are, there is no process for assigning the labels.

E

So the issue only asked how they were assigned didn't say that the method of assignment was wrong or that the none method for assignment was wrong. It only asked how they were assigned.

G

They failure this might be a case, we're just responding in the issue to say it's allowed and left up in the profile, which seems to be the answer and not change the document and just responding in the issue for record.

E

Yeah that response was provided in the issue a month ago.

G

uh It's up to the profile if a profile wants to require it or prevent it or whatever that a profile can do that right.

E

Yeah the response does not mention profiles, but I, don't think it needs to it can. But I don't think it's too.

G

You could add that comment in GitHub and close it uh I think. The text in the document is fine. Okay,.

E

Any other comment on the Su Eid labels.

E

Okay, the issue also asks if the ueid and Su Eid can be the same.

E

um The response is that, yes, they can be the same.

E

There's nothing in the document that says whether they could be the same or not, but I believe the rules for ueid and Su Eid are clear and we do not have to explicitly say that they can be the same.

E

The basically the rules imply that they can be the same, but there are, you know, rules for when an Su Eid can change and when a UE ID can change, so they might start out the same, but an s-u-e-I-d can change, and that would where a ueid can't so they might through a life cycle event that might become different, even though they started the same, but I believe the rules are clear and don't need everything.

G

Comments that I made earlier were actually on the bottom half of the slide about uh what we would do in the GitHub is commenting on the bottom half not of the top half sorry, if that wasn't clear before.

E

Okay, how would profiles relate to this.

G

All right, Lawrence's uh was choppy so I'll. Just repeat what I said. My comments before was um uh adding a comment in the GitHub and then closing it, but not changing the text and my suggestion for adding a comment in the GitHub which I could do is. Did you say that um the the UE ID and the suid may or may not be the same? The document, the the spec does not constrain it.

G

uh If there's any constraints necessary, then a profile constraint whether they must be the same or they must be different or whatever else any other constraints in the relationships. There is up to a profile, but right now there's no constraints, and we can just say that in the GitHub issue and close it because I think the text is.

B

G

B

Have agreement to close the bottom half of the slide? What about on the top half is? Is that also agree to close with the won't fix or or.

A

Nobody's on the Queue I assume, the answer is yes, foreign.

E

So this issue requests that the measurement results claim be removed uh because it is too general for a relying party to be able to interpret.

E

The you know, as you know, the measurement results claim um represents the results of a measurement. It's not the attestation results.

E

Although it might be inside data station results, foreign results claim can also be in evidence. If the a tester has the ability to evaluate measurement result measurements, and there are products out there that do this- that have the um known good values. The attacker has the known good values.

E

um So the the measurement results claim does have a simple pass fail option you can just say extension result with asked or failed and that's something that an RP would know how to interpret.

E

Also RP always has to understand the verifiers policy that just does a basic uh characteristic of the rat's architecture, so I mean in that sense every claim is in attestation. Results is the RP needs additional information to interpret any claim in attestation results.

E

The point is that the measurement results claim explains why it is General and there is a whole detailed paragraph explaining this.

E

So again, um we believe the text is fine, as is.

A

D

G

Or uh when there's complete silence, if I just jump in um yeah I, think that the text is uh fine. I think the one thing that the working group may want to discuss and I noticed that Eric Voyage is not on the call, uh but I think it might be appropriate for the ar-4si document and talk about the relationship between uh this claim and the trustworthiness vectors I.

G

Don't think any change is necessary to eat spec, but I think that a discussion of this uh question in the era for us I document might actually be very useful.

E

Sounds good to me.

A

So do we do we need to create an issue in the Air, Force I repo, then to capture this.

G

Or so meaning uh in the Air Force I document right, because the profiles can convey uh what our allowed claims prohibited claims and so on, and is this one of the loud plan? I have to go look and they reference High, because I don't remember off top my head.

F

So this is Gary here can I add a label of won't fix. Keep it open indicate that on the issue that we think this issue would be better taking handling the Air Force eye document, if I close it, then you'll lose track of some of the right up. The warrants is provided in there and also the decision we made today.

E

G

Really like to close it, if we create a new issue in every reference, this one will be discoverable, even if you close it yeah.

F

We'll close it and then uh I'll put the note in there that will say AR for I will add the label we'll fix, but I'll, uh but uh we'll put in their ar-4s I will uh we'll take this issue and I'll see if I can put a link to the ar-4si repo in the in that context as well. Okay sounds good. Thank you. Foreign.

G

To this claim anywhere mentioned in the ar-4si document, and so I think it is valid to have something to say it. Is this one an allowed claim and if so, what's the relationship uh between this and the trust or the specter.

E

Okay, moving on to the next.

E

uh This is this was filed against raft. 13 draft 13 had a normative session on endorsements and verification keys.

E

That section was moved to appendix f, so this.

E

D

There was a significant.

E

Change to the document in this area, since this claim Earth's issue was filed, foreign this issue and I believe in the comments and discussion on GitHub, uh evolved to a request for specification of methods for verification keys and endorsement identification.

E

I think it was actually requesting.

E

Detailed investigations of methods, um so the response here is appendix F today provides good examples of different ways to specify the key, so you could can do it with UE ID. You can do it with x-time, uncertificates and or the cause a kid, so it appendix F describes those examples. They're. Just examples, though,.

E

Another Point here is that the means for handling verification, keys and endorsement identification are going to vary quite a bit from use case to use case so much so that it's you don't see it as possible to pin down to in particular thing again.

E

um That would be something that would be addressed by a profile and also this parallels. What happens in kosee does not specify any end-to-end. You know key identification scheme.

E

um Basically, the we think that the method for identification, identification of key or the endorsement should be in a separate document or other documents.

E

um For example, perhaps somebody create a document that says how to use x509 certificates as endorsements.

E

So we believe there are no changes to be made in the document. We do not want to strike out and provide a whole the identification scheme. We feel the examples are sufficient.

A

G

C

G

But I think you can rest.

A

Can you repeat the question Nancy.

F

Yeah, could you transcribe it if you're having trouble with your uh with your audio I'm, not sure if it's a meat, echo or the endpoint or an endpoint issue, but I couldn't understand a word that.

E

Yeah yeah Nancy I couldn't understand it, but you said either the audio broke up.

F

F

Ty, you got the answer from Nancy in the chat one.

E

Okay, I'm yeah I'm sharing my screen that we're not the king of that.

A

I can read it out. She says uh it was a comment to say you have an example in appendix F, so I think it's okay to consider this PR result foreign.

E

Objections to closing this.

A

Thanks on the queue.

C

Yeah sorry I mean two meetings. At the same time, that's a little bit difficult, um so I heard um so I was confused by the issue of the first place. To be honest, so I have no problems with it being closed and not acted on as I I think it's about a key identification standards and I. Don't think we want to go there. So I would be fine with um just abandoning the issue.

C

I think Michael, who can't defend his issue, had something for this, but I can't put his head on because I literally don't understand it well enough. So I I think uh if there's no um feedback from MCR required, because we could abandon it. Yes, foreign.

E

E

We will close this issue.

E

So that is it for um the the issues except for one that was filed a couple of days ago by Thomas, and we have not worked through this issue completely. Yet.

E

E

There were, there were a few other PRS and issues filed uh in the last week or so um I have not put them in the slides or I'm, not bringing them up here, because I believe they are all um just a procedural or wording.

E

There's one about how to form the cddl correctly. You can follow them on GitHub, but I believe they are not important or that they don't require. The discussion at the meeting here.

E

A

What what is your plan for for the resolution for those?

A

Is it to generate another revision of the draft.

E

I think we will generate a draft 19 uh in the next uh a week or two before holidays.

E

um uh Mostly it will be. uh You know, since we haven't had any anything here so far, that's required to change. It will mostly be uh just fixing the uh the spelling and the typographic errors and I think there will be one cddl adjustment but relatively more, and it doesn't change the semantics. It's just a question of how to use CDL most actively.

D

E

A

Curious on the queue.

F

There's questions for the chairs um we put out the draft 19. will.

H

F

Can we take it to iefg review um I, understand I'm assuming most working group members, particularly the ones who have gone through, who have are who have gone through the uh RFC editing process before will know that this is not draft 19 is going to be far from the last draft, but um in fact we're anticipating uh I think we can safely anticipate several several other versions and uh in addressing the iesg comments and if the, if the chairs that reasonable Way Forward can they describe what the process will be after we put out the 19 draft.

A

So I want to get a feel for how many people have read the you know 18 and are familiar with with that one uh I, don't um the Assumption being that 19 is going to be kind of wording changes. You know. Minor wording changes to 18..

A

uh So if uh but I I do want to get a sense for how many people have uh are familiar with 18 and.

A

So uh well, let me I'm gonna, think of some I'm gonna. Maybe do show hands here.

E

Well, we know honest has read that thoroughly: hey.

A

Thanks on the Hanks on the queue.

C

Yeah and that's so I so yeah, this was off timing. I mean two things here. um So the problem is that um I, you want to have a um uh a poll first, because then I will postpone my question: yeah. Okay, sorry.

A

C

A

C

A

A poll to see how many people have read uh version 18.

C

Oh in context of this issue, or in general.

A

C

C

Ldq for now, but for this issue, I think I still have a comment so.

H

I like to say that harness has read it so plus one is not it can be here. Let's see, he said you know, go ahead. Okay,.

A

H

Relay my message so I'm I'm relaying it.

A

Okay, so add one more.

A

It looks like it's about evenly split. uh Eight people participating five red through it three, if not of those that are participating. So anybody else. Okay, it looks like it's settled down, so it sounds like it seems like us. Folks are familiar with it, uh so the next question I have is um if there are only sort of minor wording, changes in version 19.

A

Do we think it's ready uh for for last call.

F

B

You putting out.

F

A poll on that that are.

A

You going to get.

F

Some mailing list- okay,.

A

Take some time to type it.

B

Clarify what last call you're, referring to Gary specifically asked for iesg last call.

G

Right, that's what I think your question should be is: is it okay to submit to the isg now.

F

Yes, I can confirm that, with my original request.

A

Okay, so try this.

A

A

Should I assume that honest was going to vote for this.

H

E

Harnesses reviews.

A

What was that again? Nancy.

D

A

Okay, it looks like it's settled down.

A

So so eight seven seven uh hands in favor one and not raised the person that did not raise their hand. Did they want to speak and explain their reason for not raising their hand.

A

A

So I'll conclude from the voting that where it seems as though uh version 19 would be a reasonable candidate for isg review, but we will want to wait until that that is out and we've had a chance to look at it before we actually uh recommend it for ASU review, but I think that can be done. You know offline on the list. I don't know that we need to meet again.

A

Hank has this hand up.

C

Yeah so um I think that's where we are through the slide deck so I was gonna. Come back to the uh um issue that we were looking at um just a moment ago. I can't see it so I have to look at it. I'll pass through at first so yeah, um so I think the problem in general that I still perceive for this issue.

C

Why I think it should be a fix, is what uh Thomas commented on four days ago that when you do this should normative statement you that it should be coming with a rationale? What the exception is when so, when is it okay to format it and or how would it be uh obvious to the uh relying party that it's not confronted with evidence and and I think that is a valid concern? To be honest, I I wouldn't know how to use that at the moment. I think.

A

Your comment is aimed at whom.

C

uh At the authors of it, the maintainers of the issue at the people who say it's okay to have a shot without explaining how to use it.

A

This one of the authors want to respond.

F

um I'll I hate to respond to question with the question, but Hank would uh would removal of the rfc2119 language uh uh alleviate alleviate the concern. From your perspective, I know: we've had offline discussions at Thomas, so uh I'm not I'm, not going to ask you to speak for him.

C

Oh yeah, no I'm, speaking for myself, so I I just want to agree with the Thomas's Point here, so that's, basically it and um no removing it would make it worse. That would allow it right.

C

That's not that's not good! Then you clearly can state that yeah go ahead. Send evidence to the underlying party, and- and some of that feels weird to me when is this okay? How can an implementer tell that this is okay? How do you mark uh evidence as such when you send it to a relying party, because it has to go to the verify in the first place? So so that is, that is things that I would like to.

F

C

Guides about yeah yeah.

F

So so, first off I've, already I've, already stated my view on this. There's a they're a perfectly valid attestation architectures, where the verifier is integrated into the relying party. So when you talk about sending something from the verifier to the relying party you're, assuming the specific architecture.

C

You're doing an e that is attestation results.

F

I'm saying I'm saying you don't sit if the verifier is integrated into the relying party. There's not the concept of sending yeah is is not clear to me what that means.

C

Okay, so because you'd have a verifier that I don't know, internally communicates its findings through the same system. That is the relying party also I got it so, but then you're never sending uh evidence to Airline party effectively you're, sending it to a verifier. So there's no reason.

F

You're sending it to the relying party, the relying party happens to include the verifier functionality in it.

C

Yeah, but the architecture always sends things to roles that it happens to be on the same entity once.

F

Again, you're talking about an architecture, the one that you're you're a co-editor of it's, not all the that's, not the only uh architecture that they uh yeah.

C

But he's an eid's ID and, and it's it's it's it's it's uh according to the Reds working group of things, I think that I'm confused more. So this is um how well, how is this idea informed not by the architecture of Reds? Then it.

F

No I think it's meant to cover more than that right, I think! It's that's why it's a that's! Why it's softened to a should requirement.

F

H

Saying two receptions.

F

The exceptions are quite uh are quite lengthy when you would, uh when you would not want to do this, that I don't think we can give the proper uh I personally speaking as an individual, don't think we can do proper Justice to this discussion and uh inside of the eat document. No.

A

uh So we, the Dave Thayer's on the Queue I'm, just learning how to disposition this conversation I'm wondering if we need to capture it in the in the issues uh tracker. um It's.

F

Already there yeah my.

G

Comment has done a different topic, so okay, I raised my hand about the same time that uh Hank hit it up on a different.

A

Time, okay and then Carl raised his hand. Is it related to this current topic? Carl? Yes,.

B

It's on this topic. um Okay, so should I just go ahead. Dave.

A

Go ahead, yeah.

B

Yeah we're just gonna come back today. I was just gonna say: we've we've talked about four possible uh ways to to settle this. One Gary mentioned removing the the 2119 language one. The must not that that's in the issue that was the the initial submission one was to to add a brief rationale to the should, not that that says, except where a profile identifies, why you're you're you're not adhering to this, should not, and then the other option was just no change at all.

B

So those are the four things we've talked about, and only one of them got talked about here. Really so I don't know. If, if either of the the other options, there Hank catches your ear differently, but that that's where we are on addressing the issue to my eye anyway,.

A

So we do expect that that any address any rest drafts will adhere to the rat's architecture.

A

So so, if that's the.

F

Case based on what uh what sorry, but that's uh when you say, expect.

F

That's not necessarily going to apply. You know you, you know usages of that, but, for instance, a web often uh registry, which is an ietf RFC industry-wide security A.D, if it's not possible for an attestation format to actually uh to actually Define distinguish claims between evidence and that and results that is uh I'm.

F

An expert reviewer on that uh for that particular registry I'd have to reject it, because the architecture has no contemplation of such a uh such a distinction so and one of the first one of the first things me and Lawrence were asked by the security IDs at the time when we submitted uh eat for consideration by the ITF was, can this? Can this be applied to Fido? I? Have to look it up? It was an email from about four or five years ago, but uh but so I, don't really uh I.

F

Think if you, if we're going to say that I think eat applies in the context of the ecosystems in which it's uh in which it's being used, not eat, is not strictly adherent to the rat's architecture.

A

So I think the way I, don't I, don't think any I. Don't think that that the rats architecture is trying to say that eat can't apply to Fido.

A

What we want to get to is an explanation for how the rat's architecture is interpreted such that it that a phyto implementation of it makes sense. In other words, the just the description of the application of rats. Is you know, aligned with or in harmony with the rat's architecture? Okay, I, don't think. That's an architectural thing.

C

Yeah yeah, and that stated such in the um in the eat document right. It says the eat format follows the operational model described in figure, one. That is where evidence only evidence only goes to the verifier threats architecture. It literally says that so now, you're telling me it is not doing that.

F

Yeah, but it can, but it doesn't happen, it doesn't have to right. I mean it's: uh it's there we've uh the document is, is written as such, but let me rephrase this. The document is conformant with the architecture document I. Think in this case, though, the language was kept, vague and I. Think it's the right thing to do to allow for the different exceptions. Let's leave aside Fido for a second um evidence can be evidence. Cannot it can be complete conveyed, not in a token format.

F

Oh sorry, results can be conveyed, not in a token format, but in a UCCS format. The working group made a very made a decision to remove all mention of UCCS in this document. Therefore, it's up to the UCCS saw editors to decide when, for instance, an IAT claim or any other claim, that's the Providence of the token issue gets conveyed as evidence. All right, sorry results.

F

Sorry, you get confused. So if we can't really discuss that here in this document.

C

I so um so I I want to highlight, but I already forgot what it was that the second option card I mentioned was nice, but I literally forgot what it was to the discussion right now: um I'm confused about the non-u, the mentioned the relationship to UCCS I'm, also unclear now about.

F

um No, no we've removed everything, but we're not. It doesn't prohibit the IAT claim from being passed through as a result in the form of a UCCS right.

F

So how that gets handled is up to the UCCS document. Ary.

A

Is Lawrence your comment on this threat? Also, sorry, sorry, Dave.

A

Lawrence, do you have audio.

A

I, don't I don't hear Lawrence if he's speaking Yeah.

E

Okay uh to me, uh the rats architecture does inform eat and really now you know, we've worked with this architecture: the rats architecture for a long time years, um I, don't I, think it is flexible enough to accommodate Fido I. Don't think um we have to rearrange anything I'm here um and I'm kind of puzzled. By this discussion, I mean uh the the whole document. Is you know it does use the rats architecture terms. um I really would not I don't want to revisit the whole rats architecture.

E

Discussion about the separation um I also point out that all of the claims are neutral, I, think every single claim is neutral as to whether it goes into evidence or attestation results. I mean I was intentionally trying to do that. Yes,.

A

I think I think what we need is to capture the there were like four Carl mentioned. There were four different options that hadn't been discussed. We should capture those four options in the issue uh and then have whatever appropriate discussion is needed. I I think that you know the conversation around hey. We don't have to be compliant with rat's architecture uh is just I I, don't uh it's.

F

Moved it seems.

A

Like it's a mood or false yeah.

F

Yeah, it's a red herring. Don't worry about that.

D

F

The document is, the document is conformant. Even the last time, I had a little discussion with David Hank on the GitHub repo. It was already resolved in favor of the rats architecture. So it's so don't don't worry about that. Carl's already said he's going to put the four options in there into the repo.

F

The problem is that I, don't know how you gauge consensus, I'd like to get this into the dash 19 draft. If there is a resolution on this now, um can we just uh can? Can you delegate this to a smaller team? Like you know,.

H

F

I meant the guys are already looking into the repo just as the reviewers. That's all yeah yeah.

A

F

A

People the people who care about the issue should work together on it.

F

Yeah, that's that's what it's meant Lawrence, it's not it yeah! Okay, yeah! Well! So.

E

What we'll do is.

F

When we come up with the mutually Global PR, if we do we'll designate the reviewers to sign off on it and then.

A

Okay and it'll be great.

F

A

Yeah that'd be great yep.

F

Lawrence, are you okay with that, too.

E

I'm, basically, okay, with almost any option here just to get the document through. You know, with the understanding that uh you know there, there will be further drafts here. So I I really don't want this issue. To hang up. Hang us up for iesg review. That's the that's my top priority, so whatever we can do to get through into into IHS iesg reviews sooner rather than later is what I want to do.

F

A

uh Dave thanks for waiting.

A

Dave Taylor.

G

Yep sorry I was just getting off of mute here.

A

G

Last one I'm in the same spot as Lawrence as long as it doesn't regress. Anything then I don't care so um uh the issue that I want to talk about, given, though I think we have all the relevant parties on the call right now is I, didn't see in the slides unless I missed it issue 343, which is the IAT definition that Thomas mentioned, and so there's a discussion between Thomas and Carl and Lawrence, and if I uh understood right.

G

uh The issue is that the cddo allows time to be in either integer or float format, but the eat uh disallows it and we needed to decide how to do that in the document. There's a current pull request open that proposes doing so in prose meaning the ctel still allows float.

G

But the text in the document says: don't use, float, use, integer um I approved it this morning and it looks like Thomas just did as well, but wanted to give visibility to the rest of the group, um but I didn't quite follow why we could not do it in the cddl, which were a lot better too in things, but I have no objection either way. So I just want to give a minute for people to discuss this issue.

E

Let me respond quickly. Is that I didn't put it in the slides, because I thought it was just a cddl issue that didn't require. uh You know, there's no controversy about what uh what the document you.

H

Know the the intent.

E

Here it is just what is the best cddl to represent this and I Carston made a comment about uh dot within and Dot, and that I was going to look into, and maybe that solves it.

B

Well, I I'm, the only voice, that's been against doing it in cddl, so I I'll speak and the reason I'm saying this is this claim isn't defined in this document, it's being reused from somewhere else and I. Think if you, if you claim to reuse something, then change the definition and change the pros. You're, no longer reusing, and if we get to the point where cddl Imports are used now, you've got kind of this weird animal where eats using it one way and other stuff is using it some other way.

B

So that's that's the only reason and it's kind of pedantic, but that's why I I was against uh changing the cddl profile.

A

B

D

B

Was exactly the way to do it, I think right. In my opinion,.

G

uh So the question I was going to ask Carlos, which document is this being reused from.

B

The cwt RFC, which inherits from the jwtrfc.

G

Gotcha, thank you um so I I did not understand your comment in there. So yeah I already approved it, which is doing it in prose, but didn't quite understand uh why I didn't object either way. So just wanted to thank you for the elaboration. Okay.

A

So we're going to have to cut the conversation we're at time. uh So, um let's resolve offline uh captured in an issue, work the issue through the list and, uh let's hope, let's shoot for getting that resolution before version 19.

D

A

uh We'll move forward in that direction, so thanks everybody for attending and we will look forward to continuing this conversation on the list. Any last uh remarks from any of the other chairs foreign.

A

Thank you to the notes. Takers, I, appreciate that, and uh you know, have a good rest of the month and and have a happy holidays.

H