Internet Engineering Task Force RATS Architecture Design Team, 12 May 2020

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: RATS Architecture Design Team, 2020-05-12

Description

RATS Architecture Design Team, 2020-05-12

A

A

So have four, maybe there's five. Six now all requests I noted to Hank that his pull request. What was it number 86 or something.

A

77 I think it was an 86 update. The same paragraph I think we should look at Ned's 86 first, that works I had no problem with it, except then I got to this last part and then I bogged down and trying to figure out how much of this was.

B

The last I checked didn't been updated since last call, whereas 85 has been okay, let's go to 85.

A

There we go: let's talk about this stuff by the way you're, not sharing your screen. Oh I was sorry. You.

C

Know it's updated correct.

B

Yeah I, look at the other ones and hadn't been updated when I checked so does.

A

That work shared now.

B

You're starting to share content.

C

B

Only works when you have a net.

C

A

Well, maybe you have to let someone else share that then.

A

That's the problem, just don't know what the what what address at thinks it wants to use.

B

Well, if you like, since 85, was mine, I can share my screen 485 and.

A

B

The other ones please.

A

Do that and I'll see if we'll try to see if mine works again in a moment, yep.

B

Hey somebody say when they can see my screen. Yes, so it's my screen sure is working. Ok, so.

C

I was it works because, in this sequence of commits it shows two github IDs and it's actually three. So if you click to the latest one and then press next, which Michael was so polite to indicate to me, you see the finder and that's actually the big change of this series of comments, which is weird okay, so.

B

Here's the intro, if you remember we talked about this one last week when it had a different title right, where I tried to invent a term which some people like there's a bunch of discussion and the in conclusion was in general, we couldn't get any consensus on there being a term there's as much things. People didn't like and so I basically rewrote it and rewrote it rather than defining other term I read out it the way they did not require defining a term hey and so I was trying to find a way to do that.

B

That did not require defining a term for these three potentially different things right, because these could all be different things right if your firmware you're, biased or you could be different things, and so instead, what I chose to do is I chose to pick a specific example and then I only need to use one of those terms in that particular example. And so that's why you'll see a bunch of text changes, but it's probably because I was rearranging paragraphs and doesn't do a great job of showing text is.

A

B

Because this sentence, this point was actually made about here and I kept that sentence. I may have changed a word in it or something, but this fixed me was about layered attestation, which is this section here and so I deleted this, because it was actually done.

B

I am gonna, go in to display their rich gif to let you guys read this okay. Here we go cool.

A

B

A

Just played if yeah.

B

So just play the security. Now, of course you know you don't get things like that references, yet so I try to do this. The old diagram, but just so you know, the old diagram has exactly three levels: there is a tester environments, a B and C, sorry, there's environments, a P and C right, and so there's only two layers.

B

If you will right a it has to be, you know a measures b and b measures, C, okay and so I had to make sure that the example used exactly three levels, and so the example that I picked used a BIOS, Jordan ROM BIOS could be ROM or an nvram or something so I picked a BIOS sort and aramis.

B

Why you wouldn't have to go into the issue of who measures the BIOS, ROM and then I kept the rest of the discussion that Hank had whether the next level was the bootloader and the next level after that was a kernel and so I said a was ROM, which in this case was a bias in Iram II was the bootloader and C I could have gone with four levels.

B

If I had you know hardware and BIOS and NVRAM, but I would didn't want to have to change the diagram and so I kept this particular three level one, but I always made it clear that this is an example and then I had a paragraph at the end saying there could be other variations yet.

D

B

I wanted to walk through to give the full thing with that particular example. So I guess I'm going to pause the people when I read it or you want me to read it for you or what would be the most useful to you guys.

B

Actually, this.

D

Is actually sorry because I'm.

B

Seeing that some of the these things are strike through, and so that might not be easy, read this yeah.

C

So, showing more of the strike, through contrast, the edit good well.

B

The problem is everything had this, you know 1/2/3 thing, all of those are removed right so here is gone. Here is gone.

B

One example not trying to combine all three examples right and so as get rid of all of these use cases here, which is what the fix me was about, how do we get rid of, or how do we come up with a term for all of those? And so that's why I tried to rewrite them using it reference to those three levels, the diagram and so I'm gonna skip over all the red here and I'm gonna. Let you read the green here, I'm, keeping in mind that I'm, where these sentences will be similar.

B

Maybe with the word changed like you saw the UV BIOS firmware change to a different phrase.

C

So, looking at the arrival reading looking at the this commit basically commit batch. You also address the signing question of Thomas yeah, correcting.

B

Thomas responded, Thank, You Thomas for hearing so quickly and in Hanks original text. Heat I had this phrase here and when I rewrote it. I had missed that part and put in the part of all signing, and the point is both are valid ways to do it, and so that's the or here, where I took this phrase, that Hank and put it back in.

A

B

Sorry, I'm just pausing what people read unless they want to bring up something or I.

A

Just I just was worried that I had broken something at my end. Well, you stopped talking I I, believe I, believe that I believe that WebEx pinups ec2 is in order to do stuff and, of course, I'm not an added towards them.

A

B

People tell me when you want me to scroll or feel free to stop and talk about some sense in the screen.

C

It's site-by-site version on my screen, so I saw.

E

A difference between shielded and secure.

B

You are I, don't know, and so I left the phrase as Hank head it because I don't know the answer to that.

B

I think the word that Thomas used when Thomas you comment. Did you use the word sealed but Hanks text, which I was trying to preserve as much of the original text as possible unless I thought it was wrong, and so here I didn't know enough to say it was wrong. So I kept it.

A

Changes at the bottom is the delete of the time RR in the example. One time based passport model example. Is that just a goof or did you intend to put that in that I.

B

A

B

You're asking about this part time I intended to, although I didn't intend to in this same PR, yeah.

D

B

In there, in a previous edit, and it.

D

B

Unrelated to the rest here, this right here is not referenced in any of the text in this section and so right now all the text hits all the other ones, but this one wasn't, and so it was not necessary for understandings. That's why I deleted this one right here, cuz, it wasn't necessary in that section.

C

So maybe the the reference was just not captured or the Texas the guiding principle I. Don't know. Probably people draw the drawing in diagrams to know what they're trying to express but sometimes forgets to write. It's speaking to us, I think.

B

It was my cut and paste error or okay from a different one, and then, after reading, the text I didn't notice that that of it anyway, okay defined.

E

B

The table anymore, because it's still used in a different example right. That's why there's a cut and paste area Congress time at which the result was relayed at by the ax tester in the passport model.

B

The event is still there, but the time is not used for anything, and so it's not necessary and then the picture to go along with the text and that in that particular example, all right, wala.

E

C

Coming back to next question: yes, I think there is a difference between shielded and secure.

C

She could be the I as a secure. Of course, the methods are secure and, and of course shielding is a specific kind of method. So, yes, there is some overlap. Yeah, but I'm not going in to shield it here directly, because sometimes you have like a secret. You will never I never ever intended to see that is not just securely stored, it's literally fielded, so you can never really see it. You only operator on it, so you have this puffs or some other things that exists in nature. Yeah.

B

It sounds like they weren't shielded in that sense would not apply to claims which actually got any evidence. This.

C

Can never go into evidence, only operate Leon and be used for creating evidence and such but shielded means. Maybe we should elaborate on this, but I didn't want to go to the depths of it. So.

D

Can we instead use an temporal because that's how I.

C

B

Think we roots you allude from this sentence and then I think you were saying the suggestion was instead of just saying in a secure manner. Can you say in an untenable manner, which is more precise than a secure I mean.

C

Something third here are not only secure but intemperate, but also never obtainable, but.

D

Then, how can you use it as a claim right.

C

Exactly you cannot use it as okay. This.

B

Is about things that are claims.

B

Suggestion, oh yeah.

C

You you do you and disentangle the sentence with the environment a and a stuff. So yes, but it's even.

B

The old text- it was no context of frames and so I think the comment would have applied in your text here. Crops.

C

So maybe I was conveying the meaning here so signing could be used. You could use a shielded secret that you never ever be able to see only.

C

Check that okay I have it on my screen. What I try to convey here was the signing operation might using secrets that you will never be able to, or that are never intended to be a visible. That's a very strong prerequisite and it's an example of how you can do things. You can also store them securely. I can enter a key store. Well, you have to open that, and in the moment it's opened, it's it's still more attackable, as she in secret by default, is intended to never be. Let attack avoid so.

B

This is saying: how do you make sure that the next environment here be can't tamper with claims that were measured by a okay and so says? There's two ways to do that: okay, you can either have a sign them, and so it's signed by a key or a secret doesn't have, and so you could tamper with it. They would invalidate the signature and you'd fill to attest right. So that's the first part or you can take the claims and put them somewhere bearbie, just can't possibly tamper with them right.

B

It's not it's stored someplace, then you just put it in the evidence directly out of Sabine ever has a chance to touch it right. Those are the two cases or and when.

F

B

F

You say standing here, you're not talking about signing as in attestation. You just start talking about it as an implementation technique for protecting those things. I.

B

Was talking about it in the signature that could appear in the evidence itself, but it could be either if you're asking which one I was thinking. When I put the text in there, it was actually the the case where it's actually a signature, and that signature appears in the evidence.

B

That's actually how we implement dice I, don't know, throw their dice implementations, that's how we actually implement dice ICO.

C

Dice look this way, but other claims might not be huge in that way. That's.

B

Why this phrase is trying to accomplish yeah? There's many ways to do it right and some people are familiar with the second technique. Some people are familiar the first technique and the point is it's layer? That's a station to get the concept it doesn't matter which one you're doing it's the same concept right. That's a measures, BB major, so yeah.

B

Apparently I can't commit the suggestion. I don't know why, but maybe I have to finish the review. First I think Michael Aranda to that last night, all right I'm going to go back to that wallet, I'm going to go back to the wall of text and then scroll down past there, all right so before I commit that I'll. Let people finish reading that wall of text or pick honor.

C

A

B

Guess and get line breaks are significant and it can't deal with too many moves right, because it can only show pluses and minuses it can't show. If you reorder to paragraphs, one of them will always show is new.

B

B

Please check my wording here because I'm using signs here and I want to make sure there's not some other variation, where this text is only correct for a particular example, I could say it's an example. If that's the case, but there's a better wording. Let me know on the the measured and signed by here is a better wording. I mean I think that this is because it's how I and familiar with doing dice, but if there's another way that I need to change the text, happy teeth.

C

You'll remove a root of trust reference in the last paragraph, such as idiomatic, so I appreciate the.

B

Because I was taught I picked an example where it was in Ramah, and the previous text was talking about how you need a hard word of trust. And if you have wrong, then you can argue that the ramens or ever heard of trust, but I can I admitted that phrase. You are correct, I, don't know whether that's good or bad.

C

There you don't want to use it in any case, so I'm not really married to that. So, let's see how this you have to have the polishes at the end. Okay, so you can leave it out now, because our premise is to remove them by possible.

B

B

And this reference at the end here is the diagram again.

B

I think I'm gonna scroll here. If people want to stop reading for a second, you can see. The diagram is referenced here, hey and prior to my text. The old text never actually referenced the diagram number. So we never say you know, figure six or whatever the number was and so I referenced it twice right. One is right. Above it you know six or whatever depicts an example and then at the very end, that's the same reference.

B

When I say this could be extended at the other variations all right.

E

So the only talks about the environment where the code is running or loaded and is.

E

D

E

It's the claims are signed or they're stored. I, don't think, that's the same as any kind of description on where the code is actually running.

B

Okay, that's an observation. Is there the.

E

Observation is theirs, mmm it's just maybe more gentle general for any of these diagrams, but there's the observation that or NASA an expectation that the environments in which the target environment is that both the the attesting and the target environments are actually running is as meaningful.

E

Isolation, properties.

B

E

B

Say anything about that and I: don't know if there's something we need to say about it, so it didn't occur to me to put anything in when it wasn't in there in the text before it's.

E

Just it's an assumption that maybe it's obvious without explicitly stating it when.

B

People finish reading the wall of text yet or anyway, they're still reading the little tips on the screen and I scroll.

E

B

I'm, just gonna go up here and I'm going to commit this suggestion. Then.

B

B

That had to address this well, the to fix means right. Those one fix me that was technically already done before, but this was the main fix me here. We're looking for a better term then- and that's the one I tried to address here so I- claim that both of these are done this one. If you read the wall of text, you probably spotted it where it talked about in one of the maybe the second paragraph.

B

Here this this sentence right here, is the one that I kept. That did that fix me at this stage of this cycle of the device. The claims collected typically cannot be composed into evidence right that event. That cements was argued here before, and that was the one that actually had done. That particular fixed me before right requirement measurements the early stage, you're, not evidence yet such does not cover that. Yet that was that sentence that actually did that.

B

So, okay, so that's it didn't get their comments or should we merge this will on it go on to the next vlog? What do you guys think I think.

E

It's ready to merge Bridget.

B

Agree all right, I'm gonna go ahead and merge this one and I after I merge it I will stop sharing. If so many of us wants to take over just going to sharing to the next one and I'm deleting my trash okay, I'm gonna stop sharing Michael. You want to try again now and you figure it out. Let's try sure I doubt it'll work, but let's try.

A

B

Not yet, it still should starting to share content, which makes it sound like yeah we're trying to negotiate with stone or something, and it's not completing I get.

A

I guess I actually get ya complaints from browser of a stun things actually not getting the right now.

B

A

Is gone, yeah yeah, just if you'll just bear with me for one second: what what what ticket are we gonna talk about? Let's, let's talk about that question as I just try and see what happens here.

A

I'm curious what UDP packets are streaming out of my system.

A

Alright can't just dump everything it's just as too much right, there's nothing leaving of important all right well, anyway. Okay, so we had. That was one full request: the one that I'm really I, let's go to Ned's pull request 86 when you pull that one up, Dave I.

B

Can get me to share again yeah, please.

B

Yeah when I look at from the bottom up, the last things that I see are comments by people, and you know seven days ago, which was our meeting last week and there's been no commits the since, before the last meeting, so I don't know they stuff to talk about on this one or not. Yet, okay,.

A

So, as my recollection, we talked a lot about this and kind of ran at a time, so.

A

A

I, don't I, don't know that I mean you, you had you had comments. We had some discussion about service and service provider. I think that was the major one. That I went through a kind of review of what I thought and so Lawrence had a bunch of comments that I guess we need to walk through or Lawrence needs to provide, suggest some changes or argue some changes with dead. There.

B

Are two commits? Maybe it was during the meeting last time. I don't know this is like seven days ago. You see those two ones and then law rested or review act at that and generate I went down.

A

Here: yeah, that's right! So so those are I, think we're done. Lawrence comments were after the meeting and so I, don't think Ned you haven't responded, but let's, let's see if we can deal something with this because I think there was a fair bit of content here. I might.

B

Go to the files change these, you can see the comments in line.

B

So I guess you want to walk through Lawrence's comments here.

F

B

So, who wants to try precision warrants Michael? You want me to do it I'm happy for one of you to do it's.

A

Probably easier for you to point at your screen and tell us what we're talking about um for the moment. um uh I went through them about 40 minutes ago, or something like this and I guess my comments about most of the comments. I was like I. Don't I can't really see. uh I can't really see one way, the other, that it makes sense. I mean I, don't care right.

A

You know like so, for instance, here, where am I I'm in the comments near the top I guess we had some comment about. You changed the entity back to it, because I think it was clear now hold.

B

On you talked about somewhere in at the top okay, here's, the top am.

A

I, looking at the right thing, eighty six files changed whenever files changed instead of yeah, we hit files changed. Instead, if I was looking at diff.

D

Maybe we maybe.

A

We maybe I'm looking at the wrong detail level of detail here. Sorry, the point is I couldn't I couldn't reading it. I was like you know what I can't really feel any different one way or the other, so I don't have any objection to the new taxes at this point. I didn't really feel any any any attachment to it rolls I, yeah sure I.

A

Feel like we beat it to death at this point and I, don't have any further opinion.

B

Kind of spot the difference here. This is a case where yo there's some line breaks added, which is you know, thank you for breaking long lines into multiple line like this first one here. It shows up as a diff I can't help. There's a word change or not. I'm I, don't think it's a change. I think it's just a line. Great know: Harrison works, changed I.

A

Always like to put those as separate commits even push the line, the whitespace change into the master and then there, but when we do files changed, we see that merged together. We're.

B

Gonna, either way because, like here there's two sentences in the old line not find 94 and the wording change was done to the second sentence. The first sentence is engaged.

A

Yeah so I said I don't feel I I'm at the like, looks prettier to me. Go ahead. I can't see any I, don't see any objection to it. Anything here at this point.

B

A

Has some suggestion I'm gonna reread it again, which.

B

I can't do easily while driving the screen so.

A

Lords, do you want to comment on some of what you commented on.

F

Yeah, that's fine, I mean.

F

F

And was on evidence the definition of evidence. The current definition of evidence just says: it's a set of information.

B

A

A

B

Okay found it go ahead, scroll.

F

Down so you can see more.

A

So you can see what.

A

You need to go. It's further up is common or up.

B

138 right, which is right.

F

B

To the end of this second comment, you.

F

Got it so I mean to me trying to be a lot more specific about these terms and detailed and show how they relate, because there's a lot of information and helpful in understanding how they relate to each other.

B

Is actually incorrect, I think.

E

B

The information itself is not necessarily about the trust or administers its information about it, but whether it's actually something that it's when you say it's a search. The trustworthiness yeah.

F

Dave I'm, sorry, and let me let me talk here a little bit um I'm proposing rewriting the whole thing.

F

So if you read the comment and and there's some suggest, I suggest some text there so get rid of the term. Trustworthy refer to the term claim say how it relates to you know where it comes from and where it goes.

F

Suggesting it's a set of claims because we need to mention the word claim. That's really important in to finding evidence, because evidence is made up of claims. So definition of evidence without the term claim is, is kind of a very poor definition.

B

In comments once you're done soon,.

F

Okay, so you see my proposed wording there, it's just it's a set of claims, not information that claims that characterized a target. So that's what so. You know that that evidence is about a target and then at the evidence, is conveyed from the ax tester to the verifier. So you kind of know what the processing is of it or you know where it starts and where it goes, and then what happens to it when it gets to the verifier.

F

That's my proposed.

E

I'm gonna do with with changing information to claims. I think this proposed change is actually trying to add semantics to the the the use of the word information, because it isn't clear what information, what type of it you know what the information is, but, as Dave suggests, trying to ascribe trustworthiness status to few claims is implies a you know, some there's some appraisal, that's being done by the tester. Are they testing environment, which is not technically correct, but Lawrence is an agreement on.

B

That part right, absolutely yeah, so I think Lawrence and I, and you all agree that the text here that adds trustworthiness status is more trouble than it's worth and to avoid that I I think we but I disagree.

E

By changing the word information to claims, you accomplish the same goals. What I was after.

F

C

Again, let me interject remove trust with me. That's fine! You have to insert something else. Otherwise, a claim set, there's no different from evidence and there's a difference and what what quality do we add to evidence that distinguishes it from a claims set? It's very important, I think, and that is missing.

C

B

Why it's important yeah.

C

It is important because we can eliminate it from evidence. If it's not different, then it's all claim sets.

B

C

No exactly no, but so there's a difference, and that is important. The.

B

Difference is who the endpoints are. I would say that any any claim set that goes between a tester and a verifier is called evidence. Any claim set that goes between a verifier and relying party is called attestation result. Yes,.

C

I think very wrong. Sorry, I think this is not no. It's on. This is just Hank emitting the opinion that this is very wrong. Evidence has a certain specific quality that think of it like a CIS. So if you do not collect it correctly back it correctly and take it correctly, the the the forensic will never be able to use it. It will be not variable in court. That is evidence.

C

So, if you're not using an interesting environment, can that somehow has trust invested into it and as capable and and responsible and trusted to do this job? It will not be evidence, it will just be a simple claim set with no level of insurance. The elevating part here is the attesting environment is doing it. That is the one is creating the evidence on the adjuster side and therefore enhances the level of assurance. If we drop that remote attestation procedures in principle make no sense so.

F

You don't have to put all that paragraph in the definition of evidence you can describe it elsewhere. Yes,.

C

That is correct, but we have to have a linking word. That is maybe not trustworthiness. I was using trustworthiness for that exact context. You could.

B

You can take that argument to say you should never put a GPS in there if you can't personally vote for the GPS value.

C

The claim value again.

B

Put it into the definition is not the right place to do it, but I am listening. No.

C

No, no I think it's very fundamental to highlight that a claims set it's different from evidence and no not dancing around that. You are typically assigned that stuff at some very good, protected secret or a secret. That is very known to be some variants. Okay, you know not stolen and not compromised and stuff. So that's the reality yeah.

C

So we're dancing around all these terms, not talking about root of trust, not talking about shielding of secrets, not everything, but there has to remain a simple, fundamental exclamation that this is different from just a set of claims. Otherwise you would use theses. Yes, all the time.

B

C

B

Let me ask you a question: if, regardless of whether we're talking about the old wording or the new wording, where it says a set of information or a set of claims, would your concern be addressed or not addressed? If it said a set of signed information or a set of signed claims? Is that sufficient? No.

C

It's too specific: it should be because, yes,.

B

C

Could put it in a very secure channel and with the signature, which is an ongoing discussion and therefore it's not signed instead evidence. He claims.

E

Asserted by an a testing environment- yes,.

F

B

You here we have to define it without defining the details of the environment. That's in a waiter section. If.

C

You omit certain degree of I, don't know you can believe this but polling a trustworthiness. You can believe this due to some measures. I, don't know if you omit this from the term evidence you're breaking the whole thing, maybe I'm reaching out to Paul.

E

C

The squad- it's not.

C

E

It's really a style, it's a question of style where this is or the section is early in the document you're trying to give the readers. You know, introduction introductory level understanding of what all this stuff is and the, but you don't want to overwhelm them with you know, with court proceedings or legal briefs or whatever you know this, it's intentionally vague because we're going to happen, we're going to add you know the details later. I agree with that net yeah.

C

And if this text can be interpreted scientist or to specific, if you do not find text that distinguishes claims, that's from evidence, we strike evidence from the terminology period. There is no way that I will I.

E

Think we're in agreement we're in agreement with I, think you're. The wrong thing, I think I. Think the I think that there's general agreement that you have to have a crisp understanding of and definition of evidence, but I think that right now is in the document is the wrong place. To put it yes,.

C

This is the early on stuff and if I cannot distinguish a difference between evidence and claim set without using the term sign, but just specific, it's not useful and I would remove the complete definition because it is useless. It does not add anything if you're not adding anything.

C

Can I design site is too specific signed is wrong and I think Thomas would agree with this. That scient is a too strong of a statement here, because it's sometimes it's not signed. It's simply said not correct. I. Think.

B

Even in the text that wasn't the word about the implicit that the channel stuff I forget even in that stuff, it may still be sign, but it's signed by the channel by the channel encapsulation, not by the yeah, not by.

C

E

So I took I took one assignment to go by wording for that idea. But it's later in the document.

E

C

Don't understand how.

A

This this came up from this one change here, that's what I I guess it was because we had a conversation about about attestation results not being evidenced right. The.

F

The the the public West refines that a lot of the definition so I went after what I thought should be fine for the definition and and and my I mean my proposal- is actually not either of the text that's showing there. But the text in the comment later below, which is summary right right, so I I've, taken.

A

Your I took the text and I put it in a suggestion: I'm surprised that didn't refresh on your screen already, but maybe you have to explicitly reload okay, but on.

B

Lawrence's I have no preference on this one being I am fine with your proposal here. I agree with this one, certainly, and this one I disagree with that's all I wanted to say before, but if you want me to refresh second refresh here, do you agree with Dave defining a new term I hate, defining new terms and I think that I think actually confuses me more than not having the term slide up a bit yeah, so yeah I'm trying to find the right spot here. We.

E

Don't introduce the terminology around testing environment and target environment till later in the document this.

B

Is what um you show Michael yeah.

A

So I just put put the text that that Laurence suggested into an actual suggestion. So so what the question I had about that thing was whether the word attestation before evidence was meaningful.

B

No, it should not be mean you're right, that's right, yeah, okay, so I think I can edit.

A

B

Pairing of words somewhere else, maybe in a text they rewrote and I really.

C

Dave you're right education should be not capitalized here that is artifact I've.

A

Just updated the suggestion removed that word.

C

Sorry that I'm a hardliner on the evidence stuff, but literally we cannot use the term signed or trustworthy, find some.

A

It's not in this, it's not in this line. So let's not worry about it. There.

C

A

Let's focus on this line because that we got to make small steps forward here right, so do we have objections to this suggested change that resolves this particular point here.

B

Meaning this term right here, yeah.

A

Okay, so you object to the attestation target week there. Okay just wanted to be clear. Thank.

E

You yeah, and that isn't even the term that we use later in the document term,.

A

B

I, don't know what you mean by this I, don't even know what the intent was, because you know we said there can be multiple target environments right, just like in the whole point of layered attestation. The evidence is the collection of all of them.

B

That's why there's the parentheses s on here, presumably but I, don't think in the text here in the definitions, we should define evidence in a way that has to use terms about the internal implementation about you know the target environment of tested in the testing environment or whatever it is in a much later section. I want a definition here that is more there's more abstractly r-spec how things are implemented internally. It's.

E

You need something like it characterizes the behavior or something the.

B

Old takes which I did not okay anyway, the old text was that somehow was about the trustworthiness status of an ax tester, which is better I, don't like that either, but I like in fact better than this.

F

But evidence is not about an a tester. It's your.

B

Steers the entire box of stuff and.

E

B

On in the later sections, the.

E

Tester includes.

B

A role that seen from the outside stirs.

F

The ax testing environment plus the attesting target, as.

B

Used in the diagrams later, yes, it's the entire, you know like or entity or whatever term that you want to use when you're talking about it. So.

C

The question was I think to rephrase it is the testing environment and the target environment. Part of the tester. Yes,.

E

Are the claims about the ax tester, a testing environment could be a question to ask and I think we get a divided answer, because in in some cases the ax testing environments claims are asserted by endorser.

E

Some argue that they could be self asserted, in which case okay, maybe there's.

B

Things that come from the ax tester, which is the evidence- and there are things that come from an endorser which is the endorsements yeah. So here the you know a set of claims created by the ax tester, all good. So far, if you ended it with a period there, I would be happy. Hank would be unhappy, but I would be.

B

Second sentence is perfectly fine, because that's basically just a rephrasing of this with no changing meeting I have no problems with it. It's only this phrase right here is the problematic phrase.

A

Where what do we, what I, don't still don't know what you would prefer so B to make Hank unhappy, but I don't know how to make but what it? But what? What can.

C

A

Suggestion so we can find out how Hank is.

B

I'm not trying to make you unhappy I'm saying my preference is the one that would make you unhappy sorry and.

A

We can we rest of us understand that that suggestion better. Can you edit that suggestion to put your preferred wording in maybe Hank.

C

A

Unhappy, as is you think,.

C

Removing that I'm also against the last one you.

A

Just put a period there: okay, yeah.

C

A conceptual message is yes, of course, we've defined. We give it more more indication why? Where is it created and who is the ultimate consumer, but that is not defining how the message looks like, which is my point so I think yes, why? The second sentence is true: ish without relying parties, because we rely stuff. So it's not true in the literal sense, you do not send it to the verify all the time you send it to a relying party.

C

Sometimes, but I can look beyond that, but still it does not get the point of that. The actual payload is different to stand out, claim sets and that's the point and there's different semantics to it, and that is my my Hank.

F

We understand your point and we agree with it. We just want to put the text that says what that is somewhere else. I.

C

Understand if it's not in the terminology, it's a malformed terminology and therefore it should be scrubbed, because it's an undefined term no.

A

I I, don't agree, the terminology needs to introduce the terms and the terminology could say you know a set of claims, as described in Section. Four right could.

D

A

Have twelve.

D

A

Know twelve pages of explanation: okay,.

E

We were writing the dictionary, then Hank would be correct, but we're not right and.

A

We're not writing a dictionary. Yes, exactly so.

C

Yessiree and have a reference and defining section.

A

C

A

The point is the point: is this to say here's some terms and we're using them specially. So when you see them, don't go, get your dictionary. Go look up, go use our definition so that people know what's going on and that's what we were also chose to capitalize them.

C

Widely missing the point of this discussion and I'm sorry for sneaking 50 minutes of your time. If that is your point, that you can just put it right. Yes, if.

A

If that I don't know where that goes, but if that's necessary, then let's write argued about that section, not about the definition where it goes here. uh What what I think is that we have this? This request has reached the point of diminishing returns and I have.

B

I, don't know whether this is a gonna make things be really easy to resolve, or it's gonna make things harder, so feel free to say: let's not go there. You can see the original text here did not talk about claim sets.

B

It sounds like his that Hank your primary objectives main making sure that people aren't confusing a regular claim set with a the type of claim set that becomes evidence for whatever the type is you're transferring this or whatever the original one did not use the word claims, and so, if there's a set of information about, you have the same objection to the original stuff because it doesn't use the word claim set.

B

Now you can see Laura and I said I, don't have a strong preference either way about information versus claims here, I can go either way, I just don't know whether it would make you feel any better. If this said a set of evidence create a set of information created by an ax, tester, that's conveyed, and so on. Would that make you less unhappy here? If the word claim wasn't used in the definition, would you still think it says I'm heavy I think.

C

It claims good because we are using claims that a different men are already here. It's very fine I.

B

Just wanted a sequel is easier to make the thing go away by not using the word claims there and different claims until later, but it sounds like it doesn't help so backe, Michael, okay,.

B

A

B

You were trying to say something.

B

Ding returns: are you there Michael.

A

Not as strong as Hank has, but.

A

So I can't can we walk through the rest of the these things quickly in the last five minutes and see if what other big well actually what things we can agree on and what things if we can't agree on them? Let's just move on to the next thing.

A

They're mostly Lawrence says comments at this point.

F

Okay, so you're, looking at the definition of relying party there.

F

So I don't think a relying party is typically an application I. Think of it, as you know, legal entity or a service like an enterprise, a network or a bank.

B

What do you mean by that? Because evidence has to be parsed by a computer and a bank is not a computer, we're.

F

Talking about the definition of a relying party, yes,.

B

And in this context, it's enough with me I agree that it is an application. I mean the typical application I'm trying in the context of how relying car D is used in this document. How is it not an application but I'm trying to figure out I can understand their baby. Other uses of relying party might refer to a bank, but as we're using it here and 4949 I think it is an application. I mean.

F

I'm going for the this is a high-level architecture, die a discussion and trying to characterize the consumer of the whole, the whole attestation operation. So it's not about the software. That's used, like you know. Ten different different parties could use the same application software, but they might be but they're different, relying parties and then one I.

B

Will agree with you on is that it's the discussion of whether it's an application is something that we should not need to get into now. Maybe somebody will think it's necessary, but I don't think it's necessary, and so, if all you're trying to say is it's possible to write this without using the word application and here's your how about this then I'm willing to go along with that I. Don't think it that the green text here was incorrect in any way.

B

But if you're saying well that just begs questions that we'd better not get in to you I'm fine to go along with that, and you can go on and talk about your how about this? This wasn't my text in the green, so.

E

Ceases to provide some intuition about what this sentence you know, how does this relate to the outside.

B

World it's like a verifier is typically an application to write for some definition of application. Point.

A

Here is that it's not a human, it's a piece of software that is going to.

E

A

Legal entity so.

B

Greater is a human, that's. What I'm saying is I agree. That is not it's. It's not.

A

A human because the human doesn't doesn't do this, doesn't look at the signature, we don't calculate the the validate the signature ourself, we look at it. We have an application, do that and so the the application speaks to me. I, don't do the work for it.

F

I'm trying to go for a lot more high-level sense making, then just what software connects to the this software connects to that software connects to that software I'm, trying to really describe the much more broad perspective. Certain.

A

Saying a relying party is an organization or or legal entity like an enterprise, etc. That does something and so but I, but I think that that most other people are saying, but it's a specific thing that validates the attestation results and and appraises it using the appraisal policy. Well,.

F

It makes the end decision about I mean it's I. Think it's really important, because the relying party is the end consumer of the attestation. It decides whether the the the guy is going to be led on the network or the transactions going to be approved, and that that seems just really important in trying to understand what attestation is all about. The.

E

Intent of using the term application is to provide some intuition that it's it's.

E

It could be anything that that is performed in information processing, in other words, if you're familiar with terminology used in, like you know, information processing, data processing database management, they talked about an application being as an abstract term, for it could be anything. Maybe that doesn't translate. I, don't know, but it's it's.

F

It doesn't I mean I, don't I'm, not really a so of so concerned about the term application I'm just trying to go for a much broader, larger view of what a relying party is.

F

So someone coming at it coming at this trying to make sense of what out of station is all about. Has it has? It has some and you can add to the.

E

List I suppose I mean it's just the either either you you get specific about you. The question is: how much intuition does the reader need right, and so maybe they don't need any intuition. You just leave it very vague and they can infer themselves what it means, or we add some. We add some clarity to it. So.

F

I proposed some text there, a couple of sentences that trying to say what I'm saying.

E

This exemplary right he's not trying to limit it to that.

B

Right now, Lawrence's definition is anything that consumes. The attestation result is called a relying party, regardless of what you use it for I. Think that's the gist of it is even if you don't for an authorization decision, if you purely use it for logging. If all you do, is you take the information and you store it? He says: that's still, a relying party I'm, not sure, I, agree that definition I'm, not sure I disagree with it either I'm still thinking about it. I think the.

F

Point I mean I. Think the one of the key points here is the whole attestation process. Everything we're talking about is in service of the relying party they're, the they're, the ones that right so so.

E

Originally yeah, so if we removed the parenthetical, the remaining text says what you just said in one sentence except.

B

For the application word here appears again down here. You.

F

Don't say half of that, it's like.

B

E

The point of using the word application below is to say that there is some there's some actions that are not defined here.

E

In other words, the okay, so.

D

To me, here's what you're saying has.

E

No actual value that has to be integrated with something else and if we can't use the word application then come up with some other word. But we use the word application all the time in computing and everybody seems to know what it means so.

F

Take take it if you want application, just put it in parenthesis and somewhere somewhere in there and in my text and that's fine I, don't really care about that.

F

That's so much I'm, just trying to find the larger definition. So here's an analogy I mean you can write. A definition of a plane is an object that flies well. Okay.

F

Well, what I'm, trying to say is a plane is an object that flies that carries people from point A to point B, because that's where they want to go versus like a plane is an object that just you know as for entertainment, with a paper plane, or you know something like that, it's I'm trying to scrap the point of all of this.

E

We're trying to introduce the reader to you, know general concepts and we're not trying to have a dictionary definition right conversation as we had before. I have.

B

Two questions: one is it's after the hour right now: how long do we want to go and the other one is I, don't understand what was wrong with their red text, so we're trying to fix that I'd like to understand but I'm happy with understanding that outside the context of this meeting, if people want to wrap up.

E

The problem I had with the previous text is it: it uses it's it's providing intuition around the idea of providing an authorization which is too narrow, so basically that that intuition moved to the beginning into the parentheses and used the more general term application, as opposed to the specific application of authorization. Ned.

B

If I understand your answer, that changing the text as I will read, it now would have accomplished the same purpose. At least out of the partner answered, an entity that depends on the validity of information about a tester period.

B

Compared relying party that would had the same done the same fixes because you were worried about you- wanted to replace another entity, be more specific and you say the part was problematic for you. Was this part right here? Yeah.

E

If you want to just delete that, that's fine with me, I'm.

B

Just gonna capture that as another proposal down to here for her because otherwise I'm gonna forget it by the time we get to next.

F

A

F

That the wrong party is the end consumer and of all of the this in the definition, then it needs to be somewhere else.

C

Why not call it also claims here we are falling back to a information I.

B

Think I didn't think what I just copy that okay.

C

B

C

A

B

I'm just trying to wrap up because it's five minutes after now so I don't want us to argue about things if we can find. But if we don't know what the problem is were trying to solve so Thank You Ned for clarifying what problem you're kind of solved, I.

B

Don't want us to do useless black shadow, a nice to do useful, but.

B

I may not want that. My bike to be plaid or something so.

C

Bumpy I want to be comfortable so.

B

Do we want to try to continue this next time or what I want to do now over time? So I have.

C

To drop suddenly sorry, it was a good discussion, though.

B

I have put in here the things that sorry up here you can see. I said one discussion point or whatever I think Hank. Your point is not or if you'd like to enter that enter your comment yourself feel free, but I was trying to inter comment. So what I was hearing yeah discussion during me? This is what I was hearing from more than one person. If it's just one person go ahead and enter your coming or so, for if I missed it or whatever go ahead so.

C

So I know that we are blocking with one person here so I'm this time. The block are so but I think we can work around that with references or other things to edit to it. That's fine, so.

B

I am going to stop sharing my screen because I would like us to wrap up soon. Okay and if I remove the text, nobody can point at it.

B

They can read it off of their own screen, but then it's not me, that's doing it. So, okay,.

A

So we are going to come back to this next time, and uh this is probably the third hour we've spent on this poll request.

A

So Dave said: let's not bike shed too badly on this, but it seems like there is some useful work that we're doing, but I would like to maybe start please if you could Hank open to more about freshness, and maybe those are easier to close time. Considerations, 86 and 87 I haven't looked at them yet because.

B

They're only an hour old I haven't been there, I think I opened one and you know like in the five minutes before the meeting, and so maybe we start with those next time because I don't know how long this one is and.

A

82 as well I, which I opened based upon another number, 60 and I thought were pretty much done. Ned said, plus.

A

182 last I checked: it had unaddressed comments as unaddressed comments, so I'll go through them, but I think that they're minor and so, if, if you're happy with them online, then we'll pull it. Okay, 82.

B

Was one of those where Ned had the ax item because everybody liked his articulation and he was an open. The new pro requests, then I, don't know how much of this 82 would be replaced by that one, though.

B

You'd overlapped with 86 I think I see me I'm looking at 82, where it says you know I like Nets articulation, maybe the PR should just say. What's in Ned's principle and Gary said you know Ned when you take the lead on the PR revision, so I think he owns the token. Are they you two now I agree.

A

With you, Ned Ned yeah.

E

Just haven't had time so.

A

We were that was Jerry's, pull request that that we reopened, so we could edit it yeah. It's.

B

Not like a we have plenty of other things to talk about. Apparently, yes,.

A

Okay, all right talk to you guys next week.

F