Meeting minutes
<ottomorac> transcriber-bot-l, connect
<ottomorac> transcriber-bot-l, connect
Agenda
Otto Mora: Okay, um, so...
… Today, we are going to talk about the working group charter status, and um
… associated charter status, for also the did methods working group, and then also just briefly mentioned anything
… any plans specific to Tpac was hoping that Mano would be here for that. But I don't see him yet. Then we also have a proposal related to the did rubric that we wanna
… Uh, run by the group, where we wanna just request horizontal review for it
… And then we have the main topic, which is this did URL dereferencing conversation and a couple of proposals that we wanna pass in relation to that
… So is there anything else that folks want to discuss?
WG Charter Status and TPAC
Otto Mora: No. I'm seeing any… Okay? So 1st topic working group charter status and Tpac, so I'll let Pierre. Uh, update us on
Pierre-Antoine Champin: Yes, thank you. So we have announced the charter refinement period...
Otto Mora: In that regard...
Pierre-Antoine Champin: Uh, for the new charter of this working group at the beginning of the summer, we, uh, haven't received much feedback, but, uh, to be fair, the charter is just, uh...
… a continuation of what we are doing right now. Uh, so I, uh, I think we are, we are in a position to send
<JoeAndrieu> can we get a URL for the charter?
Pierre-Antoine Champin: the charter to AC review ASAP now
… Uh, and therefore, the good thing is, uh, that we… the AIC review should be over by TPAC, so if
<ottomorac> w3c/
Pierre-Antoine Champin: if it so happens that we get formal objections, then we can try to resolve them, and we get an opportunity to talk with the objectors during TPAC, which is often a… uh… a… a
… has better chances, uh, maybe to, to find, uh, to find agreements. It's easier face-to-face. So, yep, we, we are on track, uh, on this
Otto Mora: Yes, okay, so that's the DID Working Group Charter, and I think that Manu was also inquiring about DID Methods Charter. That one may be a little more… um...
Pierre-Antoine Champin: Indeed. No, no. Well, this one, we did announce the refinement period. Well, in fact, we were in refinement period since a long time ago. we...
… the charter has not evolved a lot in the meantime, but we finally found chairs, as I announced in a previous meeting. So, for this charter as well, we are ready to go to EC review, so
… Uh, so again, yeah, same, uh, same thing as, uh, as with the DID working group. We should be able to complete AIC review before TPAC, and again, in case of objections, that gives us, uh, an opportunity to, to, to resolve that, well, ideally before TPAC, and possibly. During feedback, face-to-face, if that's, uh, if we're not done yet
<Zakim> JoeAndrieu, you wanted to ask about CID spec
Otto Mora: Right on. Thank you, Pierre. Appreciate that. Uh, Joe?...
Joe Andrieu: Yeah I wanted to raise an idea that I I know we haven't been formally considering but I think we should, which is the CID spec which dids depend on is under another working group...
… And I think we should talk about whether or not we should take it over so that we have the full stack of what DIDs depend on as opposed to having it in a group that isn't advancing it because it's not particularly related to VCs
<Zakim> manu, you wanted to support Joe's proposal.
Joe Andrieu: Uh, because I think this might be an opportunity to take over maintenance of that spec. Um, but I'm not sure what the political dynamics are, how that normally happens of transferring one spec to another group, etc. So, let me just open that as a possible conversation to have
Otto Mora: Thank you. I see you might not want to also address that...
Manu Sporny: Yeah, plus one to that. I think that would be a good move. It really doesn't belong in the VC working group, speaking as one of the editors of that specification. And since we're rechartering, now is a great time to add it to the charter...
… Uh, I'll note that Michael Shea is, you know, the… basically, he is editing the spec actively, right? But there's no changes expected to be made for it, and really, because the DID spec depends very heavily and is built on top of the SID spec. Uh, plus one to what Joe was saying, we should move it into the charter of this group. That's it
Otto Mora: Okay. Up here...
Pierre-Antoine Champin: Yeah, I… plus one, that's a very good point. It makes much more… it would make much more sense that way. My only concern is that...
… and just testing the water here, what is our position if that means delaying the AIC review to not
… And before TPAC, are we willing to take the risk of that? I'm really just asking you the question. But
Otto Mora: I don't know...
Manu Sporny: Yeah, I mean, yeah, yes, like, it's the right thing to do. Um, I don't know why it would push off the review too much, PA, you've probably got an idea in your head why. Like, I mean, it feels like it's a, you know, uh, the PR could be raised today, we could have it merged by next week, and… That's that, right?...
<JoeAndrieu> +1 for getting this taken care of, hopefully without too much delay
Pierre-Antoine Champin: No, I don't have, sorry, jumping the queue, but I don't have any particular idea in my mind. It's just, yeah, this is something that we haven't mentioned before, so, and it's a...
<pdl-ASU> +1 to taking this from VCDM
Pierre-Antoine Champin: significant change to the charter. But I agree, it can be made quickly and we still have some time before the DAC review would span after TPAC. So, okay, let's do this. uh… Um
… Anyone volunteering to make that PR or I can try to do that
Manu Sporny: I've got some time today, I can do that...
Pierre-Antoine Champin: That would be awesome. Thank you very much...
Otto Mora: Thanks, bye now...
… Okay. Okay, so
… I guess that, uh, clarifies regarding the charter, and you said you wanted to discuss
… maybe something in relation to TPAC, Manu, or does that, uh, what we just discussed, uh, satisfy the
Manu Sporny: Sorry, PA, so the...
… I'm having to read the backlog. Um, the DID Methods Working Group Charter will be ideally put out to a vote before TPAC? Is that… What I heard. Excellent. Okay, that's it
… Great. Awesome. Thank you
Otto Mora: Thank you. Uh, Will?...
Will Abramson: Yep...
Pierre-Antoine Champin: Both charters, yes, yes. Currently, we are in a good shape to achieve that for both...
Will Abramson: I just wanted to talk briefly about TPAC. I mean, we have, uh, Tuesday of TPAC will be the DidWork...
… did working group day. Um, so, I mean, this is really just… I think next week I'll put together, like, a spreadsheet so you can sign up, and it'd be great to know who's coming, like, it'd be great to have you there. And then also, if anybody has topics that they think. You know you want to have face time for like it would be great to hear
from you. That's it
Otto Mora: Yep...
… Mano?
Manu Sporny: I suggest we put aside time for the charter discussion, both charter discussions, you know, maybe time on Monday for did methods...
… uh, chart discussion, and then Tuesday, um
… for the DID working group, if we need to, right? I mean, I'm just presuming there'll be things to talk about. Maybe, um
Otto Mora: Mm-hmm...
Will Abramson: Yeah, that's good...
Otto Mora: Okay, perfect. Uh, before we close up this topic, uh, there was, uh, this...
… bit here with the folks from solid. I honestly, I didn't have a whole lot of background. It seems very interesting. I guess it's like a portable data storage of some kind, some kind of standard around that
… But, um, there's this session on, um, identity on the web. Um… So, it seems interesting, they want to address identity. I don't know if one of us specifically wants to represent
<ottomorac> w3c-cg/
Otto Mora: the group there? Or do we just table it, for you know, next week we can discuss it in more detail. But just wanted to flag that case
… Also shared a link to that issue here. So folks take a look. I know this is kind of last minute, but I wanted to flag it if there's any
Joe Andrieu: What...
Otto Mora: I'm interested in that. Yes, Manuel...
… Yes
Manu Sporny: Lots of thoughts. Can we delay talking about that? Because it's not a simple discussion, I don't think...
Otto Mora: Yeah. Yeah. Well, okay. Fine. Yeah. Yeah. I'll I'll save it for next week. It's fine. I just wanted to make sure we're all aware that they'...
… They're proposing this. Perfect. Okay
… Um, okay, perfect. Okay, um
DID Rubric proposal
Otto Mora: The next topic
… is the… the rubric proposal, so
… As as you folks heard in the last meeting. there was an interest in getting the did rubric through a horizontal review. And let me also try to find the link to the right. Hope folks have it here. Oh
Will Abramson: Yeah, actually...
<Wip> w3c/
Otto Mora: I have a little spec...
Will Abramson: I can get it...
Otto Mora: Okay, thanks. Yeah, I'm just trying to… just fumbling here with the...
… the spec itself. And then where's the repo for it?
… Okay, thanks. Yeah, thanks. So yeah, so the the rubric. Yes, Will and and Joe have been working on on making some updates. And and now we would like to get a a review on it. So
… The proposal, let me promote the proposal. is to, uh
… Request a horizontal review for the rubric as part of the transition to a W3C registry
… Does somebody have any comments before we run the proposal by the group?
… Uh, yes. I see… okay, first Mano, and then Joe
Manu Sporny: Yeah, I'm a bit, uh… unprepared for this. I don't feel like we've talked about the updates, or what's happening here in the group. I'm supportive of moving this work forward, but I don't know if we've… Got enough conversation happening in the group to request a horizontal review at this point. Like, I don't… unless I missed
something, maybe I missed, like, a couple of meetings where we talked about all the updates and what we're doing with this stuff in detail?...
Otto Mora: Yeah, uh… I'll let Joe speak...
Manu Sporny: I'd be, just to be clear, I'd be a minus one at this point because this is taking me by surprise. We haven't talked about this...
Otto Mora: Sure, sure, sure. Uh, yeah. I'll let Joe speak to that...
Joe Andrieu: Yeah, so you're half right, Manu. And so if we want to delay a bit, you know, I wouldn't be offended. We did have another call where we presented this work and we asked for feedback, but we did not take a call where we walked through sort of section by section showing what the changes were...
Otto Mora: Okay, bye...
Joe Andrieu: Um, and so if that's what you would like to do, I wouldn't be opposed to that...
Manu Sporny: Yeah, Joe, I appreciate that, and I know I'm just one person. I mean, if everybody else is like, no, I definitely know what's going on here, and we're ready to ask for horizontal review, like, that's… that… that would be fine with me. Um, I would prefer to be wa… you know, if you could wa… if… if the folks that have been
working on this can walk the group through...
… all of the changes, I think it would help us feel more comfortable to say, like, yes, this is something the group is getting behind, and we want to push it forward as a registry, and so on and so forth
Otto Mora: Yep, probably doesn't hurt to have a more detailed walkthrough. No problem. Okay. Um...
DID URL De-referencing Discussion
w3c/did-resolution#344
Otto Mora: Okay, so next topic is, uh, TDRL dereferencing discussion, uh, and this specifically relates to… PR 3, 4, 4. So let me just. subtopic… that
<danpape> +1 for rubric walk-through
Otto Mora: Yes, so I know that in that Pr. There's specifically a a
<ottomorac> w3c/
Otto Mora: A set of detailed comments also from Marcus. um
… which I'll link there. Um, I wanted to, uh, do the following, so… We will, um
… we will run a couple of resolutions, but before we run the resolutions, um, I wanted to allow each party to
… Um… kind of voice… voice their, um
… their side. Right? So like we, we just try to please keep it short 5 min, Max Max each, and please try to keep it civil. So that's that's kind of what I wanted to do. So first, st I'll let Marcus his concerns to Pr. 3, 4, 4
… Then I'll let Joe speak his side, and then we will run the the 2 resolutions related to to this. If that's. If that's OK with folks. Maybe, Marcus, do you wanna… Sorry
Markus Sabadello: Uh, yes, I will… I will try. You're… you're right that, uh, yesterday I...
… posted another long comment, just summarizing the concerns with this Pr. I think everybody knows that I'm not very happy with this. I don't know how much time, or how many, how many details you want me to. to mention now, I think on… On one hand
… I'm not happy with this, because a lot of explanations, a lot of rationale that I've heard
… for this PR, I think, don't make much sense. Uh, we've… we've heard things like the current algorithm, uh, requires the references to be a separate component or remote endpoints
… or things like that, uh… so… I've heard a lot of statements that I think have no… are not really
… backed by what's in the specification right now, right? Explanations why we need the new referencing algorithm that I think are not valid from my point of view
… One example is the removal of the dereference function, right? This PR removes the dereference function that we've had since the DID 1.0. Recommendation
… Uh, Joe said at some point that there was a resolution in the Deep Working Group to remove that. I couldn't find that. I found… discussion in a Google Doc, where
… Uh, two people were in favor of removing this function, and four were against removing that
… Um
… this is one example where I think there's been a lot of miscommunication. All that this function does is document and express the inputs and outputs of an algorithm. It doesn't do anything else, so… I disagree with removing that. Um
… I have a section also in my
… comment about the things that I find really good about the PR. So I, as I said before, I like the idea of the handling strategies, the modularity, the extensibility. I like, uh, statements about selecting resolvers, um
… I like how it clarifies some terms, such as clients and dereferences, so there are some parts. that I think are good here
… The main problem with the PR is, I think, is that it's yet again one of these mega pull requests that change a lot of things at the same time
… found, I think, 11 breaking changes, uh, in the PR, 11
… things that the PR changes from… compared to what we have now in the algorithm
… some of which I think have not been discussed at all. I mean, I… I see a removal of the
… of processing the service type parameter. I see removing processing of the verification relationship option and the accept option
… These are topics that the working group here has discussed at some point in a lot of detail, and has agreed to
… to have, and in this PR. Things are changed, and
… Sometimes removed, without us having had a discussion about the specific
… changes. One of the prominent breaking changes is removal of processing the relative ref
… parameter. We've talked about this for a long time. Um, I've… Heard from several people in the working group that, uh
… The path handling approach
… is a complete replacement for relative ref. I then did a lot of experiments and concrete examples to, I think, disprove
… That claim, um
… I think I may have convinced some people that we still need the relative ref processing, and
… Other people are ignoring the statements and technical arguments
… And so this is what I see here. A lot of changes, breaking changes
… That, uh, some of which have been discussed, others have not been
… Discussed, and uh
… And, uh, we said before that these… these huge pull requests that change so much at one… one time. Really hard to
<manu> To be clear -- some of us are not "ignoring" those statements, we just don't agree with Markus' positions on many of those statements (and have discussed these things at depth).
Markus Sabadello: to decide on and to discuss. It would be… like, it would be really cool if we had a pull request that. Just as the
… general structure and the handling strategies without all the other stuff in it. Um
… And… Um… Sorry, one moment
Otto Mora: Mm-hmm...
Markus Sabadello: And, uh, yeah, it's...
… then had a few recommendations at the end, what I think should be done
… Um, which is, uh, yeah, to break these things apart. This is something we've said several times months ago, that we shouldn't have these big pull requests to change everything at the same time, often without… Real explanation, um, and instead we
… We should cover specific changes, one after the other. And, uh… And, uh, yeah, that's it. I
… I also expressed concern about the charter violation, so I want to talk about that a little bit as well
… Um, a pull request like this, if it is merged, it would make it 1.0 conformant
… The referencer's not conformant anymore. With this specification
… Which, in general, would be fine, right? Because, as others have pointed out, uh, bit resolution has never been a recommendation, so we can do whatever we want
… We can also make class 3 changes to the core. Uh, which means, uh, changing
… conformance in… in DidCore
… So in theory, we can make breaking changes
… I do not think we should make these kinds of… of, uh
… breaking changes, right? A lot of breaking changes compared to existing… the existing standard and existing implementations, so I don't see a need for… for that. And I still believe there may be a charter violation in here. Because… I think in order to… Um, in order to accept, and
… and make the changes that are in this PR, I think that would imply that we also need to make certain changes in Deed 1.1 that may be considered Class 4 changes, right? Because. Deep resolution is… It's loosely based on Bitcore
… Uh, you cannot implement the DIT resolution without implementing DIT core, and if we make such drastic changes in the dereferencing algorithm, like removing the interface definition, changing the inputs and outputs
… then I think either we would have to make class 4 changes also to bit 1.1
… Or certain parts of the resolution will not be
… Uh, compatible with DIT 1.1. Uh, that's my
… Consideration about that as well, but
… But in general, on a high level, to summarize, there are some good things in the PR
… But the communication of why it is needed, or the rationale for this, has, in my opinion, been often… has often been
… not accurate or without a real basis on what's in the spec and what's in the PRs
… And it just changes too many things at the same time. Breaking changes that have not been
… sufficiently discussed that I think, uh, have not been
… Resolved, and it reminds me a little bit of the situation that we had a few months ago, when. There was a proposal to change the input
… of the resolve function to… to TTRLs, uh
… Which didn't really work out either. Yeah, that's it. Thank you
Joe Andrieu: Sure...
Otto Mora: OK. Thank you, Marcus. Joe, you wanted to —...
Joe Andrieu: Um...
… So, I… I do think there's, um
… a need for us to respond formally to all the things that you've raised in this last comment
… We have not had time to do that in the last 14 hours. Um, but I think all of these are actually… they've already been asked and answered, um, Marcus. I think we've been talking about this for a very long time
… I think the assertion that there haven't been valid arguments is
… Challenging. I think many arguments have been put forward. I understand that you have not been convinced, but that doesn't make the concerns unmeritorious. Um
… In particular we talked about service type We had a resolution which I think we could find The other resolution that we thought we'd had is in fact a conversation
<JoeAndrieu> https://
Joe Andrieu: Where we did not get to a resolution, we did a poll. And I will drop that into the lounge
… From the minutes from that. And part of the challenge that the editor's team with the chairs were trying to figure out is how do we have a resolution given the nature of the dynamics where there were changes implemented into the spec by the editor, which did not have consensus of the group
… And so we were trying to figure out, well, how do we have a resolution, um, which, when it doesn't pass, doesn't become a resolution. So there was sort of a semantic issue about how do we, how do we do that dance. Um, there are other several factual inaccuracies here. Um, relative ref is not excluded. Relative ref is absolutely still supported
… What it is not used for is the default service handler. If you have a did URL that has a relative ref, it will be processed in the query step in the algorithm. So I posit that some of this is a misreading of the PR, but I think we've been talking about this for way too long
… Um, I also wanted to talk about the charter on two different levels. Um, one, when you brought up charter markets, you mentioned that this is just too big
… This is the third effort to make this smaller. The reason that this was triggered was twofold. One, I was trying to engage constructively with Stephen Curran's path handling approach
… Um, and as I was trying to help him, it became confusing how we could possibly update the existing algorithm to incorporate what he was looking for and still address my concerns. And as I started getting into that, I was like, wait a minute, there's a lot of things here that are very, very confusing
… The second thread was doing the threat modeling for resolution, in which we start with a data flow diagram of the system. Um, and that was the first time that… and it was an image that I created. This was the first time I saw an image that had a client, a separate dereferencer, and a separate resolver
… And so the fact that the spec had this thing but did not even have an image that described or presented what this architecture was, that's why there is such a big change here because it was so confusing, Marcus, with all due respect, I know you put a lot of work into it
… But the problem is that it confused tag. It confused me. It made technical and security analysis difficult, if not impossible
… So, I think we've been talking about this. I think there are good reasons for it. I think there are still some things that could be improved. Um, but I do believe that we were chartered to figure this out. This is the point of this round of our work, and the first round when we came up with DidCore
… Um, part of our negotiation with the larger community was that we would not be addressing protocols, so we were restricted from doing anything protocol-related, and all we had was sort of this, hey, we'll have this contract with the resolver, and we'll figure it out later. And so this is the figuring out later. So I believe we are entirely
within our charter
… And in fact, because it raised security issues, we also have full authority to update DidCore to bring it into alignment with the work we're doing here and to fix those security problems
… So that's that's my response today. I do want whatever wherever this resolution might go today, Marcus. I do commit to getting a line by line response to all of your things in there. We just haven't had a chance to do this. Is just the timing on your comment yesterday
… That's it
Otto Mora: Thank you, Joe. Appreciate it. Yes, and I will note that. Yeah, we we have had as chair like extensive discussions around this. I I do hope that this is not taken in as a personal matter. But I I think we need to move on...
… And that's why we need to run these proposals. OK, so the proposal that I'm — sorry, I see Will. I just demoted the proposal that I'm going to run, but I'll let Will speak to this as well
Will Abramson: Yeah, great. Thanks. I just wanted to echo something that Manny said yesterday, right?...
… this is not the end, and we can revisit this in the future, but this is about us putting something down and moving forwards, right? So we shouldn't see this as this is it set in stone forever. This is just us trying to get the next step down the line. So… That's all I wanted to add
Otto Mora: Yes, Ted, you also wanted to...
TallTed // Ted (he/him) Thibodeau Jr (OpenLinkSw.com): Yeah, I put a comment into the PR that expresses basically what I'm about to say...
… I understand the desire to move forward
… However, there are a lot of open comments on this PR, not just from Marcus, not just from me. Mine are mostly editorial
… But not all. And they're mostly, I think, made against editorial parts of it, but not all
… And I think those need to be resolved before
<JoeAndrieu> we've been working through this with Stephen's slide deck. Those will all be resolved.
TallTed // Ted (he/him) Thibodeau Jr (OpenLinkSw.com): we can make a proper decision about whether this PR should move forward
… I can't really make a decision except know about merging it without having those things be addressed because the meaning of the PR will change significantly. If they're left in some ways or if they're handled in other ways. That's it
Otto Mora: Yes. Yeah, I don't think we will be merging the PR today. I mean, that's a fairly… it's a good point. Uh, we… we just are making some decisions so that we can move forward. Um, definitely. Uh, Will?...
Will Abramson: Yes, I wanted to echo that, too. I think the PR is not up to date with our latest decision of the group, right? We've been spending weeks and weeks going through, uh...
… well-facilitated discussion from Stephen to talk through the comments that are outstanding on the PR and, like, decide what we want to go forwards with. I think, Joe, there's still some work to be done to get that aligned. What we are trying to do today
… is make some decisions in the group around what is going to be in that PR, like the proposal that Otto just demoted, so we don't keep getting stuck
… going round and round, having the same discussion without any fixed thing. The point of this proposal is so when it's made
… And the comment, and we come back to the same discussion, we can say, no, the group has, has, has
… voiced its opinion on this, please, we are not going to rehash this debate. That's the purpose, so we can actually move forward with refining the PR to a state where it could, like you say
… Uh, TADB be, like, accepted or not
Otto Mora: Yes...
… OK, perfect. Grace, really quickly, before we vote on the resolution, go ahead
Grace Rachmany: Yeah, I'm just gonna say something ignorant 'cause I'm a little ignorant, but what I do know is that a lot of times in the standards that we're working on...
… Less is more like just cutting, you know, like trying to do a really large pr is as difficult as you guys are handling. And I think, like you said, that, that it's a question like Will pointed at this and also Marcus, like which parts of this can pass?
… And there may be some parts that need to wait. I think that rushing to pass everything is
… yeah, you're already creating a situation where it's difficult to pass, and so I think it's much more about selecting the parts that aren't controversial than trying to pass the whole thing. At least that's how I would approach it, uh, in order to show progress, um
… that might not be practical in W3C, but that's kind of my perception of how standards move along. It's little bits at a time
Otto Mora: Yes, well, yeah, we we tried that. We we did have smaller Prs. And then this Pr. In particular, unfortunately, did requires larger changes as as we evolved the discussion. That was the original intent. But yeah. Um, Marcus, uh… Last comment, please...
Markus Sabadello: I just wrote in the Zoom chat, I'm sorry, I'm on a bad connection, and I may have to drop early, I cannot see the...
… Chat right now, I know this is not… not a good time, but uh… if I drop, don't take it the wrong way, I… Can I then vote afterwards, or
Otto Mora: I'm happy to also record your vote myself on the chat, no problem...
Markus Sabadello: Okay, and I agree with Grace. It would be way better to discuss individual changes at a time. I don't agree with the characterization that for some reason all of this has to be done together. I think it would be very possible to, for example, have individual PRs for all the individual items...
Otto Mora: Okay...
… So with that, let me pause the transcriber
<ottomorac> transcriber-bot-l, pause
<ottomorac> PROPOSAL: The DID Resolution specification will not define an abstract Dereference() function signature with specific inputs and outputs as a separate interface from the Resolve() function.
<manu> +1
<JoeAndrieu> +1
<ottomorac> +1
<pchampin> -0.5
<swcurran> +0.5
<Wip> +0.5
<TallTed> +0.5
<pdl-ASU> +0.5
<ottomorac> Markus: -1, breaks compatibility with DID 1.0. Algorithms need inputs and outputs
RESOLUTION: The DID Resolution specification will not define an abstract Dereference() function signature with specific inputs and outputs as a separate interface from the Resolve() function.
<ottomorac> transcriber-bot-l, resume
Otto Mora: Transcription. The second one that I wanted to...
… run here is we may. We may want to make some changes to it. I mean, obviously, we I don't want it to reflect like Pr. 3, 4, 4 is going to be merged without proper discussion and response to all of Marcus's
… uh, points that he raised. But, um, yeah, it says that we want to update that resolution to represent the consensus of the working group. um
… And replace the abstracted reference function, that's it, and it's HTPS binding with a procedural algorithm that's noted there
… uh
… Any comments on that?
… Oh, you also have… yeah
Joe Andrieu: Yeah, I was… yeah, I put it in the chat, but then I realized maybe I did it wrong, because if that gets auto-swapped, that's a problem. Um...
Otto Mora: Decision...
Joe Andrieu: But I think there's a debate as to whether or not we have consensus. I think we have dissent and a decision. So maybe instead of consensus, it should say decision of the working group...
Otto Mora: Okay. Fine...
… Position of the work from
… Yep. Okay, cool
Joe Andrieu: Um, and… sorry, also, maybe the… I think as represented in PR344 seems to bind to the decision...
… But I think that's a little bit of a step too far. Like, we aren't… the PR right now, it still has a bunch of edits, because we've been talking about it for a couple weeks, and we need to incorporate that. Um. So I think… Oh, being worked on? It's not bad?
Otto Mora: Okay, so you want me to remove the reference to 344 and just, in general, say that we'll be replacing? Oh...
Joe Andrieu: No, no, no, the man who's updated proposal, uh, sort of got that as I was talking...
Otto Mora: Oh, okay, okay, okay. Being worked on. Okay, yeah, that's fine. We can do that. Okay, okay. And then before we vote, yes, Marcus...
Markus Sabadello: I don't really understand the language in the proposal. It says replacing the...
… abstract function with a procedural algorithm, I
… Don't think that makes sense, because there is already a
… procedural algorithm. I I feel like it should say, replace the
… previous algorithm with the new algorithm, or something like that. There was already a
… a vote just now on the dereference function, and… and replacing that with an algorithm, I think it's not what's happening, because there is an algorithm, so this is
… a new algorithm with new rules and new inputs and outputs, replacing a previous algorithm with all the, in my opinion, with all the many breaking changes
Otto Mora: Any reaction to that, Joe, or do you want to try to run it with the… Okay, I see it might have updated. Bigger place, and that's your phone...
Manu Sporny: I agree with what Marcus said. I tried to update it accordingly while still staying true to the heart of the proposal...
… I… yep
Otto Mora: Okay...
Joe Andrieu: Yeah, I'm good with those changes and updates. I think the main intention here is to say, hey, we just agreed that. Um, we don't want a dereferencing function signature, so let's update the spec...
Otto Mora: Okay...
Joe Andrieu: That's really what we're saying here, right?...
Otto Mora: Alright, uh, yes...
… Marcus, last comment on this. Go ahead
Markus Sabadello: Well, it still says replacing the abstract dereference function with a new proposed algorithm. It should say replace the current algorithm with a new algorithm. That's what it should say...
Otto Mora: But then no, it's not procedural. Okay. Sure...
Markus Sabadello: Well, they're both procedural. There has been, I think this is a separate question from the dereference function. There has been a dereference algorithm and now there's a new dereference algorithm that has new inputs and outputs and new...
… new rules and new behaviors, and I think that's what the vote should be about
Otto Mora: Yes, we'll...
Pierre-Antoine Champin: Beautiful...
Will Abramson: Yeah, I mean, I think that's fine. I was gonna say, you know, this is really just about saying we are, as a group, are we on board with the direction that we have discussed that is kind of in 344, but actually is more in 344 and the shared discussion that we've been having...
… facilitated by Stephen over the past, you know, maybe month or so, right? There's a whole slide deck, which I can maybe drop in, that has a bunch of decisions that the group has made based on comments that have been raised against 344. And we're really just trying to get sort of
Will Abramson: Gage, is the group for this decision, this sort of direction that we are going down? This is the slide deck
… Yeah, I don't know if that helps
Otto Mora: Yeah, this is the site that's still an ongoing conversation, yes. OK. Mano, and then Ted...
Manu Sporny: Yeah, what you just said, Will, is not something you can do a proposal in resolution for, because it's way too murky. There's way too much work being done there. That is not, I don't think, the proposal that we should put in front of the group...
… I think we should continue working on the language here. Uh, we specify concretely that we're working on this in 344, and that is the direction that we're wanting to go. We're replacing the current dereferencing algorithm and its HTTPS binding with a new procedural algorithm
… With new inputs and outputs, um, and then the, the bit at the end, I think that's concrete enough to, you know, get, get a direction from the group, but we can't say things like, and everything that we've been discussing for the past six months, like, no, like, we can't, you know, like, what does that mean?
… Where's the references to it? What are we agreeing to, right? So… so I get… I get the gist of what you're saying, Will, but, like, this proposal needs to be made, and it needs to be clear enough so that it can stick. I think the way that we do that is by reffing PR344
… and the direction there. And
… and then, you know, the bits at the end there. Um, so, um, that's it. It would be nice if we could get a, you know, again, more tight… tighter proposal, but I think this was
… Good enough for now to get the direction that the group wants to head in. That's it
Otto Mora: OK, folks. Let's try to quickly have an intervention so we can vote before the call is over. Ted?...
TallTed // Ted (he/him) Thibodeau Jr (OpenLinkSw.com): Yeah, I don't think trying to rush through this is going to do us any good. So just that being said first, I mostly agree with what Manu said...
… This sort of offline discussion that's in… in that… slide deck. Um
… is not referenceable in a useful way. Uh, I can't see the discussion, for instance, easily, that came up with a decision that's on… recorded on one of those slides. I can just see that it started with something, and it finished with something. Um
… And that's that's insufficient for anybody else trying to pick this up
… Any decisions that we're making are provisional at best
… And very hand-wavy, especially if we're trying to rush them through. Um
… To some degree. the
… breakdown of different questions and therefore different decisions on those slides would be useful as issues
… each one their own issue, and discussion on… in the issue, and then a resolution in the issue that says, here's what we talked about, here's where we came to, and then a PR based on that issue
… Focused, small, relatively, and hopefully clear
Otto Mora: Mm-hmm...
TallTed // Ted (he/him) Thibodeau Jr (OpenLinkSw.com): And again, the decision that we make here is always provisional, and anybody who's not in this call, but on the mailing list, a member of the group, can object to that, and we'll come back to it next week, so...
… No matter what decisions we make here, let's not go away from it thinking, well, we've solved that, because there's still an opening. And I'll leave it at that
Otto Mora: Thank you, Ted. Marcus?...
… Marcus?
Markus Sabadello: Uh, could we… could we remove the part of the proposal also that says, describing how clients...
… and so on, because that's not really a differentiator. That's something that's in the current algorithm and in the proposed new algorithm, so all the language about clients calling resolve
… Prepare to call resolve, call resolve, handle the response
… Uh, I think we should remove that, that too, because that doesn't… that doesn't change
Otto Mora: Okay, so that's the bottom one, just taking out that last part, right?...
<Zakim> JoeAndrieu, you wanted to suggest less details about 344, and just say update the spec to that previous resolution
Otto Mora: All right. I believe Joe was coming after that. Joe?
Joe Andrieu: Um, yeah… I actually was going to propose that we remove references to 344. I think 344, as it currently stands, is...
<Wip> +1
Joe Andrieu: Um, a work in progress, and doesn't currently represent, um, any consensus of this group
… Um, the slide deck was a conversation to engage on all the comments that people had made
… Um, in an organized fashion. Um, and Steven did a great job sort of breaking out all those comments with the back and forth, so we could have a focused discussion on these calls, and that was really helpful. But I appreciate that that slide deck is sort of outside of our milieu of referenceability
… Um… so, I… I think… I think the only point of what we're trying to do here is to say, hey, update the spec
… To get rid of the dereferencing function
… Um, how we do that is, uh, um, what 344 is proposing, um, and it is not ready for us to make a vote on it. So, I'm not even sure we need this proposal. Now, given the proposal for, uh, removing the dereference function signature. Um… Uh… I think
… proposing a PR that does that is potentially the right next step. So… If
… If the details of this proposal are bogging us down right now, maybe we don't need this proposal
Otto Mora: Mm-hmm. Or a simpler one like this that will just replace it, but effectively a PR is the same. Good point, okay. uh… Here...
Pierre-Antoine Champin: Yeah, I think indeed the wording of the proposal book is bringing us down...
… Maybe what we could agree on, what I'm hearing is that we are agreeing to
… Not use the current algorithm as a baseline, but go for something completely different. Uh
… Of which many points have been discussed earlier, some of them Marcus agrees with, according to what he wrote, some others he doesn't
… Would it be possible to, because again, I sympathize with the concerns raised earlier about the size of the PR, would it be possible to start from scratch, just remove everything that's on resolve and then
… build incrementally the new algorithm with some first PRs that would create an incomplete algorithm but focus on one particular design choice and then a separate PR and so on and so forth. I think the reason this PR remains so big
… is that it needs to get rid of what is considered unacceptable in the current text, as well as introducing new design choices. If the agreement here is
… Let's start from scratch, then maybe it makes it possible to work more incrementally
… Or am I too idealistic?
… Okay, I get it. So yes, I'm being too realistic. Sorry
<Zakim> manu, you wanted to note "start from scratch"
Otto Mora: We tried all of that. I'll respond. We tried all of that, and it didn't work. We tried. We tried. It was just, yeah...
Pierre-Antoine Champin: Okay...
Otto Mora: Yeah, sorry, it's just… one cannot be separated from the other, maybe that's what the… Joe will speak to it as well. But Mando, go ahead...
Manu Sporny: Yeah, huge minus one to that, PA. That is what we have been doing for 6 months. Uh, I appreciate, you know, Grace, your suggestion of breaking it down into smaller pieces, PA, yours as well. We have tried that for 6 months, and we are at the point now where we're gonna exhaust the editors, and they're gonna give up...
… and walk away from the specification. So, huge, huge minus one for that approach. I usually am the one going, like, let's do small PRs and whatnot
… that's not where we are. We're 6 months past when we said we were gonna do that, and we are where we are, for a variety of reasons. Um, uh, I do think maybe one… we have a… we have a concrete resolution here that helps
… Um, I don't think we have enough. I think the chairs and during our next special topic call, we should focus on the concrete proposals and resolutions that we need. We should spend call time doing that so that we can re-raise this tomorrow
… Will you refer to the slide deck? We have made agreements, kind of, based on each one of those points in the slide deck. Each one of those things could be migrated into a proposal, and my expectation is that then it will align with 344
… Uh, or whatever 344 becomes. Um, if we want to break the proposal resolution process down into something more bite-size, that's something we could do
… but we… that is not, you know, but we're gonna end up in the same place, right? Um, I think we're just struggling with trying to figure out how to write down these proposals so that they are atomic and clean, uh, and
… effectively end up, um, implementing what's in, uh, you know, a variation of 344, plus the discussions we've been having. So, uh, my suggestion is we spend
… special topic call time next week, purely dedicated to breaking these down into smaller proposals that are more concrete, um, that we can pass. Um
Otto Mora: Mm-hmm...
Manu Sporny: That's it...
<Zakim> JoeAndrieu, you wanted to speak about consistency and the endless catch-22
Otto Mora: Okay, then, Joe, and then we'll have the last word...
Joe Andrieu: Yeah. Unfortunately, man, I'm huge minus one to that. In fact, it it may push me over into resigning as editor. Like, I really don't think what is going to help us is to have more issues or more proposals. What we're doing is we're taking every single little byline...
… And we are expanding it into debates that we have to have over weeks and weeks and weeks. I'm positing that that
… pattern of imagining that there's less work if we break it into smaller bits has… is not true. We should have known it wasn't true when we first got there. Um, and I'm… I'm worried about taking what was a healthy conversation that Steven has migrated
<manu> I agree with what Joe is saying FWIW
Joe Andrieu: Um, as managed, facilitated for us. There's a bunch of changes in there that everyone in the call was good with, and those are gonna go into the PR. And if you make me wait for proposals next week
<ottomorac> +1 to Joe
Joe Andrieu: And we have to debate every single proposal. We are adding at least a month, if not more, to getting through that PR. And that would be incredibly frustrating. And I don't know that I have the patience to put up with it. But I also want to speak to, um, that the
… I need to push back, sort of generally, against this idea that big PRs are fundamentally bad, because the problem in the spec is about consistency and confusion
Pierre-Antoine Champin: No...
Joe Andrieu: And so we have gone through this process of trying to break it down. My initial PR was pushed back against. I broke it down into a handful of other PRs. Some of those were adopted. Others were, wait, this is still too big. And I was asked to break that down again. So this is the broken down version of the dereferencing all by itself.
And now that we're here, we're getting pushback because it is inconsistent with the rest of the specification...
… The reason it's inconsistent with the specification is because we had to do it in isolation. And so what we're trying to figure out is what is a consistent, coherent narrative that expresses and resolve expresses our, our, our decisions
… Um, uh, and is understandable and usable. And so there's… I'll let it go at that. I
… I'm curious what we think we need in terms of a second proposal here
… Because I think there is some sense that we do need it, but I'm not sure, given that we had the other
… Uh, resolution, why we'd need more
<pchampin> FTR, I appreciate that there is a tension between divide-and-conquer and consistency :-/
Otto Mora: Well, quickly...
Will Abramson: Uh, yeah, I just wanted to say, I think the proposal I would suggest is let's just get all the changes into PR344, and then we can, like, vote on whether we approve that PR, right? Like, that feels like the best...
Otto Mora: That could as well be the effective replacement, yeah, VR itself. Yeah, okay, so to be discussed in special topic call...
… And yeah, thank you all folks
Joe Andrieu: Thanks, Adam...