Meeting minutes
<ottomorac> please use correct zoom link: https://
Otto Mora: Hello! How are you, sir?...
Ivan Herman: Hello...
Otto Mora: Hey there...
Kevin Dean: Hello...
Otto Mora: Oh, let me… let me pause the transcriber before...
<ottomorac> transcriber-bot, pause
<ottomorac> please use correct zoom link: https://
Agenda Review, Introductions (5 min)
<ottomorac> transcriber-bot, resume
Otto Mora: Resume...
… Okay, so for the agenda review, today we're just doing a quick debrief. From the
<ottomorac> https://
Otto Mora: a special topic called yesterday, which I sent an email to the group just a few minutes before, but I'll still mention the summary of it
… Um, the link to the minutes is here as well
… And, uh, then we'll, again, talk about the URL… Uh, do referencing VR
… Then we'll talk about the DID path PR as well
… And, um… yeah, I guess so, we'll start… we'll start with the
… With the debrief, but any… any other point before we start with the debrief?
… Not seeing any notes raised
… Okay. So, let me, again, thank you
Debrief from the Special Topic call for 8th Apr \[1\] (15 min)
Otto Mora: Yes, so for the debrief, um, from the call yesterday. And in general, um
… We talked about the two approaches, um, to the DRLD referencing
… And then trying to find a consensus for a sort of a generic path handling algorithm
… Uh, that also balances the need for DAD method-specific logic
… Right? Um, some portion of the call, we talked about, uh, this idea of the precedence of the TID method
… uh, specific handling, like the VAD method can have a specific handling algorithm, versus just having a generic one
… I think Joe raised some concerns that perhaps prioritizing
… Method-specific logic, uh, could create problems. For interoperability
… Um, then we also had a discussion around whether the dereferenced URL should return an actual resource
… Or just merely a URL that points to the resource
… I think Steven then suggested that we
… move the method-specific handling logic further down the priority list. I believe that was going to be the change
… Um, that we sort of, uh, agreed there. Um
… And then we, uh, talked about some of the terminology, I think that's also, uh
… Beneficial, uh, because there were some
… I guess differences of opinion on what a client, a resolver, and a dereferencer meant
… And, uh, then I think also Joe correctly pointed out another problem with RFC 3986
… Uh, the reference resolution algorithm, because there's also some… some issues there that… I believe he pointed out that also in WhatWG, those issues persist. So that we also have to think about that
… But we didn't exactly, like, reach an agreement on all of the
… let's say how to move forward, but I think at least we are more in tune with what
… Each of the changes entail, and
… Also, clarity on the… some of the concepts, I believe
… So, that's my summary of it
… But, uh… let me see, I see Will there, maybe Will?
Will Abramson: Sure, yeah, I guess I wanted to highlight two things. Like, I think towards the end of the call, we were discussing, like, relative ref...
… and essentially removing relative ref. And at least on that call, there seemed to be, you know, like
… moving towards consensus. And I think I suggested, well, we should raise issues
… Like, really what I would like to see from today is if there are points of consensus that we can come to, then I would like them to be addressed as individual
… features, ideally, wherever that is possible. Like, for example, this relative ref change. If the group feels that that is what we want to do
… like, I think it would be good to try and extract them out of these large PRs and do them as small changes that are easier to review
… And is it… and then through that process, we will be able to identify the actual points of disagreement that this group has
… And then we can have real, like, debates to try and move towards consensus around that
… Um, because, like you said, Otto, we talked for an hour yesterday, but I don't think we made
… any, like, concrete progress towards accepting either of the PRs. That was my interpretation, anyway. Maybe it is wrong, but
… Just some thoughts, like, we need to try and make progress. That's one suggestion
<Zakim> manu, you wanted to suggest realities and ways forward.
Otto Mora: Uh, mano?...
Manu Sporny: Can we turn scribing off for a second?...
Otto Mora: One second...
<ottomorac> transcriber-bot, pause
<ottomorac> transcriber-bot, resume
Otto Mora: Transcriber thought… Uh...
Stephen Curran: Uh, can you leave it off for just one second?...
Otto Mora: Oh, yeah...
<ottomorac> transcriber-bot, pause
<ottomorac> transcriber-bot, resume
<ottomorac> transcriber-bot, resume
<manu> manu: I think we should break these issues into manageable chunks and work through the end of the charter (with the presumption that we'll be given a charter extension to do the right thing)
Otto Mora: Great, okay, yeah. Um, I also saw some proposals for Marcus, uh, I think...
… Maybe, Marcus, you want to chunk the changes? Um… And
… So, I'll let you, uh, go ahead, Marcus
Markus Sabadello: Well, just to say that I fully agree with Manu that it would be easier to have smaller PRs that are easier to review, because myself also...
… I really read through Steven's PRs and Joe's PRs, and uh
… when I look at those, I see maybe 2 or 3 things I like, and then I see 2 or 3 things that I really don't like, right? And then it becomes
… Really hard to to have conversations and to get to some
… PR that everybody accepts. Um, I created a few PRs
… today, and in the last few days, I think I have 5 open PRs right now, and they are all really small, right? So I'm… I'm hoping that, um
… Even if they don't address all the open concerns and the confusions and the discussions
… Maybe some of them can be steps in the right direction
… And, uh, that's it
Otto Mora: And just to understand, you cherry-picked also some changes from Joe's… As well, right?...
Markus Sabadello: I… I didn't...
Otto Mora: some sections...
Markus Sabadello: I didn't pick any content or commit from anyone else, but I think some of my PRs. Um...
… include some of the ideas from other PRs
Otto Mora: Okay...
… Uh, Dimitri
Dmitri Zagidulin: So I think we have… I think everybody agrees that...
… we need small PRs, uh, on individual issues. My main question is
… are the two big arts. Are the smaller bits separable?
… Or do they necessarily have to be done as a… Uh, as a group in big chunks
<Zakim> manu, you wanted to say PRs are always separable.
Otto Mora: Uh, manual...
Manu Sporny: PRs are always separable. So, you know, you can always… you can… you can do a bunch of small PRs to get to a destination, or you can do one giant PR to try to get to that destination in one go...
… I think we've tried one giant PR to get to the destination in one go, and it has not been going well. Um, uh
… So, I think that's kind of… we've got to accept that we've got to
… Go… go towards the goal that I think many of us have consensus on, you know, in smaller
… Smaller steps
… Um
Otto Mora: Mm-hmm. A little more bite-sized, right?...
Manu Sporny: Yeah, the other thing I kind of want to mention is I don't think we're meeting often enough to work through these issues. I know, like, you know, we don't want to necessarily hear that, but I think a lot of the miscommunication that's happening...
… in the PRs is because we are not… the people that are disagreeing are not on the same call working through those things in a very concentrated, focused way. Um, I… I
… I would like us to start meeting daily
… To move through these things, I know that is a lot to ask, especially for people that are not
… Um… you know, funded to do this work. Uh
… that's just a request, uh, because I don't think that the cadence that we're working through these things is… we need a higher bandwidth way to work through these things, and I think we need to be more focused and structured on the. on the… the issues that people have. Um
… And that doesn't mean open 20 issues and use the issue tracker to process through all those issues. I'm thinking of something more like
… copy and paste Joe's, these are the things I changed, you know, into a document
… Um, do the same for Steven's PR, and then potentially add Marcus's things, see if they map
… Um, and… and go at it that way, so it's… so we're not, you know, weeks long processing this stuff, so that we're
… Able to kind of focus on things that, um
… Are the easy changes to make to the spec, uh, versus the… the really
… um… contentious ones, right? And ideally, we focus on the things that are, like, these are normative changes that we feel absolutely need to be made to the spec before we go into CR, because they do change things fairly
… you know, significantly. Um, so we need help from
… the folks raising the issues and wanting the changes to identify what those things are. Um, and in many cases, it is the thing that is going to lead to a formal objection. That is the thing that we need to. you know, potentially
… Come to consensus on, on do we think we're going to be able to address this or not? Um
… Anyway, so all that to say, like, I don't really know where anybody else is, um, um, on a number of issues
… it would help to have, for example, a Google Doc that has… these are all of the changes that, you know, all of the people that are potentially disagreeing want to see made
… And then we, you know, in real time, on a call
<TallTed> +1 frequency increase to higher bandwidth than text-on-screen. Daily may be too much. 3x/week may be enough? or maybe MT+TF (off W&SS)?
Manu Sporny: with those parties on the call go through and see, is this editorial or is it normative? Does this need to be made, you know, before we go into CR? Will this lead to a formal objection? And just process the issues like that
… That's it
Otto Mora: Mm-hmm...
… I don't know all the logistics, but I will plus one that, uh
… talking, like, when I… like, virtually, like, this face-to-face, it is a better medium than the
… uh, people responding to each other's, uh, threads on GitHub, and… Some of the
<Zakim> JoeAndrieu, you wanted to clarify if there is consensus is to break up the two big PRs
Otto Mora: The tone being misinterpreted, which, you know, I think can easily happen over text. Um, so
Joe Andrieu: Yeah, two things, um...
… Uh… the first one, Manu, I'm not sure how the most recent part of your proposal
… Um, isn't what we already have, in the sense that if we… if we create
… a big list of all the things that are different. That's the diff file, and if we have one meeting to argue about that, I feel like those are the meetings we've been having. So, I think we could maybe reinvent that a little better
… The other one is maybe the other way to reinvent it, which is what I thought I was picking up some consensus tours from the leadership here
… Um, and I support it, I guess, which would be to take the two big PRs and break them up into smaller PRs
… I'm not exactly sure how to do that, but, you know, I could try. I mean, I think there are some conceptual things, like
… you know, the algorithm versus the particulars of the dereferencer. Um, so there are ways that it could be broken up. I agree with Manu's assertion that you can always break it up
… Um, but those feel like two different ways to tackle it
… That was all
Otto Mora: Okay, so… Manu, were you there already before? Or… Go ahead. Go ahead...
Manu Sporny: Yeah, yeah, I put myself back to kind of respond to Joe. Yeah, um, I'm not suggesting… yeah, I can see how it can… it sounded like two different ways. Um...
… I'm suggesting we break up the PRs, right? Um, uh, but the… so, you know, plus one, let's break up the PRs
Joe Andrieu: Yep, agreed...
Manu Sporny: Uh, I'm thinking things like, uh, you know, Joe, in your PR, there is a part where you modify the abstract of the document. That could be a PR on its own, right? Because it's, uh, it's editorial, it, you know, but it does introduce, you know, a different way of thinking about the thing. So, so that is one PR...
… Uh, and uh… and the group can, you know, evaluate that on its own, and, you know, it either goes in or gets modified and goes in, or doesn't, right? And that's separable from, like, a change to the algorithm, or a change to what the
… Inputs to the resolution algorithm is, and so on and so forth
… Um, so let's do that. Uh, what I was trying to suggest is, like, how do we
… Figure out which PR is separable from another one and at least in your PR, Joe, in the, when you opened it, you said
… these are the things that I changed, right? And I thought those were really good, succinct statements, because I think most of those could be a PR on their own. I realize that some of them touch different parts of the document, but conceptually, they were
… Uh, they were focused. And conceptually, we could have a discussion around that sentence you wrote. This is what I changed, you know, for these reasons
… we could ask the question of everyone, does everyone agree that this is a change that we need to make in the specification? Does everyone agree that this is editorial or normative? If we don't do this, is it going to lead to a formal objection?
… Um, those… and it's just a temperature check, right? It's not a, we need to get in a big giant debate over it, it's just a first-level pass on, like
… Do we think we have consensus on this particular concept, uh, or not?
… Um, so, so the first thing I mentioned was a way to get to
… a bunch of concentrated PRs, or focused PRs, and work through it in a systematic fashion. Um, that's it
Otto Mora: Okay, Dimitri...
Dmitri Zagidulin: Uh, no, just pass on to what, uh, Mona said...
Otto Mora: Okay...
… So, uh, on this one, I think so, at least, yeah, let's, let's, let's, uh
… divide them up as, you know, slice them up into smaller chunks. Let me at least propose that we try to do the
… the weekly special topic calls, right, Will? I think that could be helpful to have these more intense work sessions. Um
… Uh, yeah, I see Manu on the queue. Go ahead
Manu Sporny: Yeah, plus one, I wasn't suggesting that we cancel those, I'm just saying, like, we need a higher bandwidth place for Joe, Steven, myself...
… Marcus, uh, Dimitri, anyone else that is going to have a strong opinion on this stuff, to just
… talk things through, but on… just very focused, like, the… the, uh… and I… and I agree, Joe, like, I think some variation of that has been happening, based on, like
… certain comments raised in the PR, um, uh, but in some cases, you know, I think, you know, people are… and I have heard this across the board
… That, you know, people don't feel that they're being heard, um, and we need to make sure people feel that they're being heard on their concerns
… Um, and uh, and in order to do that, I think we really do need to focus down on one thing, because people jump from one thing to the other based on, like, what they really
… you know, the concern that they have, and sometimes these things are intermixed, and it's just led to a point where I think we're
… You know, miscommunicating on the issue tracker, unfortunately, and we just need a higher bandwidth way to
… Talk with each other, make sure everyone's, you know, feeling heard, keep it focused on
… You know, these… these, you know, specific changes that we want to make on the specification. Um, and then
… And then just, you know, get a pulse of, like, where we are, like, you know
… what is the full list of changes we want to make to the specification? I don't… I don't know if we have that
… Right, I know there, it's, it's in there, it's, it's in the 180 comments that have been submitted to the issue tracker, but like, I, in my head, I'm like, I have no idea
… Right? Um, I don't know the things that are on, you know, the top 3 things that, you know, Joe's most concerned about, or Steven's most concerned about, or Marcus is most concerned about. Um, and I think we need to know those
… So that we can identify them, and we can have a focused conversation on each one of those
Otto Mora: Mhm...
Manu Sporny: And then close it out, right? I mean, at some point, we have to say, okay, yes, we have consensus on this, like...
… we're done, we've made this decision, now it's just a PR to, you know, move that one forward. We're on to the next one
Otto Mora: Right. And… And just to understand, so I think...
… the… like, Joe's PR is the one that we can break into smaller chunks, right? Like, I see, kind of, Steven's PR as more
… specific to the… to the path, uh, aspects, right? So we… we can't break that one up, but maybe we can have a discussion next week
… about which elements of Joe's PR we can break into smaller chunks, if that… Everybody agrees with that. Uh, yes, my name
<swcurran> +1
Manu Sporny: Yeah, plus one to that, um, and I definitely want to hear from other people. I'm a bit concerned about the silence. I don't know if that's agreement, or if, you know, we have disagreement. Um...
… Uh, I think, Joe, your… this is not going to be a perfect… so I think, well, what could be done
… Uh, is if people, um… well, sorry. I think Steven's PR can layer on top of Joe's. There's some variation of it where that's true
… Whether or not people agree that we do all of those changes or not is still up for debate, but I think we can layer things, you know, Steven's PR could layer on top of just about any one of these things. It's just… it's supposed to be just about the path
… you know, feature, whereas Joe's is more about, like, you know, resolution and dereferencing and clarifying, um, you know, that sort of stuff
… Um, Joe, in his, uh, PR, has a set of, like, I don't know what they're… 12, 13 things, where he was like, this is a… this is… I made these changes, 1, 2, 3, 4, 5, 6, 7
… Um, I think we can break many of those into individual PRs. I don't know, Joe, if you agree with that or not, but there's some subset that we can
… break into individual PRs, and I think we should do that, right? Um
Otto Mora: Mm-hmm...
Manu Sporny: That's how I suggest we make some progress here is, is we take Joe's PR, or sorry, we take, we take Joe's, like, these are the, you know, 12 things that I changed...
… We see which one of those we rapidly have agreement on
… Or not. Um, and we start raising PRs for the things that we have agreement on, and start, you know, merging those down. Um
Otto Mora: Mm-hmm...
Manu Sporny: The challenge is going to be, if there is some huge disagreement, you know, some really huge base-level disagreement...
… Um… you know, we may not want to
… we may want to kind of, you know, push that to the end, or we may want to front-load it and say, well, this is really critical to, you know, all the other PRs that we'd raise. Like, if we… we don't need to raise the other 5 PRs if it's dependent on this one thing that we just don't feel like we're going to be able to get the consensus on
Otto Mora: Mm-hmm...
Manu Sporny: That's it...
Otto Mora: Okay, and then one more thing. I think it would be useful if, uh, we do this in a way where somebody's screen sharing. I would screen share, but last time I screen shared, that crashed the transcriber on my computer. So, I don't know, Will or someone else, if you're...
Will Abramson: Yeah, I can screen share, that would be fine...
Otto Mora: That would be best, yeah. But… so before we go into that, I see Will and Joe on the… Uh, queue, so go ahead, Bill...
Will Abramson: Yeah, I really wanted to agree with Manu about, like, the way that we handle these. It feels...
… most sensible to handle Joe's changes, try and work through those first, and then layer on top Steven
… just because at the moment, it feels like we're duplicating a lot of effort arguing about some of the same things, right? Joe's arguing about the things he's arguing for on Stephen's PR, and then Stephen's also arguing
… about Joe's PR, like, it feels like there are some fundamental things in Joe's PR that we need to decide on
… And then, however those land, we can then decide on how we integrate past services. It feels like
… We do have consensus that PathService is a feature that we want to get in there
… Uh, but it… it makes sense to me to maybe, like, pause that feature until we've
… Figured out, like, the foundations of that feature's gonna sit on
… Uh, yeah, and I'm happy to share when we're ready to do that. I think that makes good sense
Otto Mora: Perfect, yeah. Yeah, because that way, exactly, we can start noting down the sections. More visually. Uh, Joe...
Joe Andrieu: Um, yeah, plus Wanda, I think we can layer, I think...
… Um, the… my PR was, in fact, a response to my inability to
… sort of address my concerns within Stevens. So part of it was designed to break out the algorithm so it would be easier to clarify where my concerns are. Um, so I do think we can pull it in, and I want to, like, I want to have this sort of file service capability. Um, we just need to figure out a way to do it. Um
… I think I had misunderstood what Matthew's proposal, if we're gonna go right in now to the conversation
… Um, I was hearing, like, we can break up this PR. I think I can take a stab at that this weekend, you know, but if we're gonna talk about it right now, that would be awesome. I'd love to go through those statements
… you know, I think there are 14 of them, and get some sense of the group, that would help me break it up. Um
… In terms of, you know, if half of them seem to be well supported, then I can put those in one PR, and if there are… so… If we can have that conversation, that would be great
<Wip> +1 lets do that
<Zakim> manu, you wanted to say yes, let's go through those statements.
Otto Mora: Yeah. Yeah, yeah, absolutely. Uh, yep, and Manu?...
Manu Sporny: Even if they're… even if they're 14 that we agree on, or 7 that we agree on, please put them in separate PRs. I'm concerned about big PRs. Um, but of course, like...
Joe Andrieu: Well, I don't think 14 PRs is going to be a good idea, but breaking it up, I think, will be a good idea...
Manu Sporny: Sure, okay, we… okay, um, uh, but yes, let's, let's, uh, going back to the previous thing, yes, I, I, what I'm suggesting is...
… Uh, Will, if you could put up a Google Doc that's worldwideable
… Copy and paste. Joe's 14
Will Abramson: Yeah...
Manu Sporny: Things into that, give us access so that we can start in parallel providing. like...
… You know, input, I'm not saying I have any other plan beyond that, but just getting them on the screen and getting people the ability to, in parallel, provide input instead of queuing and going through each one and listening to everybody speak, I think might speed things up
<Wip> Here is the doc: https://
Otto Mora: Mm-hmm...
… Okay, so rather than spending time on the call breaking that up right now, we can
… Oh, let's do it, okay
Will Abramson: No, let's do it. I mean, I can share my screen too, I think it makes sense...
<TallTed> To the degree that the 14 theses are atomically breakable, I would prefer one PR per atom. If there are 2 or 3 that cannot be split from each other, that combo may be worth keeping together
Manu Sporny: No, I'm saying do it, it's just everyone in parallel, plus one, minus one, and really what we're looking for are, like, minus ones to Joe's...
Otto Mora: Okay...
Manu Sporny: suggested changes. And it...
Otto Mora: So I guess, as we're screen sharing, people can just plus one and minus one on the… On the chat. You guys want...
Will Abramson: Yeah...
Manu Sporny: Well, in the document, like, let's, let's get it, you know, in the document...
Otto Mora: Oh, okay, okay, okay...
<ottomorac> link again: https://
Will Abramson: So, this is the document, and I think, Toby, the goal is to both identify places where we are in agreement, because they're easy PRs to just get in and tick them off, right? And then also identify places where we have. Strong disagreement, so we can discuss...
… Um
Otto Mora: I see Ted already...
TallTed // Ted (he/him) Thibodeau Jr (OpenLinkSw.com): Just briefly, to the degree that these 14 theses...
… to coin a phrase, uh, are atomically breakable, it would be better to break them up into one PR each
… If there are, like, 2 or 3 that cannot be separated from each other, then keeping those together probably does make sense
… But… I'm hopeful that most of them are atomic. That's it
Will Abramson: Yeah, actually, looking at what you… yeah, thanks, sorry. Oh, mine isn't queue...
… Sorry, man, can I just jump? I just wanted to see, looking at what you are doing here, like, maybe it would be better for us to run. Like, a poll in the
… in the IRC that people can post on my
… Anyway, whatever
… Thank you
Manu Sporny: What I'm suggesting is Joe explains each one of these in a very high level and then we provide input here. I want to make it dynamic because people might change their mind as we talk...
… Um, running polls takes a bit of time, and
Will Abramson: Mm-hmm...
Manu Sporny: Do you see what I'm saying? Um...
… Let's keep it loose for now, uh, is my suggestion, see how it goes, and then we can do more formal resolutions if we want to later, but
… At this point, there's a lot to go through, so let's try and parallelize the work as much as possible
Joe Andrieu: Yeah, I want...
Manu Sporny: And again, just suggestions, I'm not, you know...
Will Abramson: No, no, that's fine...
Joe Andrieu: I just want to endorse the parallelization, like, if we can… we should put some bullet points, but… After all of these, and...
… So, you don't have to wait to put your comment in while other people are debating other stuff. I think that will speed things up a bit
Otto Mora: Yep...
Joe Andrieu: Oh...
Will Abramson: Okay, do… yeah, you go for it, Joe, I think, if you want to start with one...
Joe Andrieu: Okay. So, um, the first one, actually, I reverted out, but it is still, um, potentially valuable. Um...
… which was to define dereferencing first, because that's what you do with the did URL, and which resolution is a part of that, so
… That was about aligning the order of the end-to-end
… Um, transaction, with the order of how we talk about those parts, uh, in the… In the document
… But that was a big change, so it looked like I edited everything in every section. So that's why the second PR doesn't have that. Um… so
… I actually think we should… in terms of breaking things up, this would be, like, the last thing we would want to do. Um
… So that we can evaluate the other things on a smaller basis
Otto Mora: What do partial minus 0.03.5?...
… Okay. Okay
Manu Sporny: It means I don't know. I'm kind of like, eh, you could do… I could go either way, but I'm definitely not gonna, like, you know… like, a minus one is like, I'm gonna formally object, or I cannot live with this...
Otto Mora: Okay...
Manu Sporny: You know, versus what we're seeing there, which is like, eh, I'm kind of unsure, I want to see where things go. Marcus is like, eh, I'm pretty, you know, unsure of this...
… But I'm guessing, Mark, it's like, if we did this restructuring, you wouldn't formally object over it. Is that a proper interpretation?
Markus Sabadello: Yes, it depends on how it's done in a concrete PR...
… But I don't think I would object
Otto Mora: Okay...
Will Abramson: So I just wanted to suggest, uh, after that. Joe, I think you should just talk to each of these, like, maybe pause briefly after them, but, like, I think let's not discuss any of them...
… in the chat, right, until we've tried to get through all of them. We might be able to get through all of them
Otto Mora: Mm-hmm...
Will Abramson: By time, and then we'll have a good sense of which topics need time on a future call...
Otto Mora: Agree. Let's… go ahead, Joe...
Joe Andrieu: Yeah, agreed. Like, if I keep to one minute each, we might get through them...
Will Abramson: Okay...
Joe Andrieu: Uh, the second one, hopefully, uh, is not too big of a deal. It's just past a folded URL instead of stripping it out...
… Um, since there may be parameters in it that the resolver needs anyway, um
… Uh, number 3 was, uh, there was just a sentence about, uh, when you would use parameters and when you wouldn't, um, and it wasn't consistent with how I understood it
… Um, in particular, I would not advise that to my clients, so it felt like unnecessary advice
… Um, number 4, uh, this is, I think, the biggest, most controversial thing, which is, um
… Uh, the dereferencer introduces a new object that I don't think is necessary in the flow, and I think it complicates
… Um, the definition of the algorithms, because we're forced to go
… To say what is returned from this function, um, when in fact dereferencing may do any number of things to apply it to the current context
… Um… so I expect that, you know, is the most controversial here. Um
… I separated service and service type. Um, I do think there is an error that Marcus pointed out
… Uh, should put the relative ref follow-up in both of those
… Um, but one of the things that was confusing to me is what happens when both of those are true
… And that's an open debate, um, about how we would want to handle that. Um, my algorithm handles it, um
… by falling out. Like, if the service is there, that's what we use. And so, if service and service type are there
… Um, we never process a service type, and so whether or not we should, uh, I think is a fair question
… Um, but the… the way the algorithm currently reads
… Sort of… we toggle sometimes on… at different phases in the algorithm. Uh, and that was confusing to me. Um
… Okay, number 6, um, since I got rid of the referencing, uh, as a standalone thing, the accept header doesn't make any more sense
… Um, if we keep the dereferencer, especially as an HTTP binding, then, uh, that would make this more complicated to remove
… Um… 7
… Uh, I updated the resolution image. Um, I think, uh, we had some weird thing… well, it felt weird to me that the
… Uh, the resolver was returning… somebody was returning things that felt, uh
… counter to the way at least I was thinking about it, so I updated it such that, uh, resolving clearly returns a did document, dereferencing returns a resource
… And that resource might be a DID document, because at the end of the day, that's what it's pointed to. Um, so slightly different diagram. That made more sense to me. 8, um
… Uh, this is the main algorithmic change. Um
… Which was to say, hey, there are different things that could influence the
… the retrieval strategy, how you get the actual resource, um, and I think that's the most complicated part. Like, the linked resources
… Um, can embed things in the did document. It can even have resources that are only there by obfuscation. You don't have a way to get it
… What you have is a hash to guarantee that you have the right thing, presuming that the person who wanted you to have it let you have it
… Um, so… the… the thought was, hey, what information do we have?
… When we need to figure out the retrieval strategy, and let's prioritize, um, an order so we can have a deterministic way to walk through that information
… Um, and so that's what's presented there. Um, and so the algorithm may not be quite right, but it was an attempt to break out. Well, hey
… First, we need to figure out the retrieval strategy, then we do retrieval. So this is part of that separation
… Number 9, um, I added a step where the dereferencer
… uh, selects a resolver for a givenID method. Actually, I think this is really important. We… we don't actually have in there anywhere that
… the user should even be able to choose the resolver, and I think it's a must, actually
… Um, I think our whole security foundation is based on the users being able to choose which resolvers they rely on
… Um, 10, uh, if we're passing the whole did URL, um, then
… the client doesn't need to parse the URL, um, and pull out the parameters
… Um, so I updated that to clarify that that's for overwriting defaults. Like, if you have a did URL, and it came up as, you know, this is revoked, or the key didn't work
… If you want to go back and get the did, uh, based on aversion time, then you would do that to override that value. And so that's what that language was trying to clarify. Uh
… Name, uh, this actually, I've… I've changed, um, based on Marcus's feedback, and I don't… I don't know if it's
… now, right? But, um, let me update it, because it's now base, uh, did URL
… which is the URL without the fragment, um, to clarify that what goes through Resolve doesn't have the fragment in it
… Um, we sort of said that, but we didn't call it something different, so I wanted to name those two things, um, especially in the threat modeling
… Because those are two different objects with two different, sort of, dependencies
… Um, and if we want to change it to something else, I'm open to that, but I just wanted something other than the did URL modified. Or however we had described it. Um, 12… Um
… So, I actually spoke about this a little bit earlier, uh, this deterministic heuristic. So
… you know, if the first one works, we don't continue. That waterfall felt reliable to me
… I'm happy to argue over, you know, the ordering and the precedence. I think those are interesting questions
… Um, but my contention is that I think a deterministic flow-through, um, is… is clearer and easier. Um, and before… in the other algorithm
… I felt like we were constantly keeping track of, well, if this is that, and this is that at different stages, so it was hard for me to
… tease out, you know, am I in a situation where there are multiple service objects? I don't know
… Um, I had to keep track of that in a way that was a little… a little challenging
… Uh, 13, uh, separate step for executing retrieval. Um
… And, uh, this was simply a consequence of the refactoring, that first we figure out retrieval, then you go do it
… And it meant that we didn't have to define all the retrieval strategies, because, in fact, it is the
<TallTed> I'm not keeping up with the vocal while reading while writing. the straw polls may need interpretation a few hours later, after people like me can catch up, after our other concalls today.
Joe Andrieu: Uh, it is the object which is providing path handling services which defines that retrieval strategy
… So, if you have an HTTP URL, maybe it's a get
… Um, if you have a different kind of resource coming out, maybe you have to construct it differently. For example, if it's an inline resource
… Then you don't do it again, um, because that's… that's not how you quote-unquote would retrieve it
… And then 14 to get through it all, and then we can go back and start processing some of these comments. Uh
… Uh, I added this step of using the resource, um, which is, you know, so you've gone through the… you did
… You did resolution, uh, you figured out your strategy, then you executed the strategy, now you have this thing, um, now you apply it to the current context
… And what happens there, um, could be anything. Um, it depends on the context. Um
… So, that's… those are the 14 points
Otto Mora: Okay, so let… I'm starting to process the scoring on the...
… On the first two, here with the calculator
… And then, uh, folks can just, uh
Will Abramson: Oh...
Otto Mora: Sorry...
Manu Sporny: Otto, I suggest not doing that. Um, I… this is just kind of to get a feel, like, it's not like we're gonna, you know, do floating point addition on all these numbers and come...
Otto Mora: Okay...
Manu Sporny: do you see what I'm saying? Like, what we're really looking for is the minuses, because, like, if they're pluses, even if they're weak, it's like, go ahead, right? It's the minuses that matter, and that's the stuff that we probably need to focus on...
… Um… and again, sorry, just a suggestion, I'm not insisting on that, but
… What we're trying to just do is just kind of get a weak, you know, get some weak signals. From the group on
Otto Mora: Okay...
Manu Sporny: You know, what do we think about this?...
… And I'll note that we need to pro… we definitely need to do this for Steven's PR as well, as Marcus's… PRs that he raised, um
… That's it
Otto Mora: Okay, so we have Will, Ted, and then Dimitri. Well...
Will Abramson: Uh, yeah, great. I mean, I think this was really useful. I think one that… Maybe it hasn't been covered in these options...
… And maybe it's partly to do with, um
… the HTTP… removing the HTTPS binding, but one thing that, Joe, you've also done is, like, removed the definition of, like, dereference as a… As a function with, like, strict inputs and outputs. Like, that to me feels… sort of… separate
Joe Andrieu: Yeah, that's… that's getting rid of dereferencing as a function...
Will Abramson: Right. So I didn't know...
Joe Andrieu: So, that's number 4, the dereferencing API...
Will Abramson: Right. Okay...
… So I guess, yeah, I guess I was thinking of those two as two separate things, right? There's getting rid of the API, and there's getting rid of the HTTP binding
Joe Andrieu: Uh, sure, we could… we could separate that. Um...
Will Abramson: Okay, it is in-house...
Joe Andrieu: But if we don't have the dereferencing API, then we can't have an HTTPS finding...
Will Abramson: Yeah. Agreed...
Manu Sporny: plus one for separating the question, though, because I'm plus one for one of them, and minus one… well, I don't… I actually don't know what I am for those, um...
Otto Mora: Ted, do you want it, or maybe not?...
… So, uh, Dimitrix?
… Go ahead, Dimitri
Dmitri Zagidulin: Um, you know what, I'm gonna put my hand down, and I'm gonna come back to it when we go to the next item. I think one of the...
… Well, one of the reasons that I put a minus one on, uh, I think number 5 is I'm not sure I agree with fundamental premise, which is that dereferencing
… As laid down in the spec, needs to return the resource
Otto Mora: Uh, okay, I see Ted again. Go ahead...
TallTed // Ted (he/him) Thibodeau Jr (OpenLinkSw.com): Yeah, sorry. Um, I was not at all able to keep up with...
… listening while reading, while writing, and so my feeling on most of these is not noted. I'm gonna need at least a few more hours today to be able to do that
… Um, what other people do is, you know, what you do. Um
Otto Mora: That's probably fair...
TallTed // Ted (he/him) Thibodeau Jr (OpenLinkSw.com): While listening to this bit, however, um, I did not at all understand removing dereferencing API to mean remove dereferencing function...
… So, this again is the problem of lower bandwidth inwards on the screen
… Then the higher bandwidth of conversation, and just to put that out there
Otto Mora: I don't know...
Will Abramson: Hmm...
Manu Sporny: Yeah, plus one, let's get some clarity around some, like, are there any of these that people don't feel like there's good clarity around what we're… what we're agreeing to? Um, I definitely, you know, 4 and 4A...
… I… when we say remove the dereferencing
… API. That means that API and the algorithm that's associated with it, or does it mean the HTTP API, or does it mean
… uh, just the algorithm, right? So there's… like, I'm taking 4, remove the dereferencing API and then the function signature to
… remove the API and remove the entire algorithm, even though that's not what it says. That's, like, I'm having a hard time… like, I understand remove the HTTPS binding for dereferencing. I'm super plus one
Joe Andrieu: So, Manu, I could answer, I'm on the queue waiting to do so...
Manu Sporny: Yes, please. Yep, yep, please, go...
Otto Mora: Go, go...
Joe Andrieu: Um, so I think your language was actually incredibly consistent, because you kept adding and the algorithm. This is not removing the algorithm...
… We need to define how people who have a did URL pre-process it before calling a resolver
… how they call a resolver, and what they do when they… with the result from that resolver. Um, so that algorithm, uh, remains
… It just isn't forced to be a function that has a return value that has to be a singular thing
… What the dereferencer actually does with the result of resolution depends on the context. You do a different thing if the data you're retrieving is being used to verify a proof
… than if the data you're retrieving is defining a namespace in an RDF file
… So, uh, not the algorithm, but definitely the function
Otto Mora: No...
Manu Sporny: Okay, um, uh, got it, I think, um, so we're not talking about, we're definitely not talking about removing the algorithm, like, that's not on the table at all...
… in any variation of anything we're talking about. Is that correct?
Joe Andrieu: Correct, not in here. This was adding the algorithm in this PR...
Manu Sporny: Right, and so maybe we say this is the function signature, uh, remove the dereferencing function signature. Is that… would that be a fair way to...
Joe Andrieu: Sure. I think...
Manu Sporny: The other, the other somewhat concern I have is like, you know, algorithms usually have inputs and outputs, and I'm, I guess we'll just wait until the PR happens to...
… because I'm a little confused about, like, how we're gonna specify an algorithm with no inputs or outputs. I get at a high level, it's… Okay
Joe Andrieu: So the PR current… the PR currently says that the algorithm may return a value, and it may affect. Uh, state...
… when I'm… when I'm using something for a proof
… Like, the JSON form of that can be used immediately and locally to verify the thing, and nothing needs to be returned. That's why affecting program state is the more open-ended answer
… When you… when it has to return something, and we have to define that, but now everyone has to
… Um, then now my software, which was just doing a proof, has to expose some function that, you know, I'm not really using
Otto Mora: Okay, uh...
… Steven, and then I'll just have a wrap-up, but yeah, because it's you
Stephen Curran: Okay, um, one of the things that got me on the… on yesterday's call was the...
… removal from the introduction of what is the purpose of did resolution spec
… And I think, Joe, I think you just repeated it. I… that's what I
… object to. I think that second paragraph in the current introduction should not be changed. The purpose of this, of the spec
… is to say, hey, you get a DID URL, and implementers can know what to do and return consistent results. And I think. changing the purpose of the spec
… um, is not a good thing. So that's the biggest thing I'm worried about in the descriptions I've been hearing
Otto Mora: Okay. Joe, and then I'll close it up, yeah...
Joe Andrieu: Okay, um, yeah, I… I… without being offended, I take technical issue with the...
… The spec doesn't say that we're not going to define what you do to dereference. We absolutely are doing that
… The reason that intro changed is because of this issue of removing the dereferencing function. If we are not going to define a dereferencing function that has an explicit input and output
… then we are not defining, uh, the API for a dereferencer, which is what the opening says. So
… the new language in this PR is just aligned with that other decision, which is obviously contentious, so
… They can be and should be aligned with whatever we end up deciding
Otto Mora: Okay. So, okay, so I think maybe… so we can do, uh, like, give people till end of day to plus one, minus one, their comments here...
… Uh, on the next call, we can try to maybe have
… I would say maybe 2 or 3 PRs at most. From this breakdown. Um, but
… Then maybe do something similar for Steven's, and then maybe even consider some of Marcus' PRs
… Maybe we can have something like that, but, uh, Will, you wanna
Will Abramson: Yeah, I just want to say, I don't think we need to the end of the day. I think maybe we can share this with the DID Working Group mailing list, and maybe put a blurb at the top...
… like, I think people should take their time, because obviously… and, like, if you don't understand any of these points
… Uh, it would be useful to, you know, maybe you can just put a comment on them, like, say, I don't understand
Joe Andrieu: If I may, Will, I'd be hesitant to broadcasting?...
Will Abramson: Oh, okay...
Joe Andrieu: Um, these 14 things have been in the conversation for a while now...
<TallTed> knowing they're misunderstood is its own challenge
Will Abramson: Right...
Joe Andrieu: So, these aren't new. This was… it's the conversation here that was what was important...
Will Abramson: Yeah...
Joe Andrieu: Like, I fear that someone who didn't get this pushback and feedback and engagement...
<ottomorac> +1 Ted
Joe Andrieu: They're just gonna have no more information than they had in that PR in the first place
Will Abramson: Right, right. Fair enough. But still, I would give, you know, like...
… I guess, Otto, me and you can look at this on Tuesday
Otto Mora: Mm-hmm...
Will Abramson: So, people can take their time, and then we'll use the, um...
… stuff from this document to determine the agenda for both Wednesday and Thursday. And the last thing I want to say is, Joe, it would be super useful as well, and maybe other people, like, if there are things in here that
… that didn't get adopted, right? We've got, like, plus one from you, like, if there are things that, like, absolutely need to be adopted
… Otherwise, you're going to object to the spec in its current state, I bet lots of good… good to know. That's not captured currently. I don't know how we capture that
… Um
<ottomorac> transcriber-bot, pause
Otto Mora: Yep...