Meeting minutes
Agenda
<ottomorac> transcriber-bot, resume
Otto Mora: Okay. Alright...
… So, transcription's on
Feedback from Conversations at IIW
Otto Mora: Uh, maybe let's start with this first one. Um
… feedback from conversations at IAW. I guess, Joe, since you're here
… Maybe you can summarize a bit of what that conversation was
Joe Andrieu: Sure. Um, I'm bummed Marcus isn't here. Um… uh… We had a good conversation. Um...
… Some of what we talked about seem to move towards potential alignments
… Um, in particular, I proposed that, um
… You know, sort of towards the end of our conversation, that
… Uh, it may be a suitable option to define dereferencing independent of the dereferencer
… And if it is valuable to support the existing remote dereferencers out there
… To, um, define that as a separate thing
… Um, rather than having a dereferencing algorithm that depends on something that has a dereferencing signature
… Um, that that may be a good way to support what Marcus is pushing for, is that some people have these dereferencers
… Um, and they want them to be part of the ecosystem and how you use this
… Uh, so that was one proposal that was put on the table
… Um, but Marcus was gonna sleep on it, so Marcus wasn't quite ready to buy into that, which is why I'm bummed he's not here. So I can… I can speak to the fact that we talked about that possibility
… Um, but not that we had consensus on it. Um, another aspect, um… That came up was, um
… uh, some understanding, I believe, on Marcus's behalf, that the relative ref algorithm really, um, is a problem
… Um, in particular, he didn't realize, uh, I think until that conversation, that it also breaks the DIDCOM messaging service endpoint
… Like, the ability to use the algorithm the way it's defined right now would break expectations around how he thought DIDCOM messaging would apply that algorithm
… Um… and one of the things I think that I think we've seen in the chat, and, you know, Marcus echoed there, although he didn't make a big point of it
… Um, is he kind of just wants to be done. He's frustrated that we were almost to the finish line, and then there are all these changes
… Um, so, you know, I appreciate that challenge, but I think where we're at, the tech still needs some work
… Um… and with that, I was… I was hoping to hear from Marcus what he thought about it after he had a chance to… To let those ideas settle in
Otto Mora: Mm-hmm. Okay, that's good, thank you. Let's see if the account's race… Appreciate that...
… Um, okay, so… oh, actually… well, yeah, go ahead
Will Abramson: Yeah, I mean, I think that all sounds great, and I did read Marx's message in the chat, and it was...
… useful. I guess, like, I'm wondering how that affects our agenda today, just, I mean
… you know, the topics we wanted to get closure on were both topics that I feel like we've discussed quite a lot, um
… I was hoping to make a decision on them today, but obviously Marcus not being here. I mean, we can just make a decision without him, but it feels like it would be better
… have him involved, so I don't know
Otto Mora: I don't know...
Manu Sporny: Um...
… It would be wonderful to have Markus more involved, but he has been pretty clear that he does not have the time to engage
… At the level that we need him to engage for us to make progress
… You know, standards and specs are made by the people that show up. We are showing up
… We need to make decisions and move on. Um
… you know, and it's up to Marcus if he wants to implement the stuff that we have
… suggested. I mean, I think the biggest problem here is that he does have one of the, you know, um, most popular resolvers, and
… you know, he can just decide that he's not going to implement the specification. But, you know, at this point, not having a specification, I think, would be a worse outcome
… So, we need to, you know, I think we've talked some of this stuff to death
… And if we are fairly certain we know what Marcus's position is, we can make a decision. Um
… I don't wanna… I don't wanna continue delaying decisions because, you know, Marcus is showing up to 1 out of every… and calls
Otto Mora: Okay...
Manu Sporny: I mean, I will, I will also note that a number of us are not being paid to be here. I know Marcus is in the same position...
… But, like, show up to the meetings if you want something done, right? Um
… And I don't… and I don't think, you know, not showing up to the meetings is a good reason anymore. We're so far past our… are, uh
… Uh… we're past the end of our charter, we need to… we need to make decisions and move on
<ottomorac> transcriber-bot, pause
Otto Mora: No...
<ottomorac> transcriber-bot, resume
Otto Mora: Okay, transcription, background. Uh, okay...
<ottomorac> transcriber-bot, pause
Joe Andrieu: Actually, could you turn it back off for a second?...
<TallTed> the IRC log hasn't been written to since 13:56:50 UTC
<ottomorac> transcriber-bot, resume
Otto Mora: Yes, yeah, yeah, yeah, sorry...
Manu Sporny: Alright, uh, can we go back on the record? I want to get some of this on the record. I'm happy to… happy to state it...
Otto Mora: Go for it...
Manu Sporny: Um, okay, so, so… my concerns...
… Our Marcus continues to not show up to these meetings, uh, and it makes it very difficult for the group to make progress when he is objecting, but he's not at the meetings. I completely understand that he is not funded for this
… But I will remind the group that the vast majority of us are not funded to do dead resolution work right now. So we're all in the same position, but a core group of us continue to show up
… So, if we want this thing done, we're past our charter, that core group needs to come together and get stuff done
… Um, I don't think, uh, you know, the, the, the, uh
… current, you know, framework for editing on the spec is working either. Um
… Uh, I think, uh, you know, Joe and Steven, uh, uh, have put tremendous amount of work toward trying to
… get the spec into a particular shape, uh, that is going to work. Um, they have demonstrated that they can be, you know, co-editors, whatever. They can move the spec forward, right? I am also happy to help in whatever capacity I can. Will's also, you know, done
… done PRs, we just need to get the people that are actually doing work in the group in a position where they can drive the spec forward
… Right? And for those reasons, I think, you know, we should list Steven and Joe as, you know, co-editors, or with the power to make editorial decisions, because, you know
… Those things need to be pushed forward at this point
… Um, and with that all said, uh, you know, I expect that this will lead to some amount of consternation by, you know, people that are not showing up to the meeting. Too bad
… that's where we are, right? Like, we're here, let's move it forward. Um, uh… With that said
… let's focus on the things, on the decisions that need to be made today, and let's make some decisions. We have been talking about some of this stuff endlessly, let's just do some polls, see if we've got some
… general path forward that… that might be okay, and start making decisions so that we can start raising PRs that… that put these things. Uh, in… into motion
… I expect fully that we will get formal objections moving forward by people that are not here, and that's okay, because
… We have given many opportunities to engage, and it's just not happening, so
Otto Mora: Dimitri, do you have any particular opinion? And also, if you want the transcription of 2, let me know...
Dmitri Zagidulin: You're welcome...
Otto Mora: Oh...
Dmitri Zagidulin: I was just reaching for the thumbs up button, uh, plus one for everything on the side...
Otto Mora: Okay...
Dmitri Zagidulin: Yes, I totally agree...
Otto Mora: Okay, thanks, Dimitri, appreciate it. So, anybody object to the proposal of...
… the co-editor role being given to Steven and Joe?
Joe Andrieu: For the record, we probably should make it a proposal...
Otto Mora: Okay. Proposal, please...
Will Abramson: Uh, I don't… well, I think we can, but I… Can we do proposals in this group?...
Otto Mora: Or rather, in the main… in the main call...
Will Abramson: I feel like maybe we should queue it for tomorrow. First thing...
Otto Mora: Okay, that's fine...
Will Abramson: Mm-hmm...
Dmitri Zagidulin: So yes, but it's informative, non-binding, but yes, we can just save it for next group tomorrow...
Will Abramson: Yeah...
Joe Andrieu: Right, and we would call it a poll, right? I forgot all that sort of subtlety...
Will Abramson: Well, no, this will be a proposal we'll run tomorrow, just to get it formal...
Joe Andrieu: No, no, no, we should not call it a proposal, because those are magic words...
Will Abramson: Yep...
Joe Andrieu: That was part of my understanding, is that in the task force, this is maybe more from the VCWG, so maybe you didn't hear that, Will...
… But over there, we went through the same question. We have all these task forces, and if the task force makes a proposal, then, unfortunately, to many eyes who are just reading the transcript, it looks like it's the working group working and having a proposal
Will Abramson: Mm-hmm, mhm...
Joe Andrieu: So the suggestion was, hey, just do a poll, don't call it a proposal, and then the proper working group can take the poll as input. To a proposal...
Will Abramson: Yeah, but I think we don't even need to do that. We will just run a formal proposal...
Joe Andrieu: Sure, that works for me as well. I had forgotten about the… That whole jam...
<ottomorac> transcriber-bot, pause
Otto Mora: Okay. Hold...
<ottomorac> POLL: Joe A. and Stephen C. to become co-editors of the DID Resolution Spec
<manu> +1
<Wip> +1
<ottomorac> +1
<JoeAndrieu> +1
<swcurran> +1
<ottomorac> transcriber-bot, resume
<TallTed> +1
<TallTed> <dmitriz> +1
Otto Mora: Yeah. Well, not for… I'll turn it right back on. So, okay...
Will Abramson: Yeah, yep...
Otto Mora: So, done, okay, cool. So, Paul, uh, shows majority evidence for agreement. With the… Idea of, uh...
… doing… having Joe and Steven as co-editors, and we will discuss more formally. As proposal in the next meeting. Okay. So, with that, uh
Summary from last week's call
Otto Mora: being done, uh, from our, uh
… Last week's call, just would like to summarize here, and… We, um
… Talked about the refactoring of the dereferencing function signature and the associated algorithm
… Uh, we also did debate replacing that concrete function with more flexible
… Uh, algorithm that could define some minimum outputs and outputs
… And then we also talked about service path
… I know that Steven had submitted a detailed proposal for path handling. We also talked about, um, doing smaller PRs as a more. Let's say… Manageable way of managing… of
… affecting change into the spec, because if we do just really large PRs, that would not be as easy to get agreement on the group to merge them in
… So, Manu did provide some plus one to that, and agreement on that, and then
… Yeah, I think that was, uh, in general, the conversation was some
… clarification of some doubts that we had internally. But, uh, yeah
… Anything I missed, or
… Uh, Will
Will Abramson: Yeah, it was just, uh… it was a good conversation. I think we focused on the dereferencing algorithm, that's why we decided to put...
… The dereferencing algorithm as the only topic that we're going to talk about tomorrow on the call
… with the aim, I think Manu suggested it, to… to get to… okay, we want to refactor this algorithm. What are, like, the high-level steps in that algorithm? And then
… what are the inputs and outputs of those steps? Or, like, what's the purpose of them, and then we can get those as PRs into the specs. So we have a framework for other people, like, future PRs, to go into those algorithms
… and define the steps. That was the idea behind it, to get something into the spec that starts to move us
… in the direction that I think we all want to go in. Uh
Otto Mora: Mm-hmm...
Will Abramson: So hopefully that's tomorrow's conversation...
Otto Mora: Yep. Okay. Uh, Steven...
Stephen Curran: Is someone gonna lead that meeting, and… or could someone volunteer to lead that?...
Will Abramson: I can… I can lead it. I'll take a look and put together. Some steps, if you want...
Stephen Curran: Good. That would really be appreciated. I think that would be the best way to go...
… Thanks, Will
1 - Remove dereferencing function signature
Otto Mora: So, with that, next topic is...
… Uh, topic one, which is remove the referencing function signature
… Uh, anything in particular we want to align there, I guess, prior to our call tomorrow, I see Mano
Manu Sporny: I mean, plus one, let's remove it. Um, but I do think we do need… still need the algorithm, and that algorithm needs inputs and outputs, minimum inputs and...
… Minimum outputs, and then we should say that, uh
… You know, you can implement anything you want to, as long as the inputs and the outputs are the same as the algorithm provided
<Zakim> JoeAndrieu, you wanted to say outputs are a challenge
Otto Mora: So...
Joe Andrieu: Yeah, I think the outputs are a challenge, Manu. Um...
… I think there's likely to be effect on program state, um, but I don't think we can determine it in terms of outputs. Um
… you know, the short little scriptlet that I've shared in a couple meetings now. Um, doesn't need a return value. It's just changing the state of the application
… So… um… I still think we have that… that issue to deal with
Otto Mora: Even...
Stephen Curran: I think we've… because the resolution. It has a signature on it...
… Um, and is… and they should be binding that
… Creates complications that we're gonna have to deal with, such as
… Um, resolution metadata has to be taken into account. And that's gonna get
… tricky when we don't… if we're… if we're saying, oh, you just have… like, the ideal would be you just have a URL as the input
… And that's it, but because of what we've done already, and existing extensions and things, we can't do that. I think that gets a little tricky
Otto Mora: Right...
Stephen Curran: So, for...
… mine and Joe's comments, other than the inputs and outputs, we can just have inputs and outputs, it would be great
… That was a joke, the last one. Attempt
<Zakim> manu, you wanted to note functions with side-effects aren't great.
Otto Mora: Manu...
Manu Sporny: I'm on the queue to say I'm chuckling. That's fine, Joe. I mean, I don't like functions that have side effects, like, that's typically a bad code smell, but if people want to do it, like, you know, go right ahead...
<swcurran> +1 to manu
Manu Sporny: Uh, the return value could be null. We could say, like, you know, you don't have to return anything, and you may return, you know, these values. Uh, doesn't quite necessarily lead to good interop, but, you know, I mean, I think I'm at a point where I don't really care
… About, like, you know, like, if we can't force people to do
… things that they don't want to do. So, if we've got an implementer going like, I don't want to return something from this
… And we can say, like, yeah, okay, fine, like, you don't need to return it, you know, here's what you should return, but, you know, that allows you to kind of break that
… Um, or not do that, Joe. So, you know, this may be just a should versus may, uh, language
… Um, the same thing for options as inputs, you know, um, should versus may. URL?
Otto Mora: Can you repeat that last part, sorry? Same thing for inputs and outputs...
Manu Sporny: Uh...
… Uh, you, you, we can
… Specify shoulds and mayes for input options and return values
… So that would address Joe's thing. If Joe's like, I don't want to return anything, then we can say, fine
… Uh, you should return, uh, you know, what markets wants returned. Uh, but you don't have to. Right? Um
… Now, how you demonstrate, you know, interop with, uh, internal program state change, Joe, I think is a challenge, but, like, you know, it's
… It's a should, so we don't even need to check the return value, which is… which is not great. Like, we can't do that, like, our test suite for that is gonna be like, yeah, I called the function
… And that's it. Like, we're not testing any kind of return value from the function, right? And that means that I would imagine, on the whole
… Uh, it leads to a much worse interoperability story. Um
… But there are ways to, like, you know, there are ways to negotiate our way past this, like, you know, we don't have to take a hardline position on any of this stuff
Otto Mora: Well...
Will Abramson: Yeah, I think I agree with what Manu said. I mean, one way to look at, like, Joe, your example...
… function, right? Like, it does some stuff, and it dereferences to. a verification
… method. And then it uses that method in the function to do a verification, or whatever. I think that was the example
… But it still does dereference to a verification method in this case, so that, in a way, like, depending… it's where you draw the boundaries, like, that could be the output of the
… of, like, the algorithm that we're defining here, just your, like, the code boundary that you've drawn
… Uh… is not… is not in the same place. But I think you could still say
… I think we talked about this last week too, like, I think you could still say that does have
… Uh, and output is just being consumed in the same place that it was. Retrieved. Okay
Otto Mora: Uh, Joe?...
<Zakim> JoeAndrieu, you wanted to say resolution != dereferencing
Joe Andrieu: Yeah, I don't think we can say that function has an output. I think it does stuff, and there's… there is a school of thought in programming that functions that have side effects are bad ideas, and it is certainly one way to think through how you're going to architect your program...
… But ultimately, what we need interop around is resolution
… Um, this is sort of the whole point of my frustration with the introduction of this thing that is now a dereferencer, which now introduces a whole other set of interfaces that we are talking about interop, and does the input and output of this thing
… map how people are going to use it. For example, we have… we had in that interface that people might get back a list of URIs stripped of all context
… Right? That doesn't make sense when you need the context to decide which URI to actually dereference
… And so, sort of forcing everything through that straw of that output created a problem where the recipient didn't have the metadata to understand which URI refers to which service. Like, that wasn't even in there
… Um, so… I think our focus should be on how do we clarify how a client calls a resolver?
… That is the point where we were trying to get interoperability between the necessarily unique things that different did methods do. We have, by architecture, isolated that into
… resolution. And so, we need that formalized. That needs to both be able to be over HTTPS, we've talked about that, and it needs to have explicit inputs and outputs, because that's the wire
… That we are trying to say that can go to any one of these conformant resolvers
… to lift that burden of multiple conformant dereferencers is what I'm
… that's what I'm trying to get off our plate, because it's an unnecessary range of debates and arguments about what is the right inputs and outputs, and I'm suggesting we don't need them any more than my browser needs specific standard inputs and outputs on how it dereferences
… It's going to execute the algorithm, and we need to describe the algorithm that is executed, so we knew what… we know what to do before we resolve, how do we prepare for it, how do we actually make the resolve call, and then how do we respond to the answer
… And I think that is not something where we're trying to get interoperability, except for that resolve call
… Um, the other things are context-specific. If I am, um, uh, doing analysis over an RDF graph, and I come across a DID as a URL
… What I do there is going to be fundamentally different than what I might do in verifying a proof, or trying to display some resource in a browser
… So, I don't think that interoperability is around, sort of, that component of what you put into and get out of a dereferencing
… It is how you dereference in the scope of how do you prepare, how do you call, and then what do you do with the response. That's it
Otto Mora: Steven?...
Stephen Curran: So, I wanted to agree with, um, Will, and what we're trying to do is figure out what is the, I think, the fundamental agreement, and I think Joe just...
… or disagreement, I mean, and Joe just laid it out there, which is, where does the spec end?
… Um, I don't think it ends at the end of resolution. I think that's the easy part
… I think much more interesting is having consistent meaning that when you, for example
… Put a fragment on the end of one, not only does that identify
… a portion of the JSON LD, but the type of that thing
… defines what you should do at that point. Now
… What you should do varies, I think, in this group as in
… Is that the dereferencer that's supposed to do that, or is that the client that called the dereferencer is supposed to do?
… But I think that… coming up with, okay, how far is it supposed to go, I believe… I want to know
… I want to be able to put a did URL into a… as an identifier, for example
… And I want consistent behavior out of that, and I want the, um
… somebody implementing a DID-based system to know, oh, they got a DID URL was put in, and I'm gonna get a consi… and we're gonna have an interoperable way of interpreting that and getting something out
… And so I think it should go further than… than simply resolution
<Zakim> manu, you wanted to note we're talking about dereferencing function signature... and to say what core thing does Joe/Stephen want to talk about?
Otto Mora: Uh, yeah, man...
Manu Sporny: I feel like we are not sticking to the topic that was...
… under discussion, which is fine, but it shows us that we're talking about the wrong thing, so the topic was the dereferencing function, and whether or not we define it. It is not all these other things that we're talking about right now, so
… What is the core thing that we want to talk about and get to agreement on? It sounds like it's resolution, so let's just build from the bottom up. What are the inputs to the resolution function? What are the outputs? Like, it sounds like, Joe, you really want to talk about it
… Steven, you want to talk beyond that, but, like, can we just agree on the
… function signature for resolution, then, I guess? Like
… I'm frustrated that we're jumping away from the topic that was explicitly set by the chairs
… So, what do we want to talk about? What is the core of the foundation that we need to build on?
<Zakim> JoeAndrieu, you wanted to correct that I did not say the spec ends at resolution
Manu Sporny: And let's go from there. It sounds like it's… what is the function signature for resolution?
Otto Mora: So...
Joe Andrieu: So, Manu, I think that's changing the topic. I think, actually, Steve and I were arguing about what is the boundary of dereferencing...
… Um, the, um… What did… what did I put in my tag?
Otto Mora: Um, so you had said about the, um...
Joe Andrieu: Oh, right, and, and, um...
Otto Mora: The spec end, yeah...
Joe Andrieu: To that point, Steven, you had said that I was advocating that the spec stops at resolution. No, we have to explain what you do after resolution...
<swcurran> +1 joe
Joe Andrieu: Maybe I wasn't clear about the three steps. We need to explain what you do before resolution, we need to explain how you call resolve, and we need to explain what you do after it
<manu> sounds like we're in violent agreement?
Joe Andrieu: So, I'm absolutely not saying that we stop at resolution, but I am saying that what you do with the result depends on the media type of the results
… And so, if the result is an image, you're going to deal with it differently than if it is a node in a JSON-LD file
… Um, so we do talk about how you use it, and we need to talk about that, um, but I don't think we need to specify any further API coming out of that
Otto Mora: Well...
Will Abramson: Uh, yet. So...
… I think what you said there, Joe, you know, like, you said something like, uh
… we're gonna do different things depending on the result and the media type of the result, right? Like, for me, like, I think the boundary that we're talking about is… is that result. I mean, yes, we can say, obviously, depending on the media type of this result, right, this result. You are going to handle it differently
… But, like, I think the bit that we're defining is you call resolution, you get a resolution result
… Then you do some stuff with that result, and you get a resource, and here are the ways in which you can do things with that result to get a resource
… And… depending on the media type of that resort, obviously, you're gonna handle that differently. Like, some way to say is, like, that resource is
… quote-unquote, the output. And then just what I was actually on the queue to say is, to Manu, I think there is agreement that we want to get rid of the function signature. Where this conversation really started is
… The, you know, like, we're getting rid of this function signature, but we're still defining an algorithm, and is that algorithm having inputs and outputs?
<swcurran> Maybe we need a sample set of DID URLs that we talk through to say where the spec ends.
Will Abramson: Um, I think it's really about the outputs and, like, you know, the end of the system, I suppose, is what we
… got to discuss. And maybe the question that Manu raised at the start is, like, is it okay if we say shoulds to all these outputs? And maybe that's just a good way to move boards, right? Uh
Otto Mora: Mm-hmm...
<Zakim> manu, you wanted to ask what is the call to resolve? What do you get back?
Otto Mora: And then, uh, Steven's saying that we maybe sample set of the URLs, like, uh, sort of, like, scenario. That we can react to, right? Uh, mano?
Manu Sporny: Uh, I don't know what we're talking about anymore. I'm sorry I'm being… You know, dense, but...
… Do we have a… it's… are we getting rid of the func
… I'm having a really hard time figuring out, like, what we're talking about. Um, do we agree that we can get rid of all the function signatures in the specification
… And just have algorithms that have inputs and outputs
… And specifically, for Resolve, it sounds like we agree, at least the input to resolve is a did URL
… And then maybe you have some options to it
… And the output from that is a DID document with a single media type
… and I haven't talked about dereferencing yet, but, like, do we have agreement on that?
Will Abramson: Uh, yeah, I'm on the computer part. I think… I don't know how we've confused you, Manu, but I think we're not talking about that at all. Like, I think there is agreement from the… we're not talking about the resolve function signature. Like, for me, that's staying in the spec...
… Because that is what we want people to conformantly bind to, and it has inputs and outputs. Actually, I think one thing that we maybe don't have agreement on there is whether that… one of those inputs is a did URL, but I think we
… probably everyone in this group does, apart from Mark, but I think the thing that we're talking about is the function signature in the dereferences in the spec today
… there is a function signature that defines dereferencing as inputs and outputs, and I think there is consensus to remove that function signature
… Uh, and just define an algorithm
… And I think the debate that we've been having is about that algorithm and the inputs and outputs that it should, or… What's that?
<Zakim> JoeAndrieu, you wanted to say NO. We don't agree with getting rid of all function signatures
Otto Mora: We could, POLL, for that, but before that, Joe...
Joe Andrieu: Uh, yeah, we'll mostly just set it. Um, I don't know why it feels slippery, like we're talking about resolution, uh, Manu...
… Um, I think we are all talking about what do we need to specify for dereferencing?
… Um, and, you know, whether that has HTTPS or a function signature is about the dereferencing algorithm
… And it is not about removing all function signatures in the spec. That would… I don't think anyone's proposed that
<Zakim> manu, you wanted to note that Markus disagreed with some of this -- that's why it feels slippery to me.
Otto Mora: Okay. Uh, man off...
Manu Sporny: Yeah, the reason it feels slippery is I've seen Marcus disagree with some of these things, and I don't know who else in the group agrees or disagrees with him. So, if we can… if we can build up to, like, I'm hearing people agree with each other...
… and I want the poll to get some of those things locked in, because it helps
… at least me know what we're focusing on. If we're just, you know, if we're totally fine… like, for example, Marcus was like, no, the input to the resolution function is not a dead URL
… Do we agree with him or not? Like, that to me, in my mind, is not locked in
… Right? Um, I'm fine with there being a function signature and an HTTP binding and, like, all that kind of stuff, but, like, to me, like, everything feels like it's up in the air, because there has been so much discussion. I don't know where things are right now, so
… I would like us to… I don't know if those other things are locked in
… Right? That's fine if it's a different topic, but, like, it feels like we're touching on that stuff
<Zakim> JoeAndrieu, you wanted to say that's a different topic
<swcurran> 100% it is a DID URL. <did>?versionTime=, <did>#<frag>
Otto Mora: Mm-hmm. So...
Joe Andrieu: So, you know, other than trying to support the unspoken things that you think Marcus has concerns about. When he's not here. Um...
… Uh, that's about resolution. That is what we're trying not to talk about. So, there are things we need to talk about
… But right now, that's not what the chairs have put on the agenda
Otto Mora: Steven is saying, 100% it is a did URL, did version 10, did...
… Okay, um… Yes, madam
Manu Sporny: Okay, so I'm… I'm fine with focus, it's just...
… For… for the dereferencing function
… Do we know what the… we're saying the… what are… what the inputs are still a did URL? Or is it a did document?
… And then… are there options?
… Are they mandatory, or… or, like, a should?
… And then… what's the return type? I just want to know what, Joe, you're advocating for, and Steven, you're advocating for. It's really difficult for me to do that, having read, like, all the different things that have been said over the… Past, you know, months
… So I think I just need… it would help if I had concrete proposals on the table. A versus, like
… you know, breaking each thing down. Like, Joe, what is the ideal state? What do you want to see? And then, Steven, what's your ideal state? And then maybe we can. just
… See if other people agree with that or not
Otto Mora: Okay, before we do that… Would you agree that we run this, Paul?...
Will Abramson: Uh… can I just say something?...
Joe Andrieu: I'd rather respond before we do that...
Otto Mora: Okay, cool...
Will Abramson: Oh, that joke, though. Uh...
<Zakim> JoeAndrieu, you wanted to there is an algorithm that addresses this quandary. can I get suggested edits to that? w3c/
Joe Andrieu: Um… Sorry, Will, did you send me?...
Will Abramson: No, no, you go...
Joe Andrieu: Okay. Um...
… Uh, Manu, there's… there's a PR that has this algorithm in it
… There are, in fact, 3 PRs that have the algorithm I'm proposing in it. Um, this one was specifically just to
… Uh, update the dereferencing algorithm
… And so, the only input you need for dereferencing is a did URL. Within the dereferencing algorithm, you go and get a DID document, and then you do something with it, but that is not an input
… It is something that you do during the dereferencing algorithm, and there's spec texts there that should walk you through all of that, if you have more questions
Otto Mora: Okay, uh, Will… yeah, go ahead...
Will Abramson: Yeah, I mean, I was gonna suggest that… I think that's tomorrow's topic, almost. I was just hoping we could decide on some of the...
… like, easier things, like, we can talk about the algorithm, but I think we can first say we are going to remove this function signature. I think everybody agrees we're going to remove the function signature. The bit we're not so clear about is whether it's going to have
… inputs and outputs into the algorithm, and maybe where we're getting distracted is what those inputs and outputs. Should be… but
… That's part of defining the algorithm, which I… Oh, it was tomorrow
Otto Mora: Mm-hmm...
… Okay, let's just run the poll now, I think that's fine. Um, one sec
<ottomorac> transcriber-bot, pause
<ottomorac> POLL: Remove the dereferencing function signature from the spec and replace it with an algorithm with inputs and *outputs*. Outputs are all optional
<JoeAndrieu> +0.7
<Wip> +1
<manu> +0.75 (there is no interop on outputs if we do that)
<swcurran> ¯\_(ツ)_/¯
<TallTed> -0.2 (not blocking, it just doesn't feel right)
<ottomorac> +1
<PDL-ASU> +1
<ottomorac> transcriber-bot, resume
Otto Mora: And I think...
… Uh, Will first. Thank you
<TallTed> <dimitriz> +1
<Zakim> manu, you wanted to note you talked w/ markus, did anything change?
Will Abramson: Well, I think it would be great to hear from Chad, if Chad doesn't mind voicing what he doesn't… Like, though?...
Otto Mora: Oh, Ted, Ted, sorry...
TallTed // Ted (he/him) Thibodeau Jr (OpenLinkSw.com): Alright. To my mind, um...
… the impact of
… The function signature is not what
… People feel it is, and it's an easy way to describe an algorithm and its inputs and outputs
… Um, that is, to me, what a function signature is, but
… There are other people who feel strongly about it, so, like I said, I'm not blocking, it just doesn't feel like the right thing to do, but, you know, I'm not going to stop you
Otto Mora: Manu wants to run the queue before, uh… Okay, go ahead, man...
Manu Sporny: Yeah, I mean, I agree with Ted. I think it's a bad idea to have something that doesn't have a return value. We won't have interop on it, but, you know, I don't care enough to...
… To push it. Like, I mean, I think, Joe, you're gonna object if we… if we assert that it needs to have a return value, right? And if that's the case, then, you know, whatever, let's just
… back off on the return value. I think we should say you should return, you know, XYZ, but… Um… happy to
… I'm not gonna block, right? Just trying to… Get to something we have consensus on
Otto Mora: even...
Stephen Curran: Um, plus one to what Joe said, um, this is the big change, but I think it's… Correct, which is...
… that the input to whatever we do, the input to this specification is a tip URL
… And during the course of dereferencing that vid URL, you get a… Did Doc
… you resolve the DID to get a DID doc, and then you continue processing, and I think that
… Is the big chain. That is why we are having so much struggle, is because we're trying to not
… Separate out the two, and
… that's where the big problem comes, and I know that
… you know, Marcus is gonna disagree with that, um, but… That's what I think. That's it
Otto Mora: Okay. Uh, well...
Will Abramson: Yeah, I think, um...
… I mean, I hear what Tad is saying, like, it is a good way to describe the algorithm
… I mean, I guess the softer way of not removing the function signature is to just remove some of the language, or, you know, like
… In fact, reverse some of the languages, like, you must implement this function signature to… you can implement this. However you like
… Um
<Zakim> manu, you wanted to note function signature isn't buying us anything, is it?
Will Abramson: I think that's all I had, or maybe I had smiles
Otto Mora: Wow...
<TallTed> +1 manu
Manu Sporny: There's boilerplate language for that, Will. It is… it is, you may implement this function in any way you see fit, you know, as long as the inputs and the outputs remain the same...
<swcurran> +1
Manu Sporny: Um… I… I don't… the function signature isn't buying us anything, right? I think we are… did we already agree to remove it? And we're just moving on to, like, inputs and outputs
Otto Mora: Hmm...
Will Abramson: Mm-hmm...
Manu Sporny: I think… I think we're… I'm not hearing anyone disagree. Is that a… like, where are we disagreeing? I know Marcus disagrees, but like...
… I'm not hearing anybody disagree with the current path
Otto Mora: Okay, Steven...
Stephen Curran: I don't think I'm on the queue. Uh, I...
Otto Mora: Oh, my bad. Okay. Okay...
Stephen Curran: 5 earlier, yeah...
Otto Mora: So, yeah, just close the queue, uh… I see Joe...
<Zakim> JoeAndrieu, you wanted to say its outputs, but we can move forward
Joe Andrieu: Um, yeah, I think the remaining disagreement is not enough to, uh, you know, stop this proposal. I think it is about whether or not there are outputs...
… Um, I don't think there are, um, I don't think there should be. I think this is not the point of interrupt
… Um, but I think the devil is in the details, and I think we can
<TallTed> "These are the inputs. These are the (optional) outputs. In between is a black box which we describe as this algorithm, but you don't have to use these steps as long as what you deliver as outputs from these inputs matches what we say."
Joe Andrieu: Agree to get rid of the function signature, and then talk about, well, what does the remaining algorithm look like, and what does it need to specify?
… But just to answer your question, you were… Manu, you were wondering, you know, what is a disagreement? I think it is… you want outputs, I don't like outputs
… But that doesn't mean we shouldn't get rid of the signature. Um, so, I think there's a path forward in the near term
<ottomorac> transcriber-bot, pause