libp2p Team Meetings, 18 Dec 2018

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: libp2p Q1 2019 OKR discussion

Description

This was our Q4 OKR discussion. We will present these OKRs publicly on Thursday 12/18. The finalized OKRs will live here: https://docs.google.com/spreadsheets/d/11GKG1DBRIIAiQnHvLD7_IqWxDGsVdaZFpxJM6NWtXe8/edit

A

To a private I'm, not I, wasn't planning to post this on youtube I'm, just trying to test this process out for the future future meetings, the lit p2p bi-weekly so give me just a second.

A

A

Okay, let's yeah, let's just see how this goes. I just want to see if at the end of the meeting I can I can get the YouTube livestream saved without having to reload it so anyway, this will just be an experiment, but it's not. This isn't well actually I guess this is public in YouTube, but I, don't think anyone, but us is gonna watch it.

A

I just want Matt Matt Joyner in Australia yeah in visa.

B

Seems to be having trouble not getting onto the call he's in the Oh.

A

B

Being like where'd, everyone go okay,.

A

Wait hold on I switched the link like 15 minutes ago. So let me, let me just say an IRC.

A

A

A

Okay, I guess: I messed up the invite, but changing it it's weird. Maybe he copied it to his calendar because it seems like the right. The invite seems right to me, but yeah.

C

The invite says fine to me as well I'm thinking something must've happened. Yeah.

A

I think, okay, maybe yeah calm pitches counter something: okay. Well anyway, hey guys! Oh sorry, about that. I switched to the link at the last minute to test our live streaming capabilities for next next lid p2p all hands so you're you're live on on YouTube I doubt anyone is watching this, but you're live on YouTube. um The reason I want to record this. This format Joyner in Australia, because these morning meetings are like always the middle of the night for him. So we got to get better about making recordings at least like internally.

A

Is that feedback or some is that my feedback ER so I play music I'm.

D

Playing music so hope.

A

You knew no, it's fine, it's fine! Sorry I thought I couldn't hear the music well I thought it might be feedback from the livestream. Okay, all right. um Let's do it. Let's do it guys here I'll share my screen, one sec.

A

um Okay, so here's what I wanted to do so this was my thing on this call. I think that jeaious Luke PAP is in good shape, like okay, our sink pretty clear to me overall project. I think this is what I think we should do as a theme for the year is like try to grow the user base to 1 million users. This is um the goal suggested by Juan and daveed, obviously is ambitious, but I think it's like the right I think I feel like the right direction for us to go.

A

So, as we think about these, you know the actual. Like engineering work items, we should keep in mind that our goal is to kind of grow, grow the user base for the pp. So, but the purpose of this call, though, was to talk about go since I've talked to people individually, I'll just run through this really quickly, and then we just start open discussion, um basically on dhg 2.0. Most of this came out of me and John Heisey discussing socio Johnson.

A

Let's call I, guess: Stevens, not that's unfortunate, uh but yeah me and John discussing this with information that had been fed in into the process by Steven, as well as John's own thoughts, so I'll. Let him talk more about that. Multistream, 2.0, I. Think Stevens gonna do this in q1. We think it's high priority because once well, just as let Pete if he gains more and more adopters, it's gonna get harder to change these really fundamental things like how the connection and shake works. So Stevens thinking is like let's charge us as soon as possible.

A

The swarm host, refactor I think it's language to talk about. Is that? Should that be a goal for q1 and has all these things like service brokering and I? Don't know I put dialer update under there, although I don't know, if that's quite the right place to categorize it. So we let's talk more about that one testing in the demon. This is like continuing the test. You know get to 50,000 nodes.

A

We actually put the 50,000 no thing on the on the overall project tab because it's cross language, but each language is going to have to do some additional work to get to to get to that point, and Jas also has this reflection, I think and then five was like. This is for you, visa, like the testing or whatever. That needs to be done on relay nodes and then I have a another bucket called other.

A

It's just like what else you know should we do the go mod stuff, yeah I, quit Oh, Martin, mustard, added, quick yeah and then what seven, who had seven.

A

Usa say anything about it: yeah.

C

They've been discussions of the etherium Indy theorem community, especially with here in 2.0. Some of the people there, and particularly Pegasus, is one into increments of Els a very fast.

E

C

Aggregation signature aggregation scheme that relies on UDP. Unfortunately, we don't support UDP and a word where API is are not even message-oriented, so we haven't like that's on the roadmap. It's a different topic, but for now we're advising them to be an early adopter of quick, so yeah everything that that is necessary to make quick, finally deploy a ball in the world. I guess would be. You know my ask for for Martin.

A

Okay, yeah I think Martin also because he's in Asia. So basically he can't turn this either but yeah. So we should and.

C

That's Martin also said in one issue that there is a really nice extension that's coming through in quake and the quick community for unreliable for unreliable unity like semantics on top of quake, so it comes. It goes like back up to Greg and young too, like the udb semantic support and reliable messages, and this would be pretty interesting to like ten using or implement in. Let me to be kind of like as as a testbed for for that. You know. Emerging technology, yeah.

F

There are definitely some issues open for packet protocols and I think it would be yeah I think if, like that, definitely seems like something we could prioritize sooner rather than later. Yeah.

C

Maybe that's like a larger goal here that we could start looking at, which is packet or a packet oriented communication in general right and there are some small steps that the community I know that somebody I remember whom had created a mult I think a multi gram protocol for negotiating protocols over UDP and stuff like that. So there is some. You know some I guess offshoots for this, but yeah. Basically a logical would be welcome here. You.

A

Think we should change. Seven to the title is something like a little bit of eSports packet, switching I, just gonna we can. We can undo, but something like this I mean. This is a huge like item I, don't know: if can we do it in q1, but so.

C

The items that I added there were not necessarily geared towards packet: switching okay, so maybe yeah. We do I, don't know this I can that was reading, editing whatever.

C

A

A

Mean what do people think about this overall, like any reactions on honestly the seat? To me this seems like a lot of stuff for q1, but I know what do people think.

A

You're, all in the docs I'm gonna stop sharing my screen. I, don't I, don't know what it accomplishes.

B

No question about about that stuff, which is um there's two things. One is as part of the trying to get adopters in as someone who's like been crashing. Your calls a lot, but a little external but external to loop. P2P.

B

One thing, that's I think interesting is I, think focusing on like what's gonna, make it easy for developers to get moving and something that is lid, p2p, friendly and works with, like both the NGO side and the je s, ID of things and I'm, not sure whether it's external to this stuff, how you guys generally handle this, whether it's you know internal to these things or externally, but how the you know interoperability of like what do we want to do, and when do we want to roll it out and both sets of clients, as opposed to just one like effectively.

B

What's the limp eat of peace back instead of what's the limpidity, the goal impede appear: the j/s, the p2p spec yeah.

A

This is a huge topic. I was gonna. Try to address this a little bit on in kind of introducing things on the call I fretted, but basically there's like two parallel processes that are going on right now. One is this longer term Road mapping thing where I think we probably after this quarter I think we want to stop having this division by languages and try to focus on overall project goals, but in the meantime we need to deliver q1, ok, ours like so that people know what we're doing so.

A

My proposal was that we just used the old process that we've been using in the past of dividing ourselves by language and identifying work items that we're gonna do in each language and then we'll have this team week in Porto in in January which, if even if you want, if you're particularly interested, you can join I, didn't I, didn't invite you because you're not normally the team, but I mean we love to have like people who want to use lib p2p also come talk to us, but anyway we can talk about that later.

A

So then I thought. Okay, basically, we'll use the old process to just get goals for q1, so that people know what to do. You know at 9 a.m. on January 2 and then we'll try to shake things up and align better with the roadmap and finalize that in in January. So so that's what you're seeing kind of so yeah, but there's a lot of issues around.

A

So if you go to the overall project, tab, there's a lot of stuff around documents, so, first the so the first issue with more downstream users is we need to document things better, there's a huge. It is like the number one ask we hear from everyone is about documentation, and then we also have yeah these collaborative items where some of some of these are specs, which is another facet of documentation. You need to improve, as well as the large scale testing.

A

So that's kind of I think the stuff you're asking about is kind of on the project level. Tab.

A

If anyone's missing the link to this document, I'll paste it in the chat but um but just on, go itself. I.

A

Mean any any reactions on this.

D

Yeah one small reaction, so I think we should have some higher priority on the auto-related moment, since that's major feature that were putting out I mean it's more on like, on the iPad hand, side, but it's also a little being so I, don't know roubles to give it a higher priority. Okay,.

A

I moved it up to p0 yeah everything was the p4 by default because of copy/paste but yeah. I yeah. I also agree.

A

Do you think it's phrased the way it's phrased right now relay nodes are deployed interested in the. While does that really capture? What you want do these, or is it more like yeah.

D

That's fine! That's fine! So we should say all the relay notes are deployed and okay well, yeah. It.

F

Would be cool if the you know if the daemon was in a place where it was like the de-facto way of deploying on a relay and like DHT nodes stuff like that we're.

D

Almost there with that yeah.

F

I think so, yes,.

D

Where we responders here, because you started doing something with relay option so once this has been added, then yeah.

F

D

Think that's actually ready.

F

For review so okay.

D

A

Do you think I should add something here on thirty-one? That's the cache is what you're talking about, or maybe.

F

Just a note that, like DHT and and relay nodes, are run via the LAN p2p daemon, okay. Well, it could be an opportunity to for us to like run a couple in JavaScript and a couple and go and try and just collect metrics on them to like see what the long-term health is like. But this.

A

Feels like a different item to me, though, than the three bitter there so yeah like DHT,.

C

We don't I think in the in the Latin lab week we talked about specialized Libby to be nodes in the sense that you know you could run in a very easy manner, a DHT booster or you could run a relay node and so on. This would be it right, it's kind of like exactly there's a special node in a simple binary that you can just draw.

A

um I feared my screen again: do you agree with how what I've written here calling her all in that box.

F

Yep I, it's really kind of just a detail: I guess that the idea leveraged the daemon but I think it's just a good way of like dogfooding and keeping that, like that's like a project, that's actually getting use because it is it like very generally useful and running it in the environment. This will force us to make it more usable and more stable and stuff like that. What.

A

Do you think I should do about priorities on this objective for there just again there P for by default, not because I chose that.

F

I think that would be the if, if running Auto rellenos, is it easier that should be in futuro as well? Okay,.

A

And then I think testing overall is still a high priority. So should I make this like p1 or something the right.

F

I'd say like interrupts: an integration test is definitely a p0 on diable nodes is probably the lowest priority, and then cluster would be the second. So like yeah I thought that, like roughly looks good to me, okay.

A

A

Okay, all right, okay, so we've covered four and five we've talked a little bit about, but in the two sevens seven and eight hey John. Do you want to talk a little bit about DHT? um Because we had this long conversation yesterday, but I don't know that anyone else is aware of it. Yeah.

G

So this is just the home with what I've been talking to Steven about I mean mostly inspired by by his goals originally, but I think we mostly agree on this, so we want a DHT. That is, we want to split it up into smaller components that are easy to easier to change individually than what we have now, and we also want to get some so that's sort of for the larger goal later of of of experimenting with lookup algorithms and such to see how we can we can make it more robust.

G

So that's sort of the the most important goal, and then they sort of that, along with that is, is the other sort of top level.

G

Top level thing is: is making the privacy better, but I I'm, not sure that the that the privacy features that we're proposing is necessarily that high of priority, because it's only only very part of very partial solution and I, don't know what the actual priority ought to be on that.

G

Essentially it means you can have nodes that are hidden, but only up to the point that that a DHT node could keep track. In principle of of someone's identity, it's it doesn't do anything like we haven't proposed anything like tor for now at least.

A

Okay, so maybe the privacy row, we keep us a low priority, relatively sure. Yeah yeah p3 is pretty lit. P4 means this is not gonna happen teacher, but it's there's like a 10 percent chance. That will happen. Okay, so, but on Stephens, original refactor issue I mean. Is it? Is it fair to say that this is like sort of this starting point from which we have to? We have to do this first or do you think we have to do this? One first, the sorry it's not top aligned, but the I.

A

The fact that I would say those maybe should.

G

Be combines it's okay at this point, I think it's poorly. The distinction between like rows, 6 & 8, are not very well defined. Maybe I need to define that better I mean.

A

We could just put rows 6 into row. 8 I just moved the stuff out to the notes here, but yeah.

G

A

Could just put row 6 in a row and then it's not its own I. Think.

G

That would work and then that certainly be the highest priority to.

A

Make this video and then I'll drop my clipboard since I.

G

Mean it's that sort of more the DHT github issue is sort of the high level goal of what we're trying to get, and then these other entries are are kind of how we'd actually achieve that. I would say.

A

Maybe I'd be saying like this.

A

Okay, yes, sir, is it spacing or my stuff? Let's not worry about that right now! Okay, so so is this.

G

A

So is it fair to say this is the main thing you expect to work on in q1 like this this row? What is eight now yeah all.

G

Right yeah, that seems, seems accurate to me and like something useful to work on.

A

Okay, cool, so I'm gonna delete these rows. um We should talk to Matt joiner, though about what, like I, think he can contribute a lot here um yeah. We can't talk to him right now, but I think that if it's possible to break up row with what is now row, five into multiple things like I think he could own some of those and okay.

G

Oh I'm sure we can definitely break this up. I should probably work on on breaking this up. Okay,.

A

Maybe the next you can. Can you try to tackle that as a thing care? Okay, I'm, just an assignment assignment to you? Okay,.

A

It's like all, they assume windows, I can't see. What's going on and it's.

A

Yeah, okay yeah: you can usually get a antic products on the on the you can talk to him after, like for 4 p.m. or something Pacific. So, okay.

E

A

um Works: okay, I'm, pretty sure Stevens gonna do number two I'll double-check with him, but I'm pretty sure like he like he'll. Do it I, don't know what to say about swarm host, refactor or Steven thinks we should do this step by step, but I don't really I, don't know how to break it down, and oh these oak or coal or bees. Oh, can you explain this services thing a little bit more yeah.

F

Essentially, taking we've been taking a look at the kind of notion of wrapping the p2p host in services, essentially the DHT does this and it's not particularly composable, so services are kind of using leveraging dependency injection to create a more composable service environment where services can actually interact with each other and depend on each other, and you don't just wind up having to keep track of like a bunch of objects that wrap the host.

F

hmm So we have had kind of two or three discussions about like what that would look like I think where we left it off was I. Think these, oh you correct me here if I'm wrong, but you you seemed like you were interested in taking like a first stab at sketching out what that would look like.

F

We had talked about, maybe forking dig, which is ubers dependency, injection library and yeah, and that would be kind of factoring like I guess the nice thing is it's it's not even though it's a pretty big change, it doesn't necessarily have to be like a hugely breaking one. At first like we could theoretically move services that don't depend on other stuff over to this framework, well piece by piece but say.

D

That it's priority, there's no known breaking chains, yeah.

A

Do you guys regardless is hyper high or low or medium party.

A

Also, we have used the owner visa.

D

A

You can eject.

D

Medium gray already, because we do because we have a little bit of a mess and the internals wrapping and everything and we kind of want to be able to access the battery services. Time skating like identify and now or not, and so there does give some priority. But it's not. You know like on the critical path, because it could. It doesn't effect right now. Our external visibility at all.

A

Okay, so I'm gonna leave it as p2 with visa, as owner and I'm gonna take out these top roads.

A

A

Oh someone else distract: okay, no necessary, really really. Okay,.

A

I, don't feel this title is really like quite honest anymore. It's not really complete. It's more like.

A

A

I don't know, can anyone propose something better?

A

The PDP some.

C

Please I'm internally, this baby does Freddie relates to my core elements, ultimate be to be here right. It's.

E

C

Is actually the heart open? It be to be everything that's in here.

A

F

hmm I think that's a good title, I think that describes it well. Okay,.

A

All right thanks, yeah I'll, just try thinking something.

A

Okay, I, don't know other yeah.

G

The thoughts now row 26 really doesn't fit under that heading very well. Oh.

G

Yeah yeah I, miss ya, know.

A

That's a good point. Maybe row 26 should actually be just move it to other for now. Okay, well, I was gonna, move it to relay in natural. You think that's wrong and it doesn't fit whether you like the DHT is not related. Yeah, that's, okay! Yeah! That's fair! Oh I lost a lot. uh We.

C

Don't have any any further goals for the demon right, I thought we pretty much future company. Well, the demon at this point, aside from you, know the odds. Turning for requests that we have I think.

D

We're looking pretty good the only things that I would really like that is the validator for pops up, which is a little complicated, but we can do that was probably a DeRozan okay.

A

Should I put this in here.

C

Oops validator in the demon I just I just have a keyword, demon, they're just early and.

D

Also, the notifications, we need the connection. Ratifications are very useful for implementing actual applications.

C

Mm-Hmm yeah that make some other sense here. Yeah.

F

That does that fall under like extending the core protocol for the day. Dana I think we.

D

F

D

I need to make some extensions to add both the validator and yeah. The next notification sounds.

C

Good, actually, that's that's super important right because, like if the goal for it, the goal of having the demon is that you can implement protocols in userspace, then you know many of those protocols would be determined by connection advanced right.

F

C

Space, then you need to have those those connections, yeah.

D

Let's add the connection notification for the demon into your cares.

A

Yeah you want under demon, or do you want under this these pori factors.

C

So what I'm thinking is.

A

C

We can create a section for the demon and group these three points related to the demon under it. Okay,.

A

So this this number four was supposed to be a section for the demon. Well, I, guess: I, don't draw a strong distinction between testing and the demon, but now you guys are starting to do that break it up. I think.

C

There, the there are different things: the testing heavily uses the demon, but the.

F

C

An entity in itself has yeah.

F

Testing is it's a consumer so.

A

Let's try this way so, okay, what what here, let's make this just Limpy TPS Chester ball, let go, let me be accessible and then which things should stay in here. You.

C

Say: go Libby to B's testicle in a distributed fashion or something like this yeah.

A

It's good and then which things stay in here: 23, 24, 25 and then yeah.

C

A

Minutes over okay, yeah so and then okay, so then delete these okay and do you? How do you feel about this PTO? We got a mom and we missing.

C

Two items that we added previously right for the demon would you.

A

Have a connection I think.

C

A

Because I put them in the wrong places, hold on sorry yeah, no, no yeah, here, okay! So let's move this. It did.

A

And then, where is connection notifications.

C

Got lost in translation, okay,.

A

All right, let's still just try again, okay.

A

A

F

Just extend that to be notification. Support because there are, there are other notifications as well and there's probably it's probably worth keeping it open. I.

G

Think this actually has quite a bit of overlap with the D H TV factor of hope. You know one of the goals that Stephen has for that is being able to put DHT custom, DHT functionality and use your space as well, so I think I should figure out what that's gonna look like roughly that's.

C

Interesting, do you have do you have like specifics, for what kind of functionality would go ahead and use a space.

G

mmm Not really I'm not sure if Stephen has a particular use case in mind, or if this is more of just a a high-level goal. I mean, from my perspective, the goal of refactoring. It is that the the peer routing and content routing systems that we have now it sort of each be their own module, but potentially with the goal that that would be easily extensible and you can add more later, but I don't know if that's aI, don't know. If that's you know well justified.

C

Yeah, so what's what's interesting is that I I don't know if there's a use case where you'd be using peer routing. That is not related to fear discovery right, because we tend to call it routing, but in most cases is used for discovery right right.

G

No I did peer. Routing is always going to be a more core thing. I think it's the content routing. It is potentially more of a separate module. I yeah, that's misleading I, don't think I think peer yeah.

C

The reason why the reason why I said that it's because we have created an interface for a peer discovery and potentially that part of the DHT could be encapsulated behind that interface. If you want to like break and split the DHT into several pieces, sure you know is that is.

G

That the same as the the existing peer routing interface or is there something new yeah B? So it's very similar right.

C

Because we still haven't added the ghost service while they cater this.

D

Case rather I mean basically offers a different interface, but it gets implemented through content routing from the dates date. So there is an adapter to turn the interface to take a condom, routing object and give you a discovery, object.

G

Okay, I should look at that interface. That also.

C

Do I'll send your link to the github issue where that Thanks.

A

You, okay, so maybe I'll just leave this here for now, but maybe you can clean this up John along with this one sure sure.

D

Another thing that is not in the rider right now, but I think might be important, is that there was a discussion about learning the changing the d8 steam to face the provider interface to use multi classes instead of see IDs or even put into the swings I. Don't know what it started with that, because you know kind of like ties together with C ad overhaul work in ipfs, so I don't know, what's the status with that, but man we might have a night where this will have an item for each.

A

Okay, oh no I, don't! Okay! Is that a is that a reasonable, particularly saying I. Think I just wrote what you said.

D

Lessons right, you know, we don't even know whether we actually want to do this taste and whatever usual teahouse is whether or whether to use for things, because it's kind of like backwards breaking chains and it's gonna be a bear to deploy this one. So I don't know. We really need Stephens name put on that. One cuz he's the one driving for the change. Okay,.

B

Lisa, would you be able to give me like the the TLDR on on why to change it to multi hashes instead of C IDs.

D

There is an open issue: you can just read it there: okay, so basically they're the TLDR is that those are changing. The multi base, 40 IDs and the all Shady's are raised 58, and they will change it later to do, which is gonna, break the keys when they make the chain so they're going to make it so that it takes. The movie passes the reacquaint set of the CDs or its base independent, okay,.

B

A

I'm just looking at what we have okay, so okay, so you seem pretty committed to these two rolled you so should we just leave these others and you'll work on? Okay, all right, I, don't know what to do about packet. Switching honestly, it's really big. It's really big endeavor I mean I. Think it's important, but are we gonna really get to a key one.

C

We definitely I, don't think that goal would be to implement I, guess anything related to that. But I think that we need to start thinking about it, and maybe we can have some goals in there that are related to like potential designs, which are discussions even if it's just to like identify. You know and breakdown rate, this further down into actionable. You.

F

Know I think, like some design and interface sketching would be great. hmm Some general progress, I would say yeah.

A

I think this is why I think this one's important to in the long term, so I like this idea that for q1 at least, we agree that we'll we'll have some design discussion.

C

And you know by q1 large should be back as well and I think he has also been a longtime supporter. Okay, also.

A

Yeah he's very interested in this yeah so so who does assign.

E

It to him says he's.

A

No I'm I'm, considering that but I hate to do that. Well, no I think he actually would would be on board for it. But I thought I, don't want to speak for him. He could come back after his vacation with a totally different mindset. So we don't know he said. Would anyone be willing to be an interim owner of this I suppose I could be? If no one else will do it, but does anyone want to.

A

I, don't I can't see you guys, but I don't hear anyone. Volunteering right here right here, I'll own the driving some discussion on this and that's cool I, would like to participate on some level. Yeah yeah I mean every everybody'll participate. I think like Porto would be a good place. I did invite him before I haven't heard nothing that I'm hoping he'll come yeah, but but we can also discuss some. You know their channels as well.

A

Okay, this I guess this seems like a good list of the only category that feels a little ambiguous or uncertain to me is the other section.

A

I, don't know what so so ro. What do you think about go mod like is, that is that too massive of a wreath, a current, is I. Think.

F

It's a p0 personally.

C

Oh, it's definitely it's definitely zero.

C

We shouldn't be pushing it any further. There is so there's some interesting discussion that is, that is taking place in the ipfs community that relates as well to go. Mod GX and different people are taking different positions and voicing out like different. You know very valid points, so I think when we watch that for now see you know what the different opinions in in PL are regarding to that and then we'll add to core degree.

C

I think Steve Allen was also interested and he had been sketching out a proof of concept for Sri centralizing into for centralizing interfaces across all these repositories in a single one. So that is definitely something to keep an eye on, and maybe you can I'd that add that there as a proof of concept as as a task within other, as well so below Gomory Factory in a new line yeah, maybe a proof-of-concept would be like, rather than go.

C

Mod refactoring out like in you know, in an explosive mode, I would just say kind of like a proof of concept would be the first goal. Oh.

A

Okay- well maybe, but maybe I should just add it. The would this make more sense.

C

Yeah, okay, I, would say refactoring yeah the the proof of concept coherent with whatever you know. The discussion is there and I think we does it b2b community. We should also participate in that discussion. I haven't had the bandwidth personally, but we definitely should is.

A

This is this: the discretionary frontiers there's a separate.

C

Discussion: okay,.

A

Cool all right, um are you, okay, owning that I, don't know yeah you did you make yourself on earth? Did you know I've made some noise regarding that, so it's.

C

A

Very supportive of it I I, hope we go in that direction. I'll read what people are saying on this issue, but yeah takes.

E

A

All right so Martin just dropped in quick goals without any oh hi, sorry that was me and it was.

C

Okay for us to remember to reach out to him and ask him what he wants to well, he thinks his okay one and taking account like all the discussions with me here, community as well.

A

Just so I'm just gonna ask him in the comments.

G

Just very briefly, what do you mean by putting the interfaces in one repo? Oh.

C

Yeah so right now we have all like each and the interfaces behind each component of let p2p it lives in a repo of its own, and this is makes it really hard for a newcomer to the project to actually, you know explore what is over. The p2p are are all about, because they're spread across number of rebels and if you look at the history of those reckless, it doesn't really like.

C

They haven't really evolved independently as much as one would have expected at the beginning, and that would have justified putting them in a separate red book right. So that makes a lot of sense. So what so, as a kind of like? Ideally, we have been some. There has been some discussion about potentially having you know, putting the p2p all together into a mono repo.

E

C

Know where that is gonna be headed. This is like very you, know, initial discussion, but as a like, as a proof of concept, step Ali wanted to wanted to at least try centralizing and grouping all the interfaces in a single report. So that way, we'll have I call the abstractions in a single place and then the different modules, the different implementations and alternative implementations for transport for network for DHT and so on would be was still live in separate rebels. Okay,.

G

That actually sounds really nice to me, from the perspective of when I was trying to learn how with p2p works. That would have been so much easier.

E

C

Yeah yeah I agree with that as well, and the community I would say would strongly agree as well. At least you know the community that the people that we spoke to in in for.

A

Yeah, that is I definitely agree with that.

A

A

Do people feel good about where we are on this.

B

F

Oh, that was all I'm just saying, I. Think I think we've got a pretty good spread. What were you gonna say? I didn't.

B

Yeah so I wanted to know about some of these things that are not here, but that some are like on the roadmap or work is happening on them elsewhere.

B

So one of the ones that's on the road on the roadmap is something about like offline message queuing and we, the dynamic data capabilities group and Steven, and some other folks spoke with the textile people and they have their own version of some of these things, and it may be worth putting in some time whether it's you know it's someone like me or someone like slightly outside that can help you. You guys have a little more bandwidth or internally.

B

That just says: let's come up with a bit of a plan here on what we want, so that we can figure out how to get help from partners to increase our bandwidth.

B

You know it's, it's I think that one's like more apparent than the or more more obvious in a sense than the you know, message passing our PC kind of thing, which is just helpful to get people involved and get moving or is off. My message, cueing, is something people are already doing in a sense or want to be doing.

A

Okay, I, like what you're saying can I make you the owner of which are what's your github. Well,.

B

uh Yeah I'll I'll write it. My last name is too many letters. Okay,.

A

All right, okay, I'll just say I'll, just put that even though yeah you're good having okay, I'm gonna, make this like I, don't know: PTO is kind of a generic middle ground. Also.

F

Raul mentioned earlier in the chat, rendezvous.

C

True I was just asking if we want to put some gold down there, yeah.

C

F

Frozen right now, so I would love to see that getting okay.

D

We'll start work on the spec PR, because there are a couple of things that were discussing in Scotland today when you are there in the comments right now, but they need to be added to the spec.

A

Can I do this and make it key to or something or higher or lower.

D

P two is fine: okay,.

A

Do we want to say something a little more ambitious like an initial version of Randy who is implemented or that there.

D

Is an implementation in progress which corresponds to what is written in the spec right now? I'm gonna have to update this one. So I don't know how, after you write this bunch of work accumulating, let's see how far it goes is.

A

It too bold to say, rendezvous implementation is, is complete, completed and whatever the it's.

D

At least you know it makes progress, I don't know what completed, because that would take into account JavaScript as well. Okay,.

A

Okay, well could be.

D

Implementation and nothing else, I know.

A

It's such a weak kr that doesn't really state. Ok, so you want to just have it make progress but not necessarily yeah.

D

Just for a little bit all.

A

D

I mean what is written in the spec right now. It has already been implemented. That's like pretty much ready. So there are some things: I want to do, improve the client. You provide a seamless interface, so they have a stateful client bunch of other things and also we need to split the service out out of there and the women who addition to have separate repos for the service and during the client itself. But this can happen after them when they should be I registered some it's an incidental, but really ought to do. Okay,.

A

I'm sure I decided whether to put what you just said in the notes or just leave it the way it is now I think I'm just gonna leave it like this I think we should do better Kay, ours and future quarters, but I'm not as worried about it right for q1 I think we should just get something done so by better cars. I mean like really specific about what we're gonna deliver or not, but I think it's ok crap.

A

C

I was just wondering since in 0.9 and Libby to be demon, Bros, 49 and 51. The really are leftovers right. So I think that could be collapsed in the single row because it seems like they're, yeah.

D

C

That's ok if there's ever a PR. Sorry many of these things.

D

C

D

So because this, what I would get as physical bleakness.

B

Is the complete demon thing include some sort of strategy for I have multiple things that use the p2p on my machine.

B

D

Because there's also another issue, there was discussion about having you know the host support, multiple choice and multiple deities. This is something that we really need for the demon, because, right now we don't have an any mechanism for having Fergus accident with many multi-tenancy other than implementing differently p2p cause at which fine. You know there isn't much gain over using no like a single demon per liberty, big house.

A

um Well: okay, I! Guess, since we're coming to a close the end the hour, is there anything that anyone wants to add or change or remove, speak up now.

D

Do you have anywhere from one deaths as well? You know what it's gonna be doing. Yeah.

A

Yeah I talked to him yesterday, yeah, let's come back to that. I don't have an answer yet, but ya know that is a good question.

A

It's just I'm talking to them a little bit about. Does he want to continue with the project and stuff? So let's, let's give him a little time to think about that.

H

Okay, just one finger regarding the project scope, the common tasks for the GS and ago teams, myself and Jacob ever signed, but I don't know. If anyone from there go team will be able to assign themselves to the tasks Oh.

A

Enjoy yes, you mean that.

H

No, no in the projects :, oh.

A

H

Yeah we divided from digestion, oh okay,.

A

H

But there are some tasks that alone they're: all the tasks are like cross language.

A

Okay, I'm sure Stephen I did this one.

A

Well, so visa is saying he doesn't think this will be completed, but I at least put it make him as the owner to go and speak up. If you don't like that, these are right.

D

Now I hit the owned everything yeah.

A

It's this is the same thing as the other Kia I disrespect, 2.0, whoo I, don't know who, okay, that one, it's a harder one I'm, not sure he's getting on.

A

Yes, so this one um well, these are P threes, I. Think then the near form guys might be able to work on the network visualization tool. They would need some input from us if Roley yeah, but I think that I think I was going to reach out to them about that. So let me put this as like and sort of brackets, like at your formal question mark.

A

I, don't know called you on it. You want to own the go side of benchmark Chester I.

F

Yeah I'd be happy to it.

A

Seems aligned with the stuff you already own on the other council, yep I, don't know what to do about DG spec, 2.0, I'll,.

F

Take as many P 3s as you can throw.

G

I think I can take that or yeah sure. Okay.

A

Alright, that's a pretty big one, but but alright well anyway, I'm totally yet for having help. Okay, okay, okay! Well, we can leave like this for now. um Were there any other ones, passcode that I missed it's.

H

Friend, yeah, no, it's okay,.

A

I'll reach out to Dan here for people today and I mean they may not. We never may not have figured everything out by Thursday, but it's on my looks or things to do this week anyway, right core.

B

I have a brief question for uh I. You got four I guess my can enroll, which is that if you go back to the go tab, all overalls tasks are p0 um yeah.

F

B

H

B

If it was me, I'd want some sort of some priorities within those so that on January 2nd it was there. What was.

D

Because we have too many peas, zeros I do think.

C

Yeah yeah that make sense.

C

So, let's uh for me, can you make the piece the ps4 TTL stuff, a p1 that is working and it's not at the moment causing like major problem, so that could be a p1 for me, then the POC for go want refactoring, I, think that that should be a p0. That's super important as we discussed Nick had a droid that was already part of q4 in a way, so I think it should stay on as a PZ over here. I think.

A

It's fine for one person to have two P zeros I, just think maybe more than two. It's like yeah.

C

So I have P zeros I'm, probably gonna I'm, probably gonna, be having pcs on the project tab regarding downstream implementations and native implementations as well.

E

C

That's my only concern that, like whatever we decide to flesh out here, I think should be a p0 at least on some level right. Some toss might not be, but they should that's gonna, be at least one or two pieces there, but.

E

A

C

Looks looks okay to me thanks for bringing that up and even looking for it. You know the well-being.

A

Of the people I.

C

A

Now now I I guess I am a little concerned, though, because if we think these two are both be zeros to me, it feels like more work than you may be able to handle in one quarter, but so.

C

Yeah, but these these really depend on like these are external collaborators, so it really depends on like hot, like you know how much they it's really unpredictable like where they're gonna get to right, because so I would say would be a p0 and Java would make it a p2 for now. Okay,.

D

So, just for the record, NIEM is pretty far advanced as far as I can tell.

C

Sorry come again near me.

D

C

Are they they developing a name bindings right? Oh, are they already enjoy.

D

Playing with maps up here- and you know, like sit, faith- was the one who basically ran into the serious problems with the multiple connection handling bags sweet. Oh, that's the.

C

Others have been helping, like we've, been in constant communication over Gitter as well and yeah I'm, pretty happy that you didn't get slashed by the state as dansgame, because he's doing some pretty good work and stages also like had had a few layoffs and so on that they announced previous week. So it's great that he's still on board.

A

Okay, this is good, I'm, actually I'm glad we we got this one fleshed out, hey, it is past our. So maybe we should call it for this meeting.

F

Should we have a goal of having like final priorities and descriptions by the end of the week or something yeah.

A

Well, so we need to present on Thursday, so it has feels there and then, by the end of the week, I'll say.

F

By the end of Wednesday, yeah.

A

Let's hit by the end of tomorrow, I think we should yeah cool okay, and maybe we can just kick kick take discussion over to IRC or whatever. If there's um you know more questions about what are the parties or something like that? Cool? Okay, all right! Well, thanks everyone, I'm gonna, stop the meeting.