IPFS GUI and Browsers Weekly, 6 Nov 2019

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: Policy for awesome.ipfs.io, Brave v2, cohosting - IPFS GUI and Browsers Weekly, 2019-11-06

Description

About IPFS GUI and Browsers Weekly: https://github.com/ipfs/team-mgmt/issues/790

IPFS Mirror: https://ipfs.io/ipfs/bafybeiggc5ihgevp7huikl33w2rcokun3pmk2fsb43uyt4xj6pn27tsmay/

A

Hello, welcome to GUI and in web browsers for the 6th of November 2019 this week this week on the agenda. We will talk about awesome, ipfs policy, to quick issue, review for ray 51 and then put a pin on co-hosting, we'll see how it goes. Let's start with the first one awesome I PFS policy. So in the past week weeks we been talking about ipfs, slash, awesome rico, which is also a website.

B

So this is this website, and this website is.

A

Backed by a github repo.

C

A

Github repo has always had a lot of pull requests open and it's difficult to tell what awesome and what's not, I without some kind of a policy. So I believe the idea was to.

A

Sketch policy of what things could be listed, Liston listed on the awesome, ipfs website and repo, not sure how we handle this. Should we like just brainstorm, four or five minutes and then write some points or is there a better way? You know.

D

Maybe we should do spend a couple of minutes just walking through these initial comments here, because some thought has been put into it so far: I, like Mac's initial idea or in a Victor's initial idea, in using the curt of conduct as a starting point. But beyond that I think there's like beyond things that are concerning I think there are things that just either we need to come up with a clear definition of what belongs or does not belong.

D

So is something built on ipfs.

B

D

It is ipfs involved at all step, one.

A

Megha, should we like write this on the agenda. That's a notes or yeah.

D

Let's do that so looking at so Matt's comments is basically also copyright violations, that's covered by code of conduct, also the harassment free stuff. We have a harassment, friendly harassment, free space, so that also covered by the code of conduct. So.

D

I feel like the code of conduct is, is for sure, table stakes for being included like any like any part of our anything in any more repos. It's covered by that code. So that's season.

A

D

A

Then it needs to use ipfs like actually use IP, FS or yeah.

A

Yes, I think if a repos name also might be first, then it needs to use ipfs not only likely p2p or IP LD or something right.

D

E

Also, probably sits an awesome thing: it should be used by Morton thanks people.

D

mmm Hard to measure that yeah.

E

D

So maybe it better me to ask that question is: is there a reason why we want to turn any ipfs project away aside from whether it violates the code of conduct? Who would we say no to and why.

A

I'd say something that excludes people due to financial reasons. So if it's a commercial product, it should have like a 3-tier.

A

But not sure if it's a proper way to frame it, but I feel if we publish something that's also my PFS, it should I I, don't think we should reject the projects that require a monthly payment. It's an opt-in.

D

So I guess I'm wondering if, if there's a thing to my PFS, then I think it's awesome and I'm totally excited to pay for yeah.

A

That's a good question: should we, let's say there's a audios and yeah. That's it like at this point. It's like free service, but would it always be like this? Would there be like a freak year such as like on Spotify? It has a free tier when you got like advertisements between songs.

A

That's, why I sort of sort of come up with this idea that there should be. It should be possible to use a project or a service without paying. Why I.

E

Think, like talking about that, we also have to take in account that it's not services and that you're talking about if.

D

Yeah ABB's.

E

Articles data set services tools and videos.

A

Also like there could be a data sets which are behind payroll. Are those awesome, I.

E

Don't think so so.

A

That's why I sort of like suggested they have, even if it's like a five day, trial, yeah yeah yeah. So if you.

D

A

You click on something you should be able to evaluate this and evaluate it without paying I like.

D

Maybe we described that as immediately usable and oh by everyone, so something like you can you can you can you don't have to? It is not because I agree, I, don't think we want anything. That's entirely! Paywall yeah I, like the idea. I, don't want to ban somebody who has a paid service because we want people to actually make a living, pay their rent and buy clothes and food by working with that PFS.

D

That's the end goal, so we can all do those things right, but I'm me I, like the idea that you we only want to publish things here that are immediately usable by everyone.

E

Just a suggestion: why did you go over each category in the reasons why we would something for each category, for example the videos? Why wouldn't we accept all right.

A

E

It's simpler, I, don't know.

A

Yeah, let's say some streaming service is using IP FS.

A

Would it be okay, to put it here.

E

Well, maybe new section all.

A

Right then, that would be like in services. Videos, yeah videos should be oh.

A

Man we quickly go into the ribs of like having a license for redistribution with within the awesome ipfs. No, we just link right. There's no player yeah.

E

A

There's no producing.

E

But right now, even with five videos but they're not saying against telling, like all the ipfs cam videos, for example, which.

D

Yeah, but why wouldn't it be great, though, and I think that's the question like? Is there a volume amount? Are we trying to restrict certain sort of things like limit 20 things in each category or like if we're not limiting, and why is feed busy from the.

E

D

It makes sense as a category so I will find new videos. Yeah video is the videos the number one of all media distribution. It's the number one engagement channel for developers, I like kicking. You can get your message out in your tutorial scene and your workshop scene and your materials learned about with video. So much more than any amount of articles or any amount of other type of community engagement. Then the numbers kind of speak for themselves.

D

People love watching video, I I, don't really do like reading code, but a lot of people really like videos. So I really like that. We have video, then I.

A

Think, like this point about being immediately useful to people, applies to videos. So you click on the link. You can watch the video. The other thing is that about the content itself, we don't want. Probably we don't want to have dozens of introduction to ipfs videos which are sort of like all the same. It probably should be something that adds value on top of other things that are already on the website, but it's highly subjective. How do you evaluate that? It's like a dead point on the policy.

D

Again, I want to flip the flip the question a bit because I feel like it'll, be too hard to write a policy. If we, if we don't, have an understanding of the broader constraint, so there seems to be a concern about having too much of something too many of something that too many of the same thing.

D

So then we have to because we can't really just we can't make those other decisions about what meets like how, if you have 20 introduction to ipfs videos which well what if the one we have up, there is the worst one, and then one of the ones we didn't accept is way better yep, so I get I guess from where I said.

D

I would rather make sure that we just get everything in here if it meets that basic criteria, it's about IQ best and then later we can add things that, like either you know reddit style loading that flips stuff up to the top or provides a sorting order. Based on how often it's viewed things like that, like some or how many people start it or something like that. Like later, we can add ways to maybe highlight the best unless um the most awesome awesome ideas, but I I feel like it.

D

Unless we come up with something really clear like we only have 20 of each in each category or something like that, then it's gonna be really hard to out we're. Gonna have to always be reevaluating. It always kind of redesigning these things.

A

Yeah I'm not sure about like limiting, because then it's like every new thing you want to add will require you to drop something. That's like.

D

Right exactly so, that's what I, like you can only you can go one of two ways there right like you're, just like okay, let's put lettuce much stuff, be found here as possible, get indexed by search engines all this stuff, which will be a some also, some sort of preservation of sorting function on its own.

D

It's just exponential on the website right or we do something like that, where we have a system for how we limit, what's there or evaluate what is the best introduction to IP this or we rotate really like forever for every new one that comes in in the data sets category we pop one off the bottom. Take it out I.

A

Think that was a proposal to randomize. It I think yeah.

E

I think they are already randomly yeah.

D

But it seems like the questions that we have around. Why would we say notice something aside from code of conduct? Well, we already have a bunch, it's it's, not as good as the other ones. We already have yeah.

A

So it needs to be like not noticeably unique.

A

D

ah See that's why I'm like it see it's much easier. If we just say just add everything make it be like the canonical source of all that I've give us things, and then we can layer on ways to narrow down access to the best bits. Yeah.

A

I think like right now: it's if you look outside maybe apps every time you load it. You get a different board. Basically, the more you have the different different. It is yeah.

D

Almost makes it so you, so you can't find something that you saw before ever again. Yeah.

A

But it's also like a this river like website when you just go and get something new, every time, yeah sure like Easter, because we right now we got like respect code of conduct needs to use ipfs and needs to be no.

D

A

Immediately usable I will missing something crucial here. Should it not it cause like things like a respecting copyright, all the attribution it's all under like code of conduct, so that's like huge surface covered there, yeah.

E

I'm Anna this privacy covered by the gonna, stop code of conduct. Okay,.

D

Good question like.

E

Apps privacy and data yeah.

D

How is the data used if I use this IP FS be a staff.

A

Yeah, it's I think that that's a reasonable so having this. Apart from the code of conduct like needs to come with clear privacy policy, would that be useful cuz that if it's a video, then the video is on a some service and that service comes with privacy policy and I mean.

D

At that point, like a on YouTube.

D

That's a trade-off that that we need to let the I feel like that's a trade-off. We need to the aid user, make it and also means that it's risk on its on us, as the curators embossed might be fast, then have look at the privacy policies, evaluate the surveillance and data collection practices of every single link that comes in, and that's just not not reasonable, like would that benefit that we're gonna get from that is gonna, be really low and then it's all very much to be auditing other people's Cabezas. So.

A

I propose, instead of like auditing privacy policies, just require to.

A

For a privacy policy to exist, so the people can always evaluate on their own if it's an app or a service. That service should provide a link to a privacy policy and.

D

You can put that in the issue template too, so that people have to.

D

But not all categories of things might have privacy policy like downloaded and data set won't have a price privacy policy use this library. If you want, we have the full complete code.

A

D

Like that to your license, yeah, because.

A

It's like sort of like if it's a data set or some article, then it's like a license under something could be Creative, Commons, zero or like requiring comment. Creative Commons zero. Holding to that. Unless it's something else and if it's service than there's our up, then there is a privacy policy right.

A

C

More or less should cover sounds good. Another.

E

Thing which might be more difficult so be careful about, is the if it's upstate, for example, we have a tool called IP, offense GOI, which is supposed to be a simple of Windows. You are providing a defense integration which.

D

E

Not been updated since thousand seventeen.

A

Ok, so I think we could tweak this.

D

Well, here's the thing now: if somebody wants to make a new windows-based, ipfs QE, they would find that and maybe for kids make changes to it. They might start participating. They might revive the project, so I don't want to say, remove stuff, just cuz, it's old like the best ideas explainer, it might be old, um so I I, don't know. If that's the right trade-off. There I think the sorting, sorting and ordering might help they're a little bit like if we change from random to sort by newest.

D

Always so you always everybody who comes sees fresh content. Yeah might be a way to then that that old one will get pushed way down below the fold. You know 8:32 the content Thanks.

A

Or we could like tweak this randomization sort of be a tiered like randomized, newest stuff or the past like six months or a year and then randomize everything colder than that.

A

But it's sort of like a separate, separate discussion. We got four points and I I'm, not sure I.

A

Think there should be a point about if additional information was required, so you submit a peer to the repo right and you're and you forgot to link to license or a privacy policy or something and someone making peer review asked you for that.

A

We should be able to simply close the PR without merging if we don't get response within angry.

E

A

Now how to frame that in one short sentence.

D

Well, in that case, it never gets added right, so I feel like that's a little bit different yeah.

A

I think it's a basically if, if their license is missing, it's breaking the last point. So yeah yeah.

D

A

I think that's a pretty solid list.

D

Yeah- and we still haven't solved- that we kind of do have an open concern about too many of X too many of Y, but I feel like that's a low priority concern compared to the rest of this stuff, yeah I think I. We can fix that in a whole bunch of different other ways. One yeah.

A

I think we have a bot which checks if links are still valid and we like remove that the clicks that no longer resolve. So we already have like a cleanup process in place. Wings.

D

A

Could always remove like extend it to periodically checking projects because the domain may still work, but a different thing might be behind it domain. That's that's a separate, separate problem space. Are we happy with this list, or do we want to spend more time on it?.

D

Looks good to me I think.

A

D

Think it's the right balance of how much time we want to put into running awesome, ipfs and having some clear guidance on what's allowed and what's not all.

A

Right, I think the action item here, because.

D

Later we can then say say we do end up with 258 introduction to I confess videos and we're like okay, that's a good problem, let's figure out, then, let's figure out how to you know sort them by what people like or what we're effective or something like that, but that we can solve that later down the road all right.

A

Go so we will be our driver, Polly Olsen, to the repo with those points and we'll likely find them later. Should we move to the next one brightly one issue again: yeah.

D

One quick thing: can you request review on that policy from what's-his-name there's a nice fellow from your continent that also reviews some of the PRS? No, then he pops in okay.

D

Please ask all of you to review I can't remember if you is French living in Germany or Austria and living in Belgium, or something like that. Okay,.

A

D

A

A nice yeah nice person in the 3po from Europe.

D

He was reviewing a bunch of these PRS. Oh okay,.

E

I think I know who all right.

A

Sounds like a plan yeah all right, brave, v1 issue, review I should look at the projects yeah.

D

You can walk through the set of issues that we say: okay, this is the diffuse koala Lorenzo. If we can stay like these are the definitive set of issues that we absolutely must fix and then strong line, I'm saying and call it v1.

D

Because we have a lot of options here,.

A

I was super scared that I forgot to press record button, so I can check what I'm recording sorry Avella all right, let's go with like in order in progress. This is metal shoe which I sort of like started splitting stuff out of it, because it's grew too big and you can scroll and scroll and see all the related issues in other repos. Sorry to use like a jacket like a junction point to see, relate work related to brave, but that it's mostly like a placeholder, because I'm not able to like see, see this this board.

A

So the first one on the to-do list is a local discovery story behind local discovery is that right now we are able to discover, go ideas, but in the local network using mdns. However, if you go to two brave browsers, those browsers are not able to announce themselves using mdns and instead they discover each other using centralized, WebSocket star signaling server, so they discover each other using that signaling server and they exchange local IPs and then they connect directly. So the central, the centralized signaling server is only used for this initial discovery.

A

That then, there's a peer-to-peer connection between them, so a fix here would be to see if we are able to reuse this UDP port 53 53, which is part of the spec, if not write a custom discovery or using different port. That's specific to brave, that's more or less background on this issue, so.

D

This seems like it. This is functionality to kind of be aligned with beyond regular gateway usage. So I feel like this is something that we could push to. V2 yep.

A

Should be like Magnus, somehow like a priority or yeah.

D

Or a label I think a label, for you is to something like that or maybe.

A

Maybe I should like should I, like name this brave v1, and make a separate project for brain Vito. Would that be easier now.

D

That sounds that sounds good. It makes it harder to be able to to view just for all the two issues, though so I'm kind of in favor of doing a label, no matter what and then you don't want to use a project board later. That's fine I'll! Soon, let's.

A

Do a brave Vito! For now we can always I.

E

Was going to suggest, instead of creating a new project grade a new column.

B

B

B

B

A

Let's make this yeah.

D

That's really nice, because then we can see that clear, yeah between yeah.

A

And then, if I like added to this project, if you'll automatically land in d2, because it's a board with some sort of like automation, alright, so this is a V V, then welcome screen should suggest embedded node as an option.

A

So that's definitely tier.

D

A

Yeah, that's an open question, because if we ship right now, we ship when someone installed sacrifice companion, they are asked to instil a profess desktop or going offense on this welcome screen. The question is that we remains true for v1 and we could change that in v2. However, we could also in v1 add this less prominent option to try embedded ASAP Affairs should could that be v1, I think.

D

This is a could be, v1 doesn't have to must be v1 right.

B

D

Like a nice to have I think we do want to encourage people to try the embedded node. We have a lot of opportunities to do that. Azzam, yeah.

A

Let's make it very, very low, and for now yeah sounds great, all right brave and what you are missing: values in latest column that sort of a pro open problem still I hope to. Let me find it. Yeah I was hoping that maybe I'll show it to maybe.

B

Break nightly.

B

All right and I switch to edit notes.

A

No way, if I stop sharing, it will start working I.

B

Can't believe it I.

A

Have no idea, why that's a suppose, that's a separate okay, because that's nightly it may not be a problem. Let me check.

B

That's what you get for using nightly for demo.

A

Okay, so my local node is offline. There's no ID first desktop I switched to yeah. That was Nike. I switched em, but note take some time.

B

To start, eventually to start yep and I got some peers.

B

Okay, so if I open well, you I.

D

There's some fascinating black blocks of and rendered content. That's pretty dope really.

A

Speed things up.

A

That's because I, just like immunity.

B

A

So this value for anyone.

D

A

This it's fast on the regular.

D

Laptops, it's just my luck, so the this value is is the latency of a Givens here, because.

A

I'm both recording and.

B

A

On the pure screen you can see, the latency is empty.

B

Another thing: a no idea what this is happening.

A

On bandwidth graph and what's interesting, this bug with canvas happens also, if you are using regular web UI in brave. For some reason, the empty latency column is only on the embedded node. Let me move that to a separate screen, but basically it's a a cosmetic issue like this and and the other one are cosmetic issues.

A

Would probably it does not like break? Where do I screen it's just like a missing like a sitcom.

D

This also seems to mean.

A

Are you mute it or it's just me not not hearing you at all, cuz I just realize you liked your movie, but I cannot hear you anyway. Can you say something hello.

E

A

It's me sorry guys, one. Second, a can.

D

You hear me: I can okay.

A

Can you hear me now yeah.

D

We can hear you the whole time but I think maybe you couldn't hear us.

D

Are you using the Huawei laptop? Maybe that is the problem. This.

A

Video stream is not sponsored by highway.

A

So sorry about that, for some reason my audio just got muted can.

D

A

Yeah I can hear now I understand, you're, commenting behind the scenes for a long time and I was not able to hear you.

D

This seems like a nice to have not a blocker for people all.

B

Right, let's mark it like now,.

D

The other issue where the bandwidth page also shows nothing on the canvas that seems like it might actually be a blocker for anyone. If, if with non-embedded node, we usually showed that data, then it seems like we should figure out. What's causing that to me. Oh.

B

A

I believe I created an issue for for that graph in web UI repointed.

A

A

But I should be able to add a card with this, so yeah.

D

I also have a recommendation. I say we give ourselves ten minutes to speed through the rest of these issues, because it's like taking should have turbo triage.

D

Oh I think you are familiar enough with most of these issues to that you could probably make a determination fairly quickly: okay, like switching webui to Indira I profess, probably not a blocker, not.

A

A broker at all request, throttling block action. Button, update, that's v1 for sure cuz. It's like the UI does not work. It's probably optimization. In the way we analyze things evaluate alternative data store, that's v2, the current one works also. We are waiting for performance fixes like you go preload of the high cached web UI triggers trotting Osprey old nodes that we want- and that's mostly like me, either throttling it on the extension end or talking with Olli and the Gateway be first team to make those limits throttling limits higher.

A

If it comes from companion or of something, then, where do I show simple it NFS.

D

Wait wait wait hold on back up so that last one that happens no matter what right like that has nothing to do necessarily with brave.

A

It happens if you are using and by the JCB of us, so.

D

Then it definitely sp2 I, think anything with him with embedded node and any optimization for embedded notice, v2 and the whole point of calling this v1 is now between the release of b1, which is embedded notice. Experimental it's not expected to work MV to is. We can figure out what doesn't work, don't fix it. Ok.

A

So if we are moving things that happen only if you are using committed, node, yeah yeah then also like all this stuff, who.

D

Is that also I felt? So that's why I was asking about the the bandwidth screen. Yeah.

A

So, actually only bandwidth, that's a problem in both okay.

D

Then yeah like I, feel like anything that is directly a side-effect of running embedded. Node is not a have to fix for everyone at all. It should be in v2, okay, like let's, let's identify the minimum set of issues that we can, that ones that we absolutely have to fixed you'll, say the regular I'd give us note not invented running engrave. Oh yeah,.

A

Oh, that's a super super easy lens to pass through. That's because the things in v1 are things that happen or needs to be answered, no matter if you are using external or embedded node right. So mostly we care about this initial experience. Someone is going to settings. Yeah I did twice today companion and you get this in default experience when we ask you to install a P first desktop.

A

So in that case we need to have the status page working fine with a remember that it with external go up if, as we could, suggest this option- and we also need to fix NFS size. However, this is not like specific to brave at all yeah generic bug in web UI. So that's probably I.

D

Feel like that, that's not a black or either we.

A

Probably I'm not even sure if it should be under brave, because I found it when I was working on brave.

A

That's mostly like related to co-hosting yeah.

D

They say, and that's one of a number of reasons that can cause those things to be out of sync right, like across the topology of running, desktop and running in a bedded node and running web UI. There's multiple reasons why the numbers there might not reflect what you've actually added to repo what you haven't so.

A

That's the problem. The problem with this lens is that actually all the problems happen with embedded node, that's great, so the only the only question is: should we suggest them, but note I think we could.

D

Yeah I really like that I, don't think it's a blocker for it, but I think it would be a really nice in to introduce like we've gotten this far. The extension is semi pre-installed.

D

We have some nice feet and then we have an additional set of superpowers that we have in brain that are experimental for training yep, better, not production, ready, yeah, yeah and.

A

Also, that's a part of we like improving this welcome screen right now. We got at recently added button to just open web UI.

A

If you got everything installed, the idea was to add like separate buttons for like files and other things, to make this initial experience more useful for people too, because people insta like PFS companion, that's a neat going on a tangent a little bit annexed started looking at those like on an InStyle feedback, and often people were not sure what to do or like it was not what people expected and if, on the welcome screen, we provide some options that make make it easier for people to discover. What's what I think?

A

That's it because like v1 is basically the day works, polish and and the rest is iterating on the method not supposed to be with v2.

D

Excellent, that sounds great. That means maybe we can put these posts together by the end of lab week and maybe publish early December, because otherwise, due to other releases like go, ipfs and file, current related stuff and then holidays, we would want to wait and do a blog post in January.

D

That is a requirement so that I think this. This makes our our our pathway and uncluttered very clear to be able to do that and doing that in just a if we can do it in early December I think we've talked about.

C

B

A

Sounds pretty good and DeVito's stuff, it will happen like in the background anyway. Those are things some of those things are not like very specific, but it's like breakfast first function to address them sooner than that yeah.

D

And what we should do, as well as maybe at a future meeting road map out where we, when we want to do V to edit noted goal, is for 2020. Oh yeah,.

A

Because those are sort of like a leftovers from v1 v2 will have much more yeah, like native native stuff hints on the browser level things that we've discussed in general right all right, I guess: that's done! Cool.

B

The remaining topic is co-hosting co-hosting.

D

So kind of thing super excited so much stuff to do, but but the proposal we're thinking about one of the themes for 2020 being Wikipedia and this type of flows and what I wanted to suggest was to actually stop work on co-hosting for now and tell the end of the year, so that we can focus on getting the test stuff up and running before the end of the year, which is a the priority for the quarter.

D

That now is what I wanted to propose. I.

A

Kinda agree: I mean we should finalize those open, PRS, I. Think honestly, we the only like remaining thing. We have this common line tool installable from NPM, which is acting as co-host, and there are slacker I believe. How can you open the pr2 to.

A

E

Haven't but there's the issue I'm just waiting to know. What's the stressful tres tres move something.

B

Like that, oh yeah.

E

Yeah for the 4s or twister confirmation, because I suggested wasn't 100 megabytes, which is huge and then I know that that I, if s is tight, is 10 megabytes and the blog is for 40 megabytes and Ellen wasn't expecting yeah.

A

That was interesting. Conversation cuz, yeah, turns out. The repo with the website has a lot of like super old examples which I don't think we no longer linked to, but we still publish them with education yeah, it's a professor, it's the duplicated across all versions, but still yeah, so long story, short I think we, we probably could could should add this threshold. So we warned user before they've been super big thing, because people will first think people will try to like co-host Wikipedia and we should at least warn them. Hey it's 650 gigs.

A

So we would, like you, know, check the size of entire tree and display nice warning, but I feel when it comes to time. We invest in co-hosting this quarter. That would be like the last task, the integration of like ipfs co-host as a library like both making sure it's a small bundle library and also then integrating that bundle with ipfs companion you I and type II first desktop for, like updating in background I, feel that's something that needs to happen, that in q1 or later yeah.

D

It was actually the threshold conversation that kind of made me start thinking about this, because I was like okay. There. There are I feel like we're gonna start making decisions about what the users who might be doing co-hosting want without their input and without, like I feel like that was a whole discussion.

D

The default, whether the default should be lazy or not, was a discussion about the decisions that actually the users of co-hosting need to make about how they want it to behave and where it's integrated in their system and I want to be able to have that conversation in a broader way, with the input of the people that actually want to do that archiving, like a little bit of user research and a few interviews of people who were like if I was gonna, have my browser open and I happen to have two I profess, desktop running and being installed and I'm at a website and I'm like I, want to click a button and say I want this website to be available for everyone.

D

At that point, they need to be able to make those types of decisions around what hardware capacity they have available, how much bandwidth they want to be able to share and use, and we can hang out and talk in circles to make those decisions for them without their input and without doing that.

D

Research and there's no guarantee we're gonna be right, and even if we do the research there's no guarantee right, but I want to be able to have a more informed decision making process around how we actually rolled this out in a way that makes it available like we have it, though I think the potential here is for some of the high level dreams of ipfs to really be true for end-users in an easily accessible way and I want as many people as possible to be able to take advantage of that and I.

D

Think if we're just doing that in in repose, where it's. You know the three of us and maybe Alan, or maybe one other person talking about in the context of a CLI. It's not the right, we're not going to have the best ability or visibility when we make those decisions. So that's why I was like I really want us to be able to think about this in a more product oriented way so that we can get it into the hands of millions of people and I.

D

Think it's gonna be hard to do that with lab week or testing stuff that we need to do and then the holidays, but it does fit in really well into a 2020 where we're focused for the whole year and a theme that might be orient around exactly this type of exposure. It might be fast.

A

Totally and just to clarify this discussion about threshold was specific to this common line tool which the the I believe like the key purpose of this common line tool, is to ship basic implementation of co-hosting spec. Just so, people can play with it and it works everywhere, unlike like the bar script. So that's that's the only yeah.

D

At the same time, like even the people that running and see a live, you know we're kind of making that decision. We can't just say, there's no default right, that's another one. There is no default, just like the person has to specify which the way they actually what another good there's time put into, something where we don't actually need a decision. Yet anyway,.

A

Yeah so like I I agree, we probably should that boom put a pin on any new stuff related to co-hosting, especially like this library integration to combine an under stop I'd still like to ship safer version of this command-line tool. I like I, fully understand. There is the concern of like making arbitrary decisions like we can. We can always change it. The problem is like if someone likes like enters ipfs, co-host, add Wikipedia, that's really problematic.

A

If, because we the problem is we don't have api's for tracking progress of fetching such huge data set it's 650 gigabytes and basically the command Hanks. Until you got everything on your disk, that's why I want some helpful hands to user? Hey it's 650 gigs, but.

D

Wasn't the default lazy before and then that wouldn't happen it is still defaults. It's.

A

D

A

We can just like put a pin on it. That's the part on the pause button. Yes,.

D

Is there any issues, there's probably there's always issues. There are already issues tracking things like uploaded download status, the way that you're describing file transfer status, but are those anywhere near a reality having some ability to see what happened like ipfs out a 650 game directory or file. So.

A

Like if you add data to ipfs, there's the tracking progress we got API for that I, don't think we have API for tracking fetch from ipfs itself. So if I want to fetch that CID, which is like 600 gigabytes I, have no good way to tell at what's the percentage of that on my local note, I'm, not able to tell how much it will take given the current speed and that's why we sort of like created co-hosting to surface those problems right.

A

We, those are very good problems to have when you start playing with with goo, you see. Oh, we actually don't have api for tracking dot and that's something we will need to tackle before we integrate it to companion or aquafers desktop in some way because like if we do the lazy thing, then we know we don't need to track progress, but the moment we give people option to pin entire thing. Then we need those API X, so that will probably happen in stages even in like 2020.

A

It will still happen in stages because we need those API, still an team that go and other places all.

B

D

Is it is amazing how how powerful it is, building actual gooeys and surfacing how engineering oriented their project is so far, I mean it is a protocol, so that is not really surprising, but in terms of understanding what the needs of those protocol that protocol is in order to with it. What we need from that protocol in order to do basic things like user communication- that's that's without a doubt.

D

One of those powerful parts of these types of projects spending a year on developing co-hosting into a very mature product flow I think would would be a result in likely nothing but very good things for that core protocol.

A

Yeah and then like I'm, not smart, like we are not married to this specific version of co-hosting. A messy.

D

Convention yeah.

A

It's a convention to start conversation going I believe like like the key thing that we should like focus on. Is this interaction when you are on a website like Wikipedia, and you make that decision all like the technical, the decisions and a which api's are used or added should come from that moment when user wants to help Wikipedia, but they may have limited space or they may not fully understand how ipfs, whoo and distributed networks work, and they just want I, don't know I they should be able to like.

A

Oh I should be I want to help to the degree. My local machine lets me to like write and those are very interesting questions from from, like the UX experience, all the technical aspects and problems we have for our service, which I'm.

D

A

Happy with how come.

D

A

So far because yeah.

D

A

D

That's part of the challenge is that it is exciting we're like ah there's a lot of potential here, and it would be super fun to be able to build this and we'd be able to see kind of like how what the adoption but would look like pretty early and easily, and that's a that's, a pretty powerful, powerful story, but I want us to be able to do it. We need I feel like we need to get.

D

We need to get the testing stuff out of the way first, because that's the priority for for this quarter and make sure that we at least have some like basic end-to-end tests and asier before before going on in tackling these are the bigger projects.

D

Alright, who is going to submit the PR for them or the issue of the PR for the policy? To start that conversation, oh wow, that's super easy.

D

We we did the hard part we had the conversation. um What I was thinking is maybe I just post I'm gonna paste this and that issue, because there's already an issue there for a discussion and the people that are there comment on this proposal for an initial criteria, and that would be a good start and then once we get some agreement, there's some in a PR, so I kind of go back and forth on like submitting a PR PRS have momentum.

D

So then, once there's that PR there, it seems like a little bit less of a discussion. It's like a way to totally own the conversation. Without it's that's a little bit. It's like an aggro conversation, starter, there's already an issue in a conversation, so I'll just put paces in there. Nobody says any big great, very it's very.

A

Permissive policy- oh yeah, I'm, like I, had the brain card. I meant nice for this meeting.

D

Yes, click nice I need to convert all of our there's like a bunch. Other means I need to convert to use that now.

A

I think it should be easier, because now you will like remove puffs I had to like add quarter calculation to Volkers altar notes.

C

A

There was a bug, I fix it. No quarter. Sorry correctly calculate now.

D

I just want to I was gonna, add a PR that allowed you to to have that tool, read inputs from the URL bar, so that then you can actually use it like an API and make bookmarklets and add-ons just add those parameters to the URL. And then the form will autofill.

B

Yep pretty useful I might.

A

Say all right guys we'll be stopping recording soon. So, let's wave and see, let's see everyone next week.