IPFS GUI and Browsers Weekly, 26 Mar 2018

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: Web Browsers and GUI WG Weekly Sync - 2018-03-26

Description

Tracking the Path to getting IPFS and other Decentralized Protocols Natively Supported in Web Browsers:
https://github.com/ipfs/in-web-browsers

IPFS Mirror: /ipfs/QmPTedkMw2T6sFKcR927WZX9tv5zUacWMnowoSia4b1vWL

A

A

Alright, just for the authority, this is the in web browsers and ipfs GUI working group, weekly team ketchup and the we're gonna start with reviewing the okay ours for the first quarter, checking in seeing how we're doing and then we're gonna just come up with high-level first draft of okay us for the next quarter.

B

I, just paste, sorry, sorry for interrupting, but I just paste it link to the hockey arse so that we can open I'd rather not share anything from my my laptop because I'm recording and too much for it. So if anyone wants to share that's perfectly fine, but I just won't do that. For my machine.

A

I think, as long as either people can remember them, or they have them up one up in front of them- I'm gonna just quickly whizz over the headlines. The objectives for the first quarter were the the in web. Browsers working group has a shared understanding of how ipfs should work in web browsers.

A

I still think that is a useful goal and the one that we haven't finished defining.

A

With so yeah the big, the the one that we're missing is document the user journeys that capture and ideally user experience using ipfs and daily web browsing I feel like I personally, have a much better understanding of what's needed through researching what the existing user interfaces do. But I don't know how everyone else feels about it. I, don't think we as a working group yet own the story on how idea should work in web browsers, and that was the that was the ideal outcome of this yeah.

B

Like my intuition is same, we have a lot of.

B

We did a lot of soul-searching and I feel that we we can just copy dis objective to next quarter yep and create appropriate, like cue results that are continuation of this effort.

C

B

It's like an ongoing effort.

D

A

Justin I'll see this morning: I was kind of suggesting so I think the creating of brand. The blend book was a kind of more general effort that I think a guitar is pretty happy with she's done the first implementation of that, and it seems good, but the style guide and URI kit there's this something that exists.

A

But my concern from David's feedback on desktop designs was that if we use the the UI kit as presented, we end up with a design that we're not yet happy with so I'm keen to pause coming up with like how should an input box look until we have decided what it is web UI is going to be so I do see that as an important thing, but we'll bond it at the end of next quarter. I think it's.

B

So like when it comes to key results that we would aside, let's say, I assigned to a guitar more like putting a pass on this or I, will just Renate like reframing it to let's say wire, framing web UI or something.

A

Like that yeah I think at this level, so they're the kind of objectives and key results level. The key result is going to be something around having an incredible like coming up with a metric for this. For deciding how good design is is going to be difficult. If you know what I mean I, don't think that that specific key result is having a UI kit per se, it's having a web be wise. That we think is really useable. Really simple looks amazing. People want to use it or.

B

Maybe we could even do not put any like subjective, yeah protection on that and just say that the cue result is to have at least one or two iterations of web UI wireframe, and maybe that's enough for a cure organization. That kind of thing yeah.

A

I guess I probably shouldn't take too deep in all these individual key results, because we'll be here all day.

B

Yeah, like generally speaking, we had this feedback from Mars that we should not have more than like three or five objective, and no more than like three or five objectives, so yeah yeah, yeah agreed.

A

So I think um yeah. We were kind of used to writing github issues when we first started this process. So if we were trying to write, one line for each concept, yeah I was kind of saying was more generally like I'm paused on any style guide stuff. Until we've figured out which widgets we need.

A

Yeah, okay, shared understanding. We need copy across objective to browser developers, understand the requirements of the decentralized web. This also still needs to happen. I am I'm working on. Are we distributed yet calm right now, I've just pushed conveying it to a hugo site, so that start just editing data to add more topics to it. So it's now ready for other people to start filling in more nuanced information about what what we want, what we need from browser vendors to make the distributed web a thing I think Kyle has been pulling together.

A

People who should be on a mailing list, yeah, so I think this is all in progress but I think again. This one needs to be copied over essentially where we can come up with yeah.

B

It's like especially the networking related stuff that takes time and there's this addressing content decentralized. Web I had like synchronization and catch-up with Lars last week, I think regarding Dutch. We exchanged notes and have some general discussion in the direction where this document should go, but it it won't.

B

Candice quarter like I, was to like way too optimistic when it comes to like challenges when what what we want from this document and after discussion with ours, it's like broad and scope in my hand a little bit, especially that before we provide technical details, we should provide a very good high-level case and there is a ton of notes and there needs to be done. A lot of editing related stuff like worked before. We can call it a draft so that I just move it to for next quarter.

A

Number three move towards ipfs working seamlessly in browsers, I think this one we've got the most progress on yeah. This one is fantastic.

B

We like window ipfs, was like nice to have it right now. I'd say it's like the main feature we have and it made embedded J's ipfs note much more relevant, obscure.

A

B

Without without this there was like ah it's nice we can upload but yeah. When we have window ipfs object, you can actually run distributed applications as long as you take advantage of it. So that's it's a fantastic progress. My personal idea when it comes to this is to iron all the performance-related kinks when it comes to sharing and pinning maybe not only performance but UX stuff, like user just drags and drops huge directory, it should like work, be slow, but work, and another idea is to.

B

We probably should have some kind of key results related to this sharing revamp, or maybe that's for the next one I'm, not sure. Maybe we should merge those two now that we have a low-level window ipfs.

B

A

B

A

On that, come when I say merge, you mean merge the like.

B

Sharing like we have this move toward ipfs working seamlessly in browsers. This is like more technical. One I like I feel something related to break integration. Maybe we should have key results related to to like brave and work justice, ignites and like ongoing effort, and maybe this sharing and uox two tweaks should go to the next one. Let the fourth one write like a full subjective.

B

Yeah like this, provide a nice desktop experience, cause to collect more more, like anything here, related to X, I I'd, say, should go to the last one and the third one should be like. Maybe we should provide more I, don't know more sample, apps I'm, not sure, but mostly like I'd, say performance fixes right, I,.

A

Think yeah we don't have to. We need to aim for three or four objectives, and it may be that we're just copying them over, but it may be that we need to reframe them slightly more to better capture. What we're talking about yeah point taken more sample apps would definitely be a good. How do we educate people about windowed ipfs is existence and how they can build, disappeared, apps, I think I think we need to be trying to build some distributed apps very soon, so that we dog food this new.

A

Now that when new idea, this exists like we need to try and build some things on it, so that we can write, write the codes alright.

B

Like we had this at some point, we had well at the moment when we were just writing. Github issues secure easels. We had huge odds. That said, we should implement by PFS object support within pew button. Maybe something like that: I'll just go over location.

C

We should I think we can now actually send a pull request. Appear pets like if Joe's just waiting for all of the pieces to be in place before before I could send runs their pad, but yeah I can also just back on the lots of sample apps I can totally get behind that.

C

There's two that I did already one with the pub/sub chatting and one with the peer peer map thing, but yes or more, some more would could always be good. Yeah.

B

And like where do I will be one of them, yeah.

C

That's really absolutely the.

A

Pick one that's.

C

B

Very important that it will be a core feature that we use on daily basis, so we will like catch all the problems before the hill nope.

A

There'll be no problem special.

A

Brand just well, it's highlighted the brave stuff is pools to pull. Request is open in as much as the IPS companion works in brave, but without register stream protocol. We can't get the nice URL bar handling that we want where we can literally return files directly from the embedded ipfs mode, which is where we want to get to the pool of quests that can make that happen is a sat with them.

A

They've put some comments on it, so I need to get back to them about that, but again, it's still not gonna get in for q1, but we can reframe the key result for q2 to have it to get it released like we.

C

Know the with what we've copying stuff over and reframing them I think we should definitely reframe him. I, don't I I'm curious to know how we give give off the impression that we have made process progress without like without just copying them over, because it we need to now say well, we have actually made some progress, and now now we know that these are the rather than just copying over the same thing, so yeah I strongly agree.

C

We should refrain that what we have there, even if we are sort of saying we are copying them over, but we've made progress on almost everything we.

A

Agree on the flip side, I think it's totally reasonable for a specific topic working group. If, if we came up with good objectives in the first quarter, it's probably okay, that our objectives haven't changed that much in the second quarter, but yeah. The key results may may need a little bit of evidencing the things.

B

I moved on yeah like that, this another step in the right direction.

A

But if our documentation is improving all round, so that's good I guess of note there we probably perhaps didn't do necessarily a great job of creating a key result that could be measured but I. Suppose the documentation is up to date. So I think that one's a success.

B

Yeah, like we have in companion, we have dogs directory that we started putting the dog star and there is no longer one huge readme and one huge like developer notes on the contribute page, so I think that's quite useful and with like yeah that that's linkable documentation is slowly growing. So I'd say we make the good progress. I. Think.

C

If we release companion and people will come to the site and start opening issues saying how do I do this, or what do I do that? We know we haven't, got done a good job on that yeah yeah. But if we can then just point them at the dog, then that's good yeah.

B

Like we should surface those documentation like those dogs in rails, notes and the the blog post, that will happen after so that people won't ask those questions. There was just a link there.

A

So the next day you work, em cookin me yeah companion follows draft. If I get a style guide. Well, that's not happen because I haven't committed to the style guide yet switched to Jenkins based COI process for companion. That has happened. It's done and reek desktop the ipfs Joey's. Even you have been fighting that upload. If you've made it work by force of will.

E

Yeah about this he arms the first one wasn't sent to you. The published I'm RELIGIOUS of desktop I. Don't know if the secrets, management and Jenkins is already working, so I, don't know. If you can't do that or not Victor probably knows that better than me.

E

We address correct Reporting, Initiative percent coverage. I, don't know if we should do that, because we are going to change desktop to use web UI behind. So what do you think should be well? Should I? What should I work towards 80% courage, knowing that in the future that the code will be completely different?

E

It's a good point, probably no and is the same for the documentation for desktop, because if we will change in the near future, it will be completely different. Yeah.

A

I would imagine that this this entire key result will definitely change focus. This one won't be copied across without significant changes, yeah like what we need.

B

This is called luck coming from it's missing on the ipfs companion site. That's we for the desktop app. We should have an integration tests which are like running actual application and for its electron.

B

So there are probably some tools and we could reframe this from like we want 88% of for this coverage to like create a test framework that runs integration, tests of the stock up on different platforms, because right now that now that we have Jenkins and I've it with it a little bit, it's very good idea to have jobs that runs tests on Mac on Windows, Mac Linux and see what's different, because there is always something different.

B

Yeah, that would be my idea to reframe it Laura.

C

Do you know if you can do like selenium testing with a a web extension or anything just like.

B

In past we had support at the SDK level that you were able to inspect if browser chrome like if the toolbar has the button that should appear when we define it in the API. That's no longer the case. We don't have access to those api's. They waited again.

B

Okay, like it was provided by alt as decay in Firefox, and it's just got the duplicate that we it's like a functional regression when it comes to a browser extension development, we no longer can inspect browser, internals I feel like yeah a year ago, maybe longer ticket on the like Mozilla, ripoff or web X tool that we are missing this functionality for integration testing.

B

They said that, like you, can run tests with things like mocha or maybe karma, but you won't have access to those low-level like browser interfaces and I feel that this is possible, but it requires changes in the browser like adding some additional restricted, IP api's. That will let you to access those things, but only when, like this tool, web xed is running on the same machine that your browser extension or something like that. It's executed I feel it's a security problem.

B

They explicitly made to block browser like web extensions limited and they don't want to like go back to the ecosystem. When you, you can just read stuff from your disk or access some random memory. So that's that yeah well.

A

I was just gonna say, that's probably deep enough, first, the history of with extension testing, oh yeah, and just to get it back to desktop sorry. But the point remains.

A

An Rick's point is totally valid, like we're not going to copy this. This okay are across, but there's going to be we're going to want to rebuild web UI very test, driven like we're going to want good coverage from day one because we're gonna we've built our prototype versions. This this one is like the production grade, one where everything is tested and then with desktop as the wrapper. We will want tests for for the desktop specific functionality.

A

So we will we'll definitely want a test for the the install flow not on the Installer per se, but that first new user experience of if they don't have a an IP FS repo things that are slightly I'd beyond the scope of the web UI. We would. That would be amazing to have tests for that and it's yeah, like that estimate, reeks for different platforms, free.

B

Yeah definitely.

A

All of that, but yeah not right, not this moment Burford like goals goals for key to we're going to want that yeah.

A

Yeah so I mean on balance, you know some of our ochio's are done and some of them aren't and but I don't think it's a it's not really about that. At this stage this was about finding our feet and getting used to the okay, our process, a.

C

Little said I, it sounds as though we were a little bit overly ambitious with what we're trying to do there just to give us a awesome. No.

A

Totally I, partly that I think also this we were more interested in building the apps that we already understood than generating user stories, and that's partly because of you know who we are, and developers.

B

Thing is that we were still like. We were building a team and, generally speaking, the okie are processed world's best when it's like ongoing process that you have at least one water back to look back and ponder all right. We had like yeah so I feel we did not do bad, given that we just started doing stuff and though we did explicitly now what has the higher priority than others.

A

It's really I think we did fine and we're gonna do better next Q next quarter.

A

So seeing the first two goals, the first two objectives: we have a shared understanding of how IP, if I, should work in browsers and browser developers understand requirement at the center as web I think those objectives we would probably copy across then three and four are more up for grabs in terms of how we nuance them. How how do we express our kind of focus not wanting to build web UI? We want to spend most of our time building web UI together.

A

So we want to sort of chain. We definitely want to change that. We need to reduce the number of key results and that kind of thing, but so just.

B

Thinking like the challenge was to remove the last two or maybe merge them and reframe them around web UI, because, like web UI, related work who touch both desktop and tons of extraction and even like the end users experience.

B

Maybe we should like create an objective to deliver like initial a tower or something like that of web UI by the end of the quarter like working minimal working example that takes advantage of web ipfs object and stuff like that. Yeah and right key results around that yeah.

A

Three months I reckon we should you should you should go bigger we should have. You should have a really good, whether you are deployable and go ipfs. That would be right. You know this RIT, like I, should at least be have the feature that we have to define. What's, like the smallest feature set for it to be worth replacing the existing web UI.

A

Whether do we have to have a one for one with the existing features before it's ready. Oh yeah, there's not much right now. Do I yeah yeah exactly if it did its job and it's tapered like if we make it better and just do those things I just.

E

Do as you can see about 20, why do you think it would be better to build it from zero or to reuse the current code, because I think it's a bit confusing the code.

A

But my assumption would be this would be a clean slate and that we would we would use the existing implementation as the as like a reference for. What's the minimum feature set, the code would start with a fresh repo, I. Think, okay, it's better I.

A

Think so too I think you know it's a bunch of other people's decisions that we're now gonna have to take, take ownership of and maintain and feel confident in. So you know we can.

F

A

We can do all the hard work as long as you get to have the fun as well of starting a new project. We yeah.

B

Like, maybe it's not a proper time to ask this, but should we create a new repository, or should we just work on my separate branch, like personally I'd like to keep the repository just for like historic calls that we like there's a lot of issues and discussions and design decisions within existing repository? So that's like an argument why we should create a new branch, just really remove everything or just start a new branch from scratch right and at some point you just replace master with it. Yeah.

A

Formula but we will I guess we'll know better once we've been through the wire framing process, but I think that's the general feeling yeah.

G

E

And should wait, do I depend of your I depend on BFS implementation or should web UI depend on companion in this window? Ipfs. That's.

A

A good version I think it I think what you are is gonna depend on window to ipfs yeah. So.

B

But that should be like the default and that would remove the security concerns a lot of people. So if we make web UI default to web ipfs and actually to only work with that, that would be fantastic because there would be no direct interaction with IPF. A like api port and I feel that that's very, very important. This.

C

Work, it's worth mentioning that it's not somebody using window type EFS, we're not with web UI, isn't depending on companion in that way, like that's just something that companion provides and- and this is exactly what an electron will provide- or you know the desktop will provide in the same way as companion does like it's. Not it's not tied to companion. If you see I mean, but it can be used with companion like you could just have companion installed and while we're diving it, we can use window to ipfs there.

C

But you know when it's released in desktop the electron wrapper will provide window type, FS yeah, and then you know it could. But I know we're saying it should only use window two ipfs, but it could fall back to using JSI DFS api and maybe even initially so that it can be, but just be bundled with go ipfs as it is right now.

A

Yeah absolutely I think we would the transition path and the deployment and support what what scenarios must be supported and what, where we'd like to get to it, can be a transition path, so like only ever use window to ipfs, but I think it would, as Ivan said, would bundle gaius, ipfs api with it. If window type in us doesn't exist and used before that yeah.

B

Like do what is happening right now, just refuse to work with api. If it's not running on localhost right and perfect one to one replacement, yeah.

A

But we have drifted from the agenda that feedback, so I think we've covered how we got on with key one and basically, where we're going for key to a high level. We haven't managed to come up with the wording of the third objective, but I wasn't expecting to get all that done on this call. Anyway. I think we've got a sense, it's basically web UI and how we what's our objective for web UI is essentially what we're talking about. It's that people agree with that: yeah, ok cool!

A

So maybe we could hang around on ILC after this cool and just thrash out that third objective. That would be useful, but there's some other notes in the agenda about feedback from other working groups. Are there any oakley ours that require support from QA testing and dev team enabled or any other working group yeah.

B

This is something I wanted to highlight.

B

Trust me to ask the entire working group is there anything related to like infrastructure testing tools like Jenkins without out of it amounts of websites that we need help with and so far I identified?

B

This need for providing working solution for shared secrets in Jenkins builds that would benefit of ipfs desktop for like application and installer staining, as well as browser extension, because if we had this secret solution in place, we could have as dedicated better branch and every time we merge staff to better branch, better would be automatically signed by Mozilla and it would be automatically deployed to like Firefox, store and chrome store and that way, if a bus hits me, you can still do a release. Oh no I think yeah I feel yeah. I.

B

Think Davidge also has access to even Zilla store, but still just.

A

Don't get hit by a bus yeah! That's our current solution as well. Snow.

C

B

Like don't say that yeah, but going back to this request, do we have any urgent needs when it comes to to like this infrastructure infrastructure.

A

Qa, oh geez, Jenkins a ton in the past, but it'd be good to get I, don't know where I go and find out what our particular Jenkins setup supports so be good to just get an introduction to it.

B

Yeah, like so I, think I like provide feedback to victor Judd. We would love to see some instructions how to set up like builds for different platforms, right yeah, Allen.

C

Yes, I just had a suggestion when it says QA I, just thought: it'd be rad if we could get like a whole bunch of people to use some of the things that we're building and give us some kind of feedback on the UI like how it works, what they fought about it. You know that.

B

C

Is that fall under QA III.

B

I think it's mostly this QA small from the technical side like right.

C

B

We can we can and we will do organize some people when we write blog posts that advertise our better channels and I think is.

A

Really a really good point: when we release new web UI new desktop new IPS companion, there's an assumption that other members of ipfs will test it we'll try our we dog food it, but we should probably make that a bit more explicit, like it'd, be useful and also undoubtably. Some members of ipfs will install it and do test it, but it'd be nice to make that a bit more explicit in terms of like we would like your feedback now in this document.

A

Rather than have you raised an issue in three months time saying you don't like it, you know, you know I mean like if we, if we are more proactive in seeking out feedback on our Terms when we're ready for it, then we'll get a bit less of this, like oh I, don't like tabs, or you know like feedback that that is indicative of a problem, but isn't quite as useful as feedback at the right time in the process and I.

A

Imagine we'll also want to include that at the wireframe stage, and no doubt David dears will give us plenty of feedback. Once we've got some usable, wireframes and designs, but just thinking about it as proactive rather than reactive, let's seek it when we're ready for it. Yeah.

B

Maybe some kind of a process within protocol labs, or maybe core group or generally protocol upstream, that encourages dogfooding I.

A

Think everyone's totally up for it and as long as we're a bit noisy as long as we make a big point about it at the right time, I think we'll get lots of good good information.

A

I can't think of any other I, don't really know what all the working groups are office of. My heads, I, don't know what we need from the rest of them. Yeah.

B

B

Yeah, like I, had the discussion with Kyle regarding public gateways, but I think that work will be kind of parallel to what we will do with sharing UI this quarter, so I don't feel think we will need anything from each other. At least.

B

Nothing comes up to my mind at this point, but if you like have any ideas, what we may need from other groups should free to raise that there are things like rate, limiting bandwidth, limiting and creating statistics, or something more than an entire repo that we could ask and communicate to, let's say, go ipfs team, but I feel that may happen at the end of this quarter after we have wireframes, and we have more solid examples of UI that we want to build that we can't fully built without those API sort of features of conservators I feel.

B

Maybe next quarter will have some solid, specific key results.

A

Demand demand, oh yeah, yeah, I, think I think that's very good I think it probably worth cueing them up now. It's just like I think I think just knowing how much bandwidth the specific CID has has have.

A

You have uploaded like what's your things like this concept of your ratio of like if, if you're hosting a public data set, and it sends that to be super popular and it's it's like the the peak of your bell curve of like where's, all my bandwidth going and it's like, oh well, it's because I rehearsed the xkcd archive yeah, it isn't everyone. Everyone loves that yeah.

B

A

B

Should also maybe, if we have some ideas like that, we probably should very early. Do sanity check with back-end, because, like some things that sound quite easy may not be easy and they could provide us with not as good but fine alternative to that. What we want to build so I think communication is quite important here, so that we don't run into situation when we ask back-end for something that requires like huge refactoring called huge performance penalty, yeah. So communication.

A

You strongly agree, but I've been plopping information on issues of go ipfs as a kind of first step to that. But okay, so next agenda item ipfs desktop release plan. We've only got five minutes left, but I think we've already touched on web UI revamp discussion at the bottom. There I think there's a release planned I. Think.

F

We all have yeah.

A

Yeah, exactly yeah I think I was kind of saying: we can't release it because it explodes when we add millions of files, but I. Think nrx point is the version. 3 version, oh god, sorry explodes when we have millions of files, so we haven't broken anything yeah.

E

The brother might be on going BFS itself and not on this topic. Might.

C

Be I'd seem to remember there being a problem with adding loads of files or but I couldn't find the issue, so maybe I just completely imagine Zach yeah.

E

G

Cinderelly yeah all right last issue: yeah I just send a solution. I.

A

I think we should release it as well. I might just conscious Lidl tends to be very cautious in his releases and releases them in a beta channel, and then we wait another week and then it's still in the beta Channel and then zero bugs get shipped to users, and it's great so have has there I guess everyone here has had a go at installing desktop from master, and it's not wise fine should do. We should probably get a handful of hours to write master.

A

Firstly, it's just trying to get a defined release process for desktop that similar to what companion has currently I'm so little. Maybe you could talk to it and reek about what the whereabouts is your launch checklist. Is it documented.

D

On a repo and something.

A

That we can look.

B

Like in seconds.

B

A

There's any other relevant for desktop, or is it mostly.

B

A

B

Yeah, like we have translations, we have some I.

B

Think orchestration around the way was literally signing it happens. So it's like very.

A

I think so for desktop. We've installed it on OSX we've installed on Windows even saw that on a Linux flavor would net. So now would be a good time to cut a video for release and then just tell like announce it on the All Hands tonight and then get ipfs users to install it like it. I give a screw seemed to install it. Yeah.

B

Yeah, like it's in two hours, so I do you want to sprint through every lease? That's fast I, don't.

A

Know I mean my.

B

A

It doesn't it's not quite the same ceremony as far as I'm aware, like it's a tag and I get hub release like what's the button of bundling process, yeah yeah, yeah.

B

A

No I, don't I, don't mind a fast. We do I'm, just taking an reeks positivity and running with it. Yeah.

B

Sure so, let's say let's try your day, but if you will see that it will be a rush job, then we can safely push it to the next week. I think.

A

More more to say like, let's start, that's whatever is I, don't I've never released desktop, so unassuming henrique I'm, assuming that you're gonna do it like do the do the release Intel ever oh yeah,.

E

I can go to the to the ipfs lens today, but I I will release the because it's up 0.4 and I trust. You wanna feel to tell everyone: okay, yeah.

H

A

Sure will you be able to write some release notes for that. hmm Will you be able to write the release notes for that I already.

E

Already wrote some stuff, but I don't think there isn't a lot of things to talk about in this religion. It's some bug fixes and we were just using em FS on the first step, I.

A

E

A

Said I think that's quite a big one, because it means if they've used desktop before then. Everything they used to seeing in desktop will no longer be visible. Yeah.

G

That's wonder business with me, so that's.

H

The headline yeah.

A

So I would I would take a moment to review how you want to do this release I, don't know if there's an e I, don't know. If there's any code, we could write to streamline that process because it seems pretty surprising and I think a bunch of ipfs team will have at least installed desktop at some point.

A

I, don't know, I, don't know the best way to manage that right now. Actually,.

B

B

The change between 0.3 and the current one I'd say is a big one when it comes to stability and she's at the meeting I, don't feel that we should be skittish with releasing it. It's like a different use, guys done in companion of custom like in companion. We have thousands of people like over 500 in stable channel in Firefox alone, so the same from so I. Don't think that many people use stop yet so we can be more safe, yeah Alan to.

C

Two things I just thought of it would be awesome if we had some stats on how many downloads we have for desktop how many current users and stuff, but also the my other potential high discussion point is: should we just release it as a major version number seeing as it's a big breaking change.

A

One point no I.

B

Don't think so.

B

I appreciate the low number it sets the proper expectations.

A

I think, unfortunately, by peel assembler, yes Alan but privilege in the sub 11.0 numbers as beta work, layer you're not quite sure what the app should do and everything everything is there breaking change. Surprise, more I!

A

Think more concretely, though I take needles point you shouldn't be too worried about it, but I think hearing those words makes me a lot more worried because we actually have no idea how many people are using I'd give this desktop and- and we know that we're about to break it for everyone, so I think probably we should be a bit more or considered in how we announce this change.

B

Actually, I just came up with idea of like key results or some feedback for other working groups. How should we, generally speaking when it comes to ideas, projects or protocol labs? How should we track, or even should we track how many users are like installing running our software? Oh hi Kyle.

B

We were starting one hour earlier because of daylight savings time and I think I feel Kyle in using our from activist community calendar.

F

Never near me, yeah.

H

Yeah we can hear you.

A

Little was just expressing concern that perhaps you did not get the memo that tons and changed.

D

Sorry seven yeah, like we have a summer time right now, so we moved this one hour now.

F

My calendar said a sir: that's a crab.

H

Yeah, it's okay! We just have no permission to update the cracker. Sorry for that we are recording.

B

This going back to the idea is that I was asking if we want to know how many people are using our browser extension or our desktop application. How can we track that, and should we track that I feel that's a question that someone else could answer globally for like protocol apps? We have some privacy concerns these days.

B

That would be nice to know how many people are using that our applications. We can tell how many people are using stable, Firefox version and chrome versions, but we have no way of telling how many people are running better channel, for example, on Firefox. So that's something I will probably raise and ask other working groups to address I. Think.

A

It's a reasonable question: it's more along the lines of what would we use the metrics for and if it's to make decisions around how fast and least to be with releasing, we can probably just decide what our tolerances are for. How tested is a piece of software before we launch it without tracking people, though we don't attract anyone I think companion, with its add-on, store, download stats I think that's a perfectly adequate proxy for how many people are using it for now. I think we need to install any extras and I think it will.

A

It just has more potential to annoy people than it does to give us useful information.

A

But that's just that's just like my opinion, man, anywho, okay. We were just talking previously just as Koya right after we was talking about the desktop release plan and we went from saying, let's launch it immediately, Oh dot for two. There me being like oh way, if I forget I, just realized we're going to break existing users, view of the world Oh dot, three doesn't use. Em, FS and odo four will use em FS.

A

So we need some messaging to explain that to people or we need to decide whether that's a significant enough issue that we should write code to copy across things that your desktop app. You know dot three can see into your em FS I. Think.

C

It should be front and center at the very least, we could provide a script that might do that if people really want it, I don't know as a suggestion.

C

The yeah, we don't know how many people are using desktop. So it's difficult to make an informed decision about it. But given we're on a zero point release- and we don't think we don't think there are many people using it, we might, which might just be able to get away with yeah.

B

Like I would not spend many resources on that. Just put it in the bold font in the release, notes that we are now using him FS and it's up to you to me great and I think if you still want to see.

A

It I mean on the account to that. This is a this is a project. Lots of people are interested in if anyone's interested in oculus. It's entirely plausible that they've at least Troy DS desktop, and if we, if we release it, we'd cut any release now that just loses all your files or appears to lose all your files because, of course, they'll still be in your repo. You just want me to see them. There isn't a super great user experience, yeah.

B

So I propose we create an issue. Okay first does to people to Tory and to address this like should we create a migration script, and if this script is created, it should be like embedded within application itself, and when you install the latest version, it should migrate your files automatically. Otherwise there is no point in making it just for.

B

For the sake of writing that we've created the script, it's either automatically and transparent to the user, or we just put in the bolt that we move to em FS and the your files will go not gone, but there is a different view for data right.

A

Okay, agreed, that's probably the last word on that. Do you want to talk about ipfs companion, release at all? No.

B

Like like, it will happen, and my plan is to s to make a release candidate this week as soon as I emerge this PR from Allah regarding ipfs files, API in the name, scoping I feel that important one well, if you want to ship a release with a window, ipfs object, we want to provide this name spacing out of the box so that there is no breaking change when it comes to application that people start playing with writing with this API.

B

So release candidate will happen this week, I'm not sure if today before All Hands, maybe will take it slower, but it would happen this week and maybe a release a week after that, if there are no bugs there is some work related to localization I realize that there is a long story about localization codes for weird languages. The problem with weird languages is that no one reported to me that translations do not work, because one person is using the block out, so I need to sort that out, but that's like a low priority.

B

If no one was reporting that it's not working, probably everyone is running English version anyway. Still I wanted to have localization working before the release, so that's the plan and within two weeks will should have a stable release that it's just replacing the latest one I think.

A

I think your dedication to localization is super important and is only going to get more important as as these apps get better, and it's a useful reminder that we need to bake that in in some form to web UI. Next yeah.

B

Like as as long as you start with creating English locale, that's enough to be honest because lick people will come to you and ask for localization, or they will just add that pull request. Then, when you see there's a need, you just set up some mighty web interface for providing localization, but for now just just make sure the local localization strings are in separate file like there's some kind of framework and that's all yeah.

A

So I'm aware that it's been an hour just a quick, the only other thing that I really wanted to talk about was we had a really really productive session over IRC this last week, so the general thought was.

A

Would these weekly meetings be more efficient as a time that we all agree to be in the in web browsers, IRC channel just due to kind of language barriers and pine zones, and la-dee-da I mean we can't solve time zones, but at least I'll see a slightly less cognitive load at 7:00 in the morning than an IRC chat.

A

So so often the time delays in IRC. We just have to work around those because people check in and out at different rates, but if we had a kind of time window, if we had a time a start time like a meeting, but that was conducted on IRC. That was like this and had an agenda, but we didn't have to sit on video.

A

um It was just put forward as a possible alternative to this, and I wondered what people, if that people were interested in that, if they felt that that gets to counter like do, people feel that the video chat is important and useful, or should we experiment with IRC chat, a given I think it's more like being available for conversation at a static time each week, but I'd be interested today. If anyone has any thoughts against that.

C

Anyone, yes I guess we can always jump on a cool, an actual face-to-face cool. If we decide that's needed, can't we if we start off with IRC, oh.

F

Are you guys all based in Europe right now like on this call, yeah I mean, but the 7 a.m. context. Like don't worry about me, I mean I'll, get up as I. Don't care I get up at 6 a.m. anyway, so it's like not I, think of a deal for me.

F

If you don't mind the way my hair looks at least but um ya know, don't just do it because, like it sucks for people over in Silicon, Valley aka me, you know if you think that your cars are important continue doing them, but I will couple with whatever you guys want to do.

A

Super-Good um I think it's nice to get FaceTime every now and again, but I think it it right now for us, it's more important that people feel that they can say what's on their mind and express themselves, and that seemed to work really well and I'll see the other day just having like it's really useful that we're all there at the same time. So the conversation flows.

A

Got sherry people in the office?

A

It's useful, we're all there at same times the conversation flows, but it seems to unfairly privilege native English speakers because of the way that, like real time like conversation works, so IRC seems to be fast enough that we all got our thoughts across but slow enough that we all had time to think about what we wanted to say. So that was the reason for suggesting it.

F

Sounds good to me, the.

A

Cool okay, so we'll try so we'll have the meeting at the same time next week, I'm so 7 a.m. sorry, but we'll just be on the in web, browsers, IRC channel and we'll just add that to their calendar, whenever someone with calendar powers lets us, do it I'm now super aware that we've run over the hour, but if, if there's only anybody has to dash off nope Kyle was there anything you wanted to talk about, seeing as we in secret without you.

F

No, it's fine I'm, just I'm, just listening in right now. I, don't have any like specific things to add to the conversation Rena.

A

Did did you see the chat from I'll see over the course of the week? The the major finding was a consensus around focusing on a rebuild of web UI going forwards and rather than maintaining a separate desktop app to build web, you are using window, died BFFs as the communication mechanism, so that we can reuse this code, the web UI code in desktop directly yeah. That's that's.

F

Great I mean that's generally, what I think we should do yeah I guess that works for both you.

A

Know it just seems to be a massive win in terms of effort, particularly also around like documentation when we make screencasts about how to use this thing, doesn't matter whether you're a desktop user or a web user, it's the same same screencast applies yeah I, think that makes sense.

F

Yeah I mean you know. If it's not working for some reasons, then you can always just mean two versions right. So I I am looking forward to seeing the wireframes really been published yet or.

A

We are a designer down, a couture is ill, so I've been focusing on. Are we distribute yeah calm I just poured that to a hugo site to be in line with all the other ipfs sites, but it's now driven from data, so we can start filling out filling out the call to action uniform. All right. Do you. You had a good bunch of meetings with people last week. Did you get a chance to write anything up about what their, what their needs and wants worth from.

F

Like the serve like moment you from from web browsers and stuff yeah, yeah yeah, let me uh let me let me get that sent out. Where should I put that I guess in web browsers or in.

A

Web browsers, I'll see that'd, be fine. Okay,.

F

Yep, great and I've got a mailing list I'm setting up this week. Finally, so I'll get everyone invited to that perfect.

A

Okay, anyone got any other business, no other business. Then this has been the in web browsers working, great team catch up and I'll see you on IRC this time next week.