IPFS GUI and Browsers Weekly, 3 Apr 2019

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: GUI and In Web Browsers Weekly, 2019-04-03

Description

Meeting Notes: https://github.com/ipfs/team-mgmt/pull/931

A

Welcome to the weekly in web browsers and I could ask you to sync up. As is customary, we will be working from agenda, so let me share that momentarily.

A

There's a couple of items of gender meat school time: okay, our discussions, but we shall start with a round of show-and-tell from things you done in last week, Michael top of the list top of the bill, and she shared that's the thing we did.

B

C

Okay, so we have not hear each other on this call for two weeks, so the first week we've been in this bond, doing hacking head planning and there's a threat for web browsers working group with our hopes and dreams for this quarter, and those hopes and dreams then had to be cut down significantly so things that we will be able to fit into a quarter. But if you are interested in my like entire list of hopes and dreams, dietrich published it on this fantastic thread. So thats related to oculus had some discussions around leonesse collaborations.

C

On the translation front, we were en s collaborations yeah, so like nothing solid yet, but we are probing ground on like taking different stakeholders in a secure system. So the browser vendors doubts some low-level platforms, that's already introduced in s, support and that this effort is mostly like truck by Arkady, but I'm helping him with.

B

Some technical advice for those who don't know his ENS, oh gosh,.

C

Yeah so E&S is maybe I'll share, so EMS is TV room service which lets you to publish a pointer to ipfs, Content, ID or something else entirely on the it better rock chain, and some software already shipped support for that. If you install metal mask, you are able to enter a tearoom domains which and with that e th that will use distributed blockchain service for instead of the classical dns-based, resolver and load web page for you who I profess compile, does not support that yet, mostly because we don't ship atrium clients with it.

C

But we are looking into ways we could add ENS resulting capability to software that does not ship with fully TM clients, all sort of like delegated ENS resolver. And that's what those conversation about it's just very beginning of hurting all the cuts and looking what is the smallest thing? We could collaborate together on yeah, so that's more or less moving moving to like translations. We added support for our Brazilian collaborators, who have some specific different language than people from Portugal for technical terms, and we will be.

C

We are in the process of splitting single local into two separate vocals added, some notes about alternative service header, which could enable us to support symbolize migration path, similar to the way HTTP 2 was introduced.

C

Those were sort of boring things that I don't have any visuals, but now I have visual for ya only. What.

A

C

Surface HTTP header that you can use to announce that this specific website service can be loaded using a different transport, which is very cool for web browser, supports it's not cool right now, because it requires like native support by the vendor itself. We are not able to like make use of this header from our browser extension, but if one of vendors shipped support, we that would be like standards compliant way of introducing migration paths and.

D

C

And like it's already like battle-tested fool around HTTP, 2 and origin and addressing location bar stays the same. So it's an interesting way. We should probably look at that eventually.

C

Companion in version 2.8 landed in stable channel, it includes visual refresh of main menu and preferences screen. Pinning is no longer closing this menu. It's like a lot of small visual tweaks still working progress. If you have ideas of like how to improve it, there's an issue when you can discuss that and already there are things that we plan to do next context. Actions are available on DNS link. So even if you temporarily disable redirect and keep location, original location, location bar, you will get context actions for websites that are backed by a key offense.

C

We shipped well, who shipped we open new. What UI with companion, which has very cool improvements on the file screen, so make sure to check that you are now able to opt out from redirect for a specific website. So let's say you want to watch something on d2 power or something like that, but for some reason it does not work.

C

You can set the redirect on you on that one website and if you are browsing the NSTIC website, it will switch between IP anna's path and original domain name, and there are separate toggle for disabling redirect globally. If you want to test something or just want to use other integrations, synchronized translations big. Thank you to all everyone who contributed translations and I.

C

Think that's it install it from the store it's on the stable channel available for Firefox, Chrome, chromium, brave, opera, all the usual suspects, and what alright I was able to finally find some time to fix, embedded J's gateway, so it now produce the Chabad extreme responses. I think I think I've added a like a longer explanation. What was wrong and what are the next steps but I think I will be able to give you a quick demo.

C

So I got brave running on my laptop and you can see: there's J's ipfs node under the flow training hero and if we like, open local gateway, so 1990 is local gateway exposed by the web extension itself, and this is an image that was loaded from that gateway.

C

It loaded Twitter of the problem, which is good and what is very good, is I've also managed to fix a bug in JSA PFS, which did not do streaming responses correctly because, like it was doing streaming, but the HTTP server that we've been using was catching entire pile out because it was doing gzip, compression, I think and it did not flash after each chunk. It was like buffering and tired of thinking memory and when entire thing was buffered, then it compressed and I think so.

C

I fixed that and after now, when it does the streaming after each chunk, it is flashing. The compressor and streaming response is working correctly, so I think I probably should be able to show streaming, but I think even cooler things to show is that now that we have like working gateway, we probably should should be able to just open web UI using this gateway.

C

Okay, that was yeah, it works, so this is a local gateway exposed by jacek EFS running in the extension itself, which is super super cool and you can see. I am able to like I've got a file here at it before I can explore stuff, and it's just everything is embedded I need to refresh this UI once it's, it's very good. That.

E

C

That fix that fixed everything yep all right, I won't take more time from everybody else, but that same demo, I hope.

A

That's very cool, so that's kind of our give us javascript in the browser pointing to polls. It's ridiculous, yeah.

C

We have like HTTP server in HTTP browser, which is ridiculous.

A

What musing from using companion, it's probably isn't the right place to dig into it too deeply, but there's a kind of UX and winds, which is like the the the idea of just always redirecting to the gateway or always using HTTP, it's kind of like never what I want what I want is like go for, go for the fastest thing possible like get me the web page and if it won't like, and so if that's HTTP initially, because you have to locate Pearson Baba like I kind of want that, like after the fact, also get this website into my local IP FS, so that then the next time I visit it.

A

It's super fact like the serviceworker kind of model, so that I can leave the I can leave. The redirect, enabled or I can like I can still get things from my local cache, whether available but get like not feel. The slowdown like I, don't ever want to have to turn it off for a particular website, because I'm annoyed, because it's taking too long and I know that the HTT will be faster for the first load.

B

That make sense, yeah yeah.

C

Like we, we could. We are detecting DNS link on web sites right now, so we could sort of like preload them in the background not sure like it's a tricky, because some people are using may be using it on mobile and you are duplicating data, so you've got the double.

A

C

A

You so much for today. Also so the way that beacon does it is it's like you got the edge to be one of it's like oh PSP version is available, and then you kind of opt in to that Jim. Did you have a comment? Yeah.

F

I think the stuff that Eric Lee goes ala, who's working on that's pretty much the whole idea per peer-to-peer progressive web app. So it's like you go it's. It shows up HTTP. You know it shows up. It appears like a normal website, but it's actually being loaded completely over IP FS. If you've got that capability, then if you're on your phone doesn't have that capability still works somehow so, but I think we're getting like with the companion.

F

I mean like if you got companion installed or if you're running brave, and it's just got it like it's gonna, be like anyone. That's got that capability, but yeah.

A

F

Definitely put together is pretty like iframes as a bunch of jokers is it's like. That's a real tingle, interesting.

A

This is just figure.

F

Out what what the clean way of doing that is going to be is gonna, be a lot of work.

D

A

This is just a proposal for the UX of ipfs companion that might be more pleasing day-to-day to this link. For the first time you get a site that is also as a DNS link. Then don't don't force that first visit to go buy at the local gateway. Oh yeah,.

C

So there's actually a separate issue about if we should read that actually redirect the nesting websites to local gateway, when we don't have capability to guarantee original is isolation, but that's a mouthful way of saying that we don't want to change like the domain name and I feel there's a way of making dots like of improving dots. Basically, we are able to detect if a request is for the mainframe, which is the page that has an address in location bar or is it for any of sub resources on that page?

C

So, basically, for DNS linked websites, we could change the default to not redirect that main request and keep the location, but all images, CSS JavaScript, would below that from ipfs. The.

G

Second, the.

C

Second time yeah, so that's an open discussion. There's an issue related to that. If anyone is interested to go to the.

A

Garage depends on how you perceive the viability or timeline of delivering sub-domain like local HTTP proxy within companion and coop events, I guess so: okay, that's arguably a better solution. Yeah.

C

Yeah, so the plan is like it's not very solid, but I think until we have that HTTP proxy, which creates a very nice domains in location bar, we will probably do what I just mentioned. Just don't really direct the main requests and keep the origin unchanged. There's a risk of mixing content from different origins, yeah.

A

It's not as it's not as clear cut, as makes it sound yeah.

C

A

Challenging all right that was a tangent but somebody's been rolling around my head next on the list: hey Fiona, do you wanna share them, yeah sure.

D

Okay, so how last week we were almost everybody in the team week. Listen! So that's that then this week I started to let it in Protoss co-learning a bit of you. So I did a bunch of reworking and restructuring of the project.

G

D

Our aim is to make it as simple as possible for everyone to add tutorials, and you had to to put some code in to make everything work, but the way I restructured things you just have to basically add some routes, because I haven't found a better way to do that, but I should be able to in the future, but then again, I think it's pretty better, but most of the work is done on this Jason. These tutorials you just I, can just show you it's easier right. Now we have a bunch of tutorials.

D

In our case, we have three. If you want to do with another one. You just come to this Jason.

D

Pick a new entry like or or something just going to.

D

And it should automatically appear. Okay. Everything is right here so yeah. This is cool because we're hoping that we'll get a bunch of tutorials for ipfs camp and as we make it easier for everyone, it will be easier for us too. So that's that's. The thing another thing I did was to I have two componentize a bunch of stuff here. So every tutorial is now component and we have now landing pages for each to them. So when we you previously, when you click here, you will automatically go to the first lesson.

D

Now we have like this little indexed landing page for each tutorial, so you can see what lessons are on it yeah. Yes, this is basically it's I would say it's a work in progress, but I think it's one more commit and it's ready to go so so yeah, very cool yeah thanks.

D

Okay, I'm kind of I put this under lock, but I am not really locked I'm, just waiting on some feedback to the web, UI chores that I have been working, but the everywhere everything is working and just waiting for feedback. So we can merge this and make it release it. I, don't know next up this week, I want to finish the is Roscoe commit and I'll start start to help Terry in the first tutorial and probably tackle some web UI issues that I I saw. That told me to.

A

Sleep good, that is good progress on the school, no ease all right. It's me share effect. No! Nothing super visual to share this week, but worth discussing is the we came up with a first draft of the hopes and dreams for the curry team for the next quarter. So if you are interested in what you I endeavor, that's can be done to improve I profess over the next three months.

A

You should take a look at this issue: nine to six on IP FS team management, and it has has I've proposed okay, ours, which basically is focused on protocol documentation, onboarding user experience, creating content for ipfs camphor-tree hope, it's mostly valuable to protocol and some improvements to the docs website and I can nest or io website.

A

One that we've been talking about a chunk this week is I, give us desktop being the default installed item. So there was a proposal on for the homepage. There was a proposal to simplify right now. If you click on install and I'd better stay. Oh, it takes you to a multi-step process that involves unpacking like download at our unpack at our copier binary into your path, which is basically a kind of unhelpful level of detail for experts and an unhelpful level of detail for beginners.

A

So there's a proposal to say well, basically, if ipfs is a command-line tool, let's let's suggest that you install it from your favorite package manager and make sure that that's up to date, so that the install step is one command, not several and more likely to work.

A

That way, we kind of demonstrate that IVFs is is, you know, has active community support in that it exists on many different package managers. That kind of lends a sense of credibility to a project and but also the discussion. Then we've done shouldn't. We just recommend installing IPS desktop. So there's another there's another iteration on this. So then the only caveat anything preventing us saying everyone just installed desktop all the time is most of our content is tutorials that involve using the command line and right now installing desktop doesn't also install the command line tooling.

A

So there is an open issue on desktop to make that the case- and this is our next priority for desktop- and you know, okay as Lydell and in recurved in thrashin that one out and I just you know net today that we should definitely start by copying whatever it is. Adam did so it dug around in that for a bit and I yeah. It's kind of neat, just it checks for their existence on first run and copies it to a shell script.

A

I think Lydell has some interesting ideas around not directly linking to the binary but having an intermediate wrapper shell script that can make sure that your mic FS command line is always pointing to the same binary. That desktop is pointing to which projects forward to a future feature that is desktop, should let you configure which version of my DFS you want to use, and that would include toggling between Jas or go implementations, which you probably is definitely an advanced feature. It's something that we're all internally interested in, but I think externally.

A

We kind of need to be careful about how we promote that, because it's the kind of cognitive overhead for the user that perhaps isn't interesting at this stage or at least needs to be framed in terms of what features it makes available anyway. That is a chunk of work that is going on to check out the old cars check out the antipas issue that is gonna, make installing things a lot nicer.

A

We just got. We just updated. God gave us with the latest 2.4 that for web, be a release and there's an O 420 go ipfs release that is going to rc1 in the next either today or tomorrow, and so potentially, by the end of the week, the new web UI might see the light of day with with gotta be best, which would be great otherwise, in digging into the code of JSC ID. The cornerstone of our defense see IDs and I'm, making breaking changes.

A

So if you care about see IDs in JavaScript- and you might want to look at pull requests 77 on multi formats, JSC ID occurs. The problem that we have is in Libya. We, if we get given a base, 32 encoded CID. Well, what we do is we take the user provided string and we pass it to the CID constructor.

A

So we say new CID, this string and in the current implementation of CID it coerces all see IDs like if you asked for the string representation back from the CID object, it always gives it to you in base 58, Bitcoin flavor, which, which has to change when we move to base 32, but is also, in my opinion, unnecessarily draconian in the sense that, with series flexible to the base encoding, so I've added a feature to CID, which means it preserves.

A

If you pass a CID string to the CID object study every constructor, it will preserve the base that you gave it. So when you ask for the strength for map again, it will use the base encoding that you provide it. What I mean what I mean by now, it's easier if I show it so what I'm proposing is, if I give a base, 32 encoded, Co, D string to the CID library today.

A

What I'd like to happen is if I later call to string on it that I get the same string back and or if I pass a base. 58 encoded CID, if I later cool to sharing I, think I get a base, 50 encoded version back, which seems kind of obvious way to say it like that. But up until now, that has not been the case.

A

So there's a pull request in flight to fix that and also to tidy up the constructor of clear cont because it was getting a bit hairy, because a few features have been added to it over the years and yeah neat was time to refactor. Anyway.

A

That's why I've been working on so next week, I'm going to be for the rest of the week, I'll be focusing on I've, already been looking at adding end-to-end tests to desktop, because a couple of bucks of surface this week that that's the jus the promise of using an electron app is that you get an own version of chromium.

A

So you have fewer problems with browser bugs because you know the browser that you're deploying on which, unfortunately, we haven't been testing on that specific version and it's an old version, and it has really interesting unique, bugs that chrome doesn't have didn't have like that. That release number particularly around adding directories of files again so right now in desktop, depending on how you try and add directories or files, it doesn't work again, which is great, so there's a like. We have to stop releasing the of the gooeys.

A

The core feature work if this doesn't work in them, but this this has good tests in web UI and they all pass. But when you load it up in desktop it fails because the version of chromium in electron has new and unique and interesting interpretations of the get WebKit relative path.

A

Non-Standard feature of when you drop directories on on webpages, so I have to dig into that, but I'm not going to dig into it and so I have a test that tells me that it passes because it's just getting embarrassing now and there's another one which I don't I. Just don't even understand yet, but the the graph Krong the iCub, the Explorer they lets you navigate back up to the graph. So if you drill down and drill down and drill down into an IP LD graph, you know pit Explorer.

A

If you use the breadcrumb in Firefox or Chrome works. Fine always has done if you use it in electron in chromium errors out, super cool didn't didn't see that coming don't have a test for that, so don't have a test in it. I can run an electron map and drive it. So yeah, it's been a fun week.

A

What else happened? Oh yeah I did a good write-up. If you haven't seen it, it's got a couple of photos on. If you have photos of that event that you wanna share, you should definitely set up a account with the textile, the textile photos. You can share the photos of that with us and that's me anyone else got things. They wanna share Eric's on the list. Great sure.

H

Not a whole store of IPF Estes did the Lisbon thing y'all and yeah I'm. As someone split across different projects, file coin has been one. That's been distracting me trying to iterate on that brand and give them give them some usable stuff and in terms of ipfs.

H

We have nothing GUI why's, that I've done. I am doing something ipfs wise regarding just designing some thumbnail, some thumbnails for kind of title cards for videos that we put on YouTube, trying to make some kind of a simple system for for Zach. The problem is it's essentially like postage stamp design. The idea is to to make things scannable in a way that you can tell what's different about each each of them.

H

Words are easy to read, or at least are readable, some I'm, making it I'm sort of gravitating toward a really simple, really simple layout the head that uses maybe some kind of a color coding overall colors ipfs, but then it uses a shard of color to identify. You know I DM my PFS GUI, etcetera, etcetera, not specifically GUI, but you know it's.

A

Like yes, I like the direction. I saw some previous iterations of that I'd like the direction it's going and one on one design challenge to add to the mix that you may have seen already is embedding embedding these videos in other websites YouTube ads the title of the video over the top of the slide. You can see the challenge here. These are relatively like easy to read clear, slightly title screens, but then youtubers has clematis, probably adding adding the name on top.

A

So you can configure that what we should do on this website is just configure the embed to not embed the title, but it's just worth being aware of yeah.

H

Because that's essentially, what we're covering our fan challenges in every direction and yeah next time playing the dive headfirst into getting a style guide under way.

H

If tachyons are the way to go, then that might be a good base for it. I'm gonna take a look at to the best of my knowledge, see what you know to the best of my knowledge about how ipfs stuff works, see whether I think tachyons is a good base, flexible and something that we might be able to customize without too much. You know without turning too much of it off and additionally, then you know taking a look at the couple of apps that we do have and then breaking them down doing analysis of things.

H

I might iterate on and different components that we might pull out of there and then they dropping those into a into a web page. Just a general outline of what a style guide in component library might look like just creeping towards something. That's clickable and living useful for.

H

Jung-Woo we played a little with bit, do go and Chris played a little with bit, which is pretty cool component like an automated component library, pulling out our react stuff, and that might be. That might be a call to them. It's really it's! It's really I. Did you go already Diogo? Sorry, you didn't go yet did you or did you know if you were planning on showing that.

D

I don't have that girl, because I was making that on the master branch, so I got everything yeah, but my fake points from it's every command is pretty similar to get and like in ten minutes we increase. We were able to create a new collection, but we made it private, but it can be involved with no problem. You made a new collection, exported ten components there and then, in our project, deleted a lot local components and used the ones from from it. I think the integration is a bit seamless. It's pretty good.

D

The only downside we had I, don't.

H

Remember I know: I wrote it up loading.

D

H

Learning the whole framework for each. You have two loads loaded like for each component, again: yeah yeah, exactly.

D

Yeah as we're importing it's about the global importance, we're importing patents globally in our project, but in bits which makes us imports if it doesn't work with with global imports. So in each of the components in the collection we have to manually me for two seconds and ipfs ESS, although the fungal or rather a bit, he said that they're working on a fix, so we can do global imports on collections. So that's awesome.

D

He said he'll pingas and that's an I think it's worth taking a look on bit: I like it yeah yeah they're, pretty responsive, yeah, they're yeah. They found this without even pinging them, so Mexico.

A

Very cool, it certainly seems like manage a component library. Is an orthogonal engineering challenge to pushing forward the d web and if, if it does a good job of it and I'm happy to offload that work on that yeah anything cool, Eric anything else or shall we we can who's next, you go Terry Terry, you wanna share thing: oh yeah,.

I

So mostly I've been watching, do go, be a rock star and crank out all this work. That's just on my to-do list, but I have done so.

I

There are links to the roadmap. Thank you. Okay, ours for protocol dropped in the note. So if anybody wants to look at those I think it was great to be together last week and getting some help on protocol, it's gonna be awesome.

I

So the thing that I'm in the middle of focus more on now and like kind of making progress, kind of just feeling overwhelmed stuck, is this file tutorial, which one of the problems I'm having is that there are two completely separate points of view on what we're teaching them in what order, because we have em FS commands and then we have all the none and my first commands that you can use with files just like the adding yet and most people had told me that they thought it made sense to do like there's dying, API, there's the non em FS piling stuff and then which was like this is really how I BFS works and then wouldn't it be nice.

I

If we could do all of this pretending it's mutable stuff, on top of it great then teach I'm a fascist though, and the Festus League it's the last thing you learn and Michael has been encouraging me to just teach em FS and treat it as like. The entry point you just think of it as real. Don't necessarily admit in that lesson that it's not and then later reveal under the covers. Like he's the dag thing, that's there as being advanced, he sees the adding get part on files as being advanced.

I

They would do it in another order. So I have that kind of like that conflict, going on to separate sets of commands to kind of teach myself to be able to teach them and my duck even my documentation. Reading skills are not great or the documentation. It's not create some. Some combination so I'm getting a little stuck as I try to make the code exercises. So this is the one I'm most successful with today, which is specifically like the M FS right command.

I

um You know so this one so that the upload thing it seems to be working here. So we can get files into the browser to work with and then we're, starting with just like putting a brand new file in and I have this working? If you do it correctly, and then you can see what you put in, but I have this issue where the way that these success and failure messages are generated, it's like if an error message is generated by the ipfs API that comes up instead of whatever I've told it.

I

To put so, for example, if I forget to put the file name and I just give it a folder which, in my beginner mind I, don't when someone says path, I, don't think of the file name as being included, so I would be the kind of person who would do this all right reset it to make that work.

I

So if I do this mistakenly I don't get the air I put in, which would have been oh look. You forgot to put the file name. I said I get this, which means nothing to me and if I make a different mistake, which is to forget to do this, create true. So it's just looking to overwrite an existing file.

I

I

I get file does not exist, not like the destination file. You want it right, you doesn't exist and also I had to change my files to like zebra to convince myself that this word filed, and it mean my word file. Yes, Ali yeah.

A

I agree: this is a big challenge for a school, because the current implementation of verification of code just just runs your code. It has no doesn't do any introspection of the actual text that you write in the editor. It just executes sit and then verifies the response, which makes it currently impossible to say anything useful about the code. You wrote because we don't we don't introspect it, so we would have to do.

A

We would have to look into passing the code and having an expectation of the actual, the actual ast, the abstract, syntax tree of the code that we were hoping the user would put in, which is definitely achievable, but it's totally on like it's, not something that protocol has looked at yeah. So the the feature that you need to be able to get good error messages is as a big chunk of work, but worth worth checking out and doing yeah.

I

So when I mentioned this to Michael I like given some screenshots of what was happening, it was like. Is there a way around this? Like insights, Frodo school, he said we could do a bunch of work to like hack around it and make our error messages go first or whatever, or do what you're saying the other thing we could do is if people agree that one of these two error messages I've shown, you is just plain old, not helpful. We could replace it with something we consider to be helpful in the API.

I

This is coming from the API right that.

A

Is also a good point so.

I

That's what like that's one option if people agree, but it's possible that you know some people are seeing this, who aren't trying to do the same thing and making the same mistake. I am so like inside proto school I'm, very confident about what the mistake is much more confident about what the specific mistake is. They're making so I have like I have built I built validation, things they're just getting hit. None of my validation is getting hit. It's just not actually submitting and getting to any of my validation.

G

I

This is one of the things I am stuck with, but hey I made code that passed and made it say it passed. So that is a thing I have done, which makes me feel slightly less bad about this project and then the other problem that I'm having when I tried to use the add command, and this may be in miscommunication or miss reading of the docs. But when I was doing this with Michael last night and he was trying to help me with both the ad one and the right one.

I

So I swear that, like two weeks ago, when I was mistakenly using like making the lesson and add and trying to without m FS and then trying to verify it using stuff in EM FS, which obviously the verification was not working. I had mistakenly written ipfs, dot files to add and then I thought I just passed in my file object as data and that it was like the verification, didn't work, but it like gave me back a CID right now.

I

It doesn't do that and when I looked at this with Michael, he said that he doesn't think the ad can take a file object, which is a blob. The documentation would suggest that it can't take a blob, because it's not listed here, it is listed in the documentation for right and I. Asked Diogo says that you're using adding it for shared ipfs. That is so clearly. Someone knows how to get a file object in here, but if it's a bunch of work that has to be exposed to the user, that would be bad Polly's. So.

A

We we discussed this on a previous call and in web UI we wrap it in a readable string before passing it to add its was.

I

There a way to fix it, API see, don't have so I, don't have to teach a lesson about wrapping file, objects and readable streams. Yeah Harry.

J

Let me just say one thing:.

A

When people are screen sharing, we can't see all the faces so yeah.

I

Feel free to interrupt me.

K

One of my okie ours is two things it's like. Oh.

I

That would be lovely. Thank you. It.

K

Will get them so.

I

Basically, what's happening right now is just because I made one thing: work in MSS and haven't made any things work in this one I think to make myself feel good. I'm, just gonna keep working on MF s for right now, I'm, making a Memphis lesson yeah. We can circle back to how we do this later. Yes, ally in.

A

Terms of priorities I'm, basically in agreement, it's my call like I, think.

C

A

S is a reasonable thing to start with. Okay I think the dagger commands are definitely advanced users only I would campaign in general for them. Nobody.

I

Somebody should've told me that the first thing I was learning was advance. I would have felt better about myself. Oh yeah,.

A

No, it's advanced in as much as its it's, not files like IP. This file, interplanetary file system, adding getting files like that's the whole bit I feel these is great, but I. Don't think two coefficients present that as like as important as files in this stage, Jimmy I.

F

Just want to point out, like I, actually think that ipfs add-in ipfs dot files. That ad is like PIRG anomaly wrong, like that KPI yeah.

A

F

Too easy to forget type dot files. This is a completely different API method. Then you're gonna spend like an hour trying to figure out what this.

G

Is- and this is a eight.

F

I

F

I

Because there's no ipfs step files to add in the documentation, sorry yeah.

A

This is a well-known UX blunder that is waiting for someone to you solve so, depending on which library you're talking about so the HTTP client now has moved the the files is now top-level I can I'm, not even gonna, say on the call, which is correct, because it's changed at least twice since I've worked here. But the point being. This is a known problem and we are looking for Heroes to solve it. You guys suggested that's gonna.

A

Be me me, maybe it's not currently on my okay Oz, but it's a combination of the of the problem of there's like this mismatch of entropy in old, original flavor ipfs. From way back when it was focused on, had put a file in get a CID out and get put a CID in get a file out. That's where it began and then later on, we were like. We need some kind of human friendly mapping for C IDs, because they're hard to work with.

A

So then, you get the NFS to mutable filesystem, which is just a view over a bunch of C IDs that lets you give human friendly, filecats, alias, is to see IDs yeah and, as everyone says, and then moving files around from ipfs dot files, dot X was it to ipfs top-level, oh yeah. It's the whole thing. The general consensus that we got to was that commands about files should be at the top level like they shouldn't be names based commands about. Everything else should be names based because we should make file operations. The easiest thing.

A

And no one to freak out every time. I say that. So what happened? Oh.

I

It's a good sign that you're not on the.

F

Wondering if everything should be in the same namespace, you.

A

Know what shouldn't be in the same? Nice Isis, dag an object and a little, but all those other things, because.

F

It is like certain when you first opened up the documentation page and you just kind of do one specific task. It's like ah it's like this. A lot they're.

A

F

The south yeah.

A

I would um I would extract dag and object to a separate command. I would make an IPL D command. That is something you can install if you want to add data to I. Give this I.

F

Think the proto school stuff is really helpful, because this is really going to force us to change things.

F

Derry's, completing fixing but well I. Think if you really want people to really learn these API, it's like they're going to go to this really high quality place and if everybody's crashing and burning when you go there, that's probably telling us that our API.

C

F

Is probably fine, but just the way it's organized attention that.

A

Happened once already with pro schools, so we did pro school in time for the D Web Summit and everyone was like. Why do I have to manually unpack, C IDs all over the place? This is the worst, and so we fixed that.

B

I

Know, that's me we can stop chatting about me. That's.

A

But we're talking about the UX of ipfs in general. The pro school is at the vanguard of.

I

A

Pointy end the thing that we can also have metrics, on which lessons to people freak out on and just quit, to delete their browser.

A

Anyone else got anything else to want to want to share Jim you've got a thing you want to do just.

F

Ddc working group did we do there: okay, ours, like everybody else, and just get the information out there. So there's. Basically, three people in DD C right now, so had your own ID or just gonna. Do I p NS, I BN s. Only you're gonna make IP NS fast and awesome and multi writer and all sorts of crazy things.

C

F

So we'll see how that goes. That seems like a big big job and then I'm gonna be doing everything else so and I'm gonna do lots and lots of little things. So hopefully, maybe I can do a little bit of stuff with package measures a little bit of stuff with collaborations and then do a little bit of stuff stuff for FS camp super.

A

Cool, are you the main point of contact for the DDC team? Now, yes,.

F

I, don't think I'm officially captain there anything, you know: I'm wrangled, OPRS anyways, give co-captain yeah.

A

The work attaches responsibility attaches to those who do the work, yeah random, cool, three minutes on the clock. Oh there was a thing on the agenda: Oh John, I taught Oh Hugo, of course, I'm. So sorry, you can talk, keep interrupting.

A

K

I will so what I've been doing Lisbon team week reviewing some part for lockfile stuff that never ends, but possibly this is the time lots of planning and no cure definitions for the web browser. Little brute continue the view support debugging of em flex in readable stream, and basically, we just was just gonna wait for the pull em plugs to be more John, Jessup EFS, and that will fix the most critical problem on view and yeah and I can s. Research I went through the DDC of yards and all our documents.

K

They have the the RFC from Mars, which is basically I've, been s over the units, which is not the way the delici working group is going to implement a Phoenix.

K

Ya hear that I'm walk on the pull implies request basically to finish or finish part one of my video for to cure and yeah I'll continue my IP NS research, yes Holly just.

A

Only that's not the way the Dec are gonna implement it. Last I heard a demon speak about it. He was requesting some metrics from the package managers working group to suggest to help them decide what the right solution for this is like. What are the needs of packages? How quick do they need openness to be? I saw it I, didn't Jim, okay, okay,.

F

So I think we should get like like an IP NS Tiger team together. So it's not like just Petra Nadine or two full-time resources, but like Hugo, should be on that tier for.

G

F

There's probably like other people, that's just everything, so it's sort of like I should be fairly large, so I don't think we have to do that. Okay, ours, on the spreadsheet, in there to fit into the particular groups, but I think so for something like IP NS that, like spreads across everything, I, think we need to figure out how to get people to meet together. Yeah.

K

I already already talked to either with the petrol a little bit and it's topped with a dame and yeah mmm my works more scopes to the browser context. Yeah. We need to talk with everyone and see what we're supposed to do, because from what the OPRS is kind of yeah I would rather hire implement, IPS or DNS. Then doing floods with.

E

Peanuts really but yeah, we should.

F

Talk yeah I mean I'm sort of talk to a Dean a little bit about like how it happened so fast in that protocol and that project- and you know it's basically the same things.

A

Jim, would you be happy to bubble this up to the project project, cool yeah.

F

A

Double-Check with petrol and Adeem that they're happy to be tying it up, yeah.

F

A

It's a good ideas, good suggestion, you just need to raise it with with project level, mm-hmm.

F

It's just gonna be like the site like you. Do you want to do video meetings and Raqqah synchronously yeah.

A

Perfect anybody anybody you got anything else. They want to share we're one minute over. So no is a good answer.

A

There's a thing in the agenda that says this call has moved back an hour for many people whose time zones have shifted relative to UTC. Is that a problem? Anyone as this calls seem to be healthy.

A

There's, you know, there's no deitrick, but I assume he's that's not because of because of time zones, Diego's, tolerant, tolerating it with a 50/50 Hugo, still looks happy all right, I'd.

I

Be happy to push back to the error we had to that, but this is not impossible.

I

A

Good I would like to keep it locked to UTC and just deal with it and then and then, in the same way like we should fix the API like we should fix our individual countries sluicing around relative to UCC.

I

Yeah I have a lot of control over my government, as you can tell us and that'll be enough. Emma all.

A

Right, oh well, we should have a chat offline because there's some other things I think we should get fixed really. The.

F

Whole West Coast of us in Canada are gonna switch I. Think.

A

Portugal is was talking about it yeah in two years time, two years time so Europe has kind of said, we're gonna stop dictating whether we should all do it or not, which is just going to be chaos, because it's like we should either all not do it well, we should all do it, but now it's gonna be like every country should just decide just.

E

Or just leave Europe right.

A

Okay, oh my trunk. Let's have a nice evening thanks you go. This is very idea based weekly in web browsers and greetings in cool and protocol and EDC.