YUI Library Open RoundTable, 26 Sep 2013

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: YUI Open RoundTable September 26, 2013

Description

YUI Open RoundTable September 26, 2013

A

Table it's thursday september 26 2013. thank everybody for coming. They have a full house. Today. We've got clarence tilo and ezekiel, hiding in the corner. Derek me andrew everyone on the internet.

A

So um while I get things copied over for people to see the link, um maybe you want to start kick things off a little bit.

B

Yeah, what do we got on the agenda today?.

C

um Eric has been working on um getting pure working better with power and also, therefore integrating with other application code bases. So I can talk a little bit about that stuff. Yeah.

D

B

Would be cool, learn more learn a little bit more about firework.

A

We've got um futures that we could talk a little preview a little bit about next weeks around next week. Yeah.

E

What are we talking about next week? uh Why not storage and why not tv.

C

I can also talk a little bit if time permits about this. um This proposal for improving yui's gesture events that I sent to you, I can trip, got some good feedback. um I'm gonna start coding that probably next week, so I can talk a bit about what that would look like once it's done.

D

Actually, you should just.

C

D

F

Eugene, hey: what's up guys.

A

Cool well, let's kick things off with the demo stuff. Then, if you don't have any demos, we could talk about start off with you, though,.

C

uh Sure so um so eric has been working, he was unable to join, because I think he has another meeting, but he's been working um on getting pure and bauer better integrated. So basically up until now, if you ran bauer install up here, what would happen is that everything in the pure repo, like yui, appear everything that repo will get pulled down into your power components folder, so that repo only has source files.

C

So what you have to do is you'd have to run power, install pure and then you'd have to also run grunt to actually go into pure's grunt file and build pure well. This is actually not useful, because what we want to tell application developers is that you want to you only care about purest, build files. You shouldn't be fooling around with your source files at all.

C

um At the same time, we didn't want to commit these build files to the repo, because then it's just a pain to get everything to be synced up every time you commit source. Well, you have to commit build file as well, so um we were waiting for bauer to have some way to just like solve this problem, but um but there wasn't much progress happening in that front. So basically, what we did was we just created a repo called peer release and we only every time we have a pure release.

C

Basically, I think we explode the zip of that release and we push that ex that the contents of that zip to the pure release, folder of to the pure release repo and when you run power, install appear now it links to the pure release repo. So, basically, when you run virus up here, you got the pure build files now, so that's that's! Basically how all other bower packages work, and now you can just run pure power and stop here and then pull whatever build file.

C

You want, you can get the build files for individual modules or you can also get the roll up like the puremint.css.

C

So that's the first thing um and then the other thing that helps um integrate with application code bases is that um some like, for example, we want to pull in some parts of pure into iui. So, for example, we have like css grids, which is actually pretty outdated. The pure grids is more up-to-date. um We fixed an issue which makes pure grids now use flexbox um in browsers that support it and use inline, blog and browsers that don't so, that's something! That's the peer group, that's not in css grids.

C

So, instead of you know copy pasting that code over um we want to use. You want to basically take pure grids and somehow programmatically replace css grids with that. Well, the problem is that in yui everything has the yui three class name and pure google. Everything has the pure class name. So eric also wrote this a very useful, plugin or very useful run task run, plugin actually, um which contextualizes um you know, class names. So basically it runs through um it parts to your css and for every selector it hits you're able to define.

C

um You know what selector you want to replace that with. So you can define either a prefix to put in front of every class or you can do a complete replacement. So you can say I want to replace this string with that string so using that which would that run, plug-in actually is actually on npm.

C

Now I think it's called run css selector, um so using that run, plug-in um he's actually just just issued a pull request to yui like 10 minutes ago, uh one two, four zero yeah, which basically replaces you know it updates yui's run file to use that plugin and you pull in the latest version up here. So we we're gonna kind of um formalize this a bit on the pure website, and we haven't really called it out too much yet but going forward.

C

This is how we see people implementing pure in their application code, bases that also address some of the pre-processor talk that we've been hearing. Basically, people asking: um why isn't you theory is a pre-processor and you know is pure going to use a preprocessor and the answer is not I I don't think I don't think it is. um If you want to, if you want to use a preprocessor, what you should do is pull in pure dot. Css then use whatever free processor.

C

You want to overwrite whatever classes you want, and that way the benefit of doing that, for you, rather than you know, toggling variables in a preprocessor is that let's say tomorrow we update to pier 1.0. You can basically update right away. You don't have to worry about anything, whereas if you were using a pre-process and you mutated, pure you're gonna run into issues just like bootstrap. Just right. Updating from bootstrap two to three is a bit of an issue because preprocessor involves you, don't have that issue.

C

So that's some work that eric has done can't take anybody from pretty much all him, um but I'm just the messenger since he's in a meeting.

B

So with that first thing you were talking about with, um does bauer not have the concept like npm does of post install scripts? No, I don't think so. Oh wow, okay, yeah cause. That's that's how npm gets around that issues? There's uh yeah whenever.

E

You do npm install, I think that was uh for power, no, not yeah yeah. I think that was something that they were. I.

C

Think power is very simple in the sense that you just specified like it has a server with the registry of like when you say some keyword. What do I pull down from what give people everything yeah.

A

So it's really all it's using is a simple way to fetch things based on the keyword.

E

Yeah has that been something that people find requesting, though, like bauer integration with pure? Is that something that.

C

Yes, and the reason this is actually important is um what this allows me to do is, for example, people can now make you know little ui modules. So pure has you know some basic ui like it's not like you know, but let's say if I want, if I'm making a web application, I need to use alerts like. I need some css for alerts here. It doesn't have alerts.

C

So does that mean I should just discount and at the same time it's always not a fundamental part of my web application to the point where I want to, like you know, code the css for alerts like I needed to look like this. It's just something I need, which is, which is a very popular use case. So for that, for example, what I can do now is. I have this css extras gallery module right. Well, I can now use bower to say this depends on up here, so you just.

C

You can also go bauer install pure extras, and now you have period pure extras and suddenly people can start building on top of pure and adopt the same. um Some some of the some of the same class name, the way we like smacks concepts and the class name. The way we like you, know, write class names to have this cohesive, set of separate ui components which then you can just go power and solve this card, sell that for install that and suddenly I have a set of ui components which are kind of loosely coupled.

C

I only need I only fetch the ones I need at the same time. If I fetch a lot of them, then I have a pre. You know full-fledged ui framework or css framework. At least that's a good idea. Do.

A

Things like yeoman use, power or.

C

um Yes, yes, I think yeoman is a combination of bauer, grunt and yo, which is like yeah yo.

E

C

Which is actually another thing that happened this week is uh adios money and a couple of other guys released a a yeoman generator called mobile, um which lets you generate january scaffolding for a mobile web app and it's uh using pure. uh You can well. You can set options but like if you look at like it basically has the options of using foundation pure or bootstrap, and if you look at repo they're using pure for theirs, so it's pretty cool.

E

So how much customization looks like if you can you like, select specifically like which parts of pure you want to pull in? I don't think.

C

So I haven't used the generator, but I will look into that because I'm talking about that for my open hack, usa talk um but the reason we made it you're so small. So you don't have to think about that stuff right I mean you're thinking about.

E

One kv, which is 4kb.

C

Doesn't really make a difference.

E

So, if I want to extend here so all I do now would be, is it's just like put here in my like bower dependencies? Is that how yeah.

A

Yeah and you'll be able to see how it works so with nyu shortly, so you need to use it there awesome you do it do it before you build so you're, basically changing the source files. It only happens like when pure updates right, that's like how you could do it. So you see this, um maybe for other dependencies that yy might start using, like other external libraries, yeah.

C

We right now we have.

A

C

Though the external libraries we use in our handle is normalized, I.

A

Think we might switch those more over to the power style reporting as opposed to having copies in the. I guess. We still have to keep copies in the repo right, yeah.

C

um I think handlebars is an interesting use. I don't know, I think I think I don't know I don't know about handlebars, but for normalized. I think we're using bauer. Already I don't know, I don't know what happened where's. Eric would know more about that. I don't know if we're doing anything.

E

Yeah, I think I was talking to eric about that and no not yet like. um We want to have like an easy way that what we eventually want to like move to them. So now you can like just choose like whichever version you want, but right now we don't have that we just like clone handlebars inside of our customer.

A

Oh, I see, but it's not a dependent library. It's just.

E

We are copying him over yeah there's like a little there's a few changes that we make so that works with like why dot template, which like makes handlebars uh which uses, makes every template engine like uh follow like a specific api right. So it is basically pulling.

A

Down and then doing some post-processing of it yeah exactly. I mean that's what that's that's kind of unification step in the future, so that's cool.

E

Any new stuff on the yeoman sign, like I heard there, was also like a pure, pure specific human generator yeah. It is but it's not up to date, I mean that's. I'm.

C

Just of pulled in all of pure source, so um I really like this hand this new generator, that was, that was kind of the.

G

Mobile one, so I.

C

Think it uh I don't know, I don't know.

D

C

The details of it off the top of my head, but if you just.

A

It seems like it's more practical because who would just pull in pure by itself, necessarily, whereas this is like you're doing a task, you're building a mobile app? Well, the reason I like it is actually because it does.

C

More than just pull it in because that's what the pure generator just kind of just pulls it in and starts you think of link. um So I'm just opening up the github repo. Now, let's see um okay features, scaffold support for bootstrap foundation, pure top coat, generates responsive images used for source set, uses browser stack for cloud device, testing uses fast click to avoid io, essentially commodity for full screen. Api polyfill phrasing, global storage.

C

So it has a bunch of yield stuff in there.

C

It's like mobile, first and yeah synchronized across device like loading device testing in the cloud so yeah. That's.

A

What I like it has like, it seems like.

C

Pretty full-fledged and since adios money is one of the he's one of the founders of the field.

E

C

Basically yeah so.

E

It's a good vote of confidence yeah. So with like the generators you can also like make a generator like call another generator. So I think that's where, like a benefit of how computer generator might be so when I was working on like the mojito generator, for instance, um something that might be useful is, I can like say if you want to include pure, then it'll call out to the pure generator and then it'll automatically include pure with your mojito app. So there's like some like cross um compatibility over there. Oh that's cool.

E

And does it do any post processing too, or is it yeah, so you can select whatever you want? You can basically just choose like you have like these options and you can select like options for pure options for mojito, whatever you want, and these are all just with like either like uh like some like menus that you can do and it's all from the command line.

A

Does that mean you're going to add more features to peer now, because people can sort of pick and choose what they want or you're going to try to keep it small? I think we'll keep this.

C

Core small, because we're going to form, I think the next one is to formalize this approach. You can build components that depend on pure and that's how you build up this third party ecosystem. A few. I components.

A

Do you see that living as a separate repo, maybe yeah,.

C

I see that living as similar to how grunt plug-ins living out, which, like people, can just make plug-ins and then the pure site would just have a list of all these things that people make it's like. Curated listers like a list of all of them, maybe.

A

You could move the templates there too right.

G

Yeah so yeah, so that's kind of where I see it coming and we can actually launch like period 1.0 after.

C

Like some other capital.

E

One is there any cool community stuff about up here right now, um you see you mean like live like, like no sorry like built on top here, like other, like small, like css modules that build on top of here.

C

um Yeah there are things popping up all over the place like if you search your own, you get a whole bunch of things that people have built like people have built like full mirrors for sas and less using pure people. Like you know, building pokemon stuff, it's that's cool to see yeah. If you really want to use prefessors or in pure, you can always check out one of those mirrors that have you know, sass or less or whatever compressor you want.

A

But you're working on more than just pure right and you're working on yeah, something else that has friends.

C

With estrogen yeah, do you want to talk about that now.

A

Yeah you're welcome to go ahead and go.

C

Okay, um if you feel like you're done yes.

A

C

So here's only I'm only.

D

Working on that, for.

C

Like 25 at a time, I think um I've also been working on gesture stuff for yui. um Actually, why don't we come back to that, because I'll show I'm going to hang out and I'll chosen code that I've been doing regarding test results? Just oh cool.

A

Okay, um let's take a look here, so one thing I wanted to go over real briefly was the schedule that's coming up this friday is the.

D

um Scheduled page.

A

So friday is code freeze, which means um after friday, any any check-ins you make into devmastered f3x will be, will not be in 313. It will be in some future release, 314 or 313.1 or whatever.

A

So, if you have a pull request that you want to get in before, then you need to go ahead and have those by yesterday, because that was the deadline, if you're, depending on the three-day lazy consensus, otherwise you'll you'll need to have a reviewer.

A

Have that help you put it in and then october 2nd is the commit freeze, which means no more commits to the special branch that we'll be creating after friday, and that's the plural request deadline for that, and then october 4th is the final commit freeze and then october 8th through you should have a stable release. So that's pretty much where we are with scheduling um right now. We have a call for speakers for yycomp.

A

So if you have even the biggest of ideas for wyocom talk, um please go ahead and submit those as soon as you can, and the blog post right now has one on our we'll have a new blog post up shortly with the deadline which is now october 11th. So you have until then to submit a um request to speak.

H

Do you have something on your side? I think no not really can't think of much going on this week. Oh, but there are things going on this weekend.

B

Yes, yes, so this weekend is uh yahoo open, hack, usa, I think is the is hackiers openhack ufc, uh so anyways yeah, it's a open act day, so anybody in the bay area, as we've mentioned in the last few weeks on open route table anybody in the bay area um rsvp.

B

uh I bet you can probably just go, find links on our twitter feed as well as yahoo. Developers.Tumblr.Com.

B

um There's a post on there that has an rsvp link um and yeah, so rsvp and show up on uh show up on saturday throws throw some hacks together and um when surprises.

A

It says during this two-day event, you'll hear talks from yahoo's top engineers and special guests. You can use any language web service, api library or platform in your hack. The possibilities are infinite. You just you should just do the whole round too.

A

Rockstar judges will evaluate your hacks and prizes will be awarded we're thrilled to share our seat that our ceo marissa mayer was planning to join us for this year's yahoo hack, usa.

B

So uh so yeah there's gonna, be lots of. As I just mentioned, lots of talks um so yeah learn about learn about pure tila's gonna be given one about pier. um I actually don't think we even have any yui talks. um I think we're all just so busy working on other stuff. I think.

H

We're gonna be like in the trenches right. What's up we'll be there in the trenches like yeah.

A

It's gonna, be there.

H

Yeah and I'll I'll, be there yeah totally hello.

B

A

B

Be really special, so if you.

A

Have any I mean like, for instance, if you're just around and uh yeah you're not gonna, be able to make it to like some other conference? This is like going to a major conference without having to pay a huge fee and there's other companies there like do. You know like pinterest and pinterest? Is there.

C

A

An accelerator.

C

Over there absolutely is there, uh I was gonna, say tumblr.

B

But now we have oh yeah right totally yeah giant.

A

Yes, yeah, so definitely if you are and another thing, if you should bring your kids along too, if they're.

H

A

Remotely inclined in learning hacking stuff, so I know people that went to the previous one.

D

You talked to that they were, they came when they were like teenagers and they they got really impressed. So, okay, I'm ready to show some stuff cool. Let me click over here.

A

That's my tv announcer voice.

C

So um I submitted this.

C

Yeah yeah, um so yeah I've been working a bit on gesture events and yui and um basically my initial. What I wanted to do is improve our whole gesture event suite because it's kind of out of date, you don't have all the events that I think would be interesting to fire.

C

um So I submitted this little proposal to um to the yui, contribute um you know forum and you can take a look at it there like what I mentioned. The problem is basically, we only have the only events we have are flick, gesture, move, star events and tap, and arguably we have drag and drop events, but they kind of use a different api, um and I was I was comparing it to other libraries that exist out there. So hammer gs is one that I'm. uh I really like. I'm kind of a fan of these guys.

C

um They have a lot of different uh gesture events. They're very small they're very well tested um see.

C

Yeah uh is a good example. So if you look over here and you kind of uh they work across a lot of different routes, they support a very similar. You know set of browsers that we do so. You kind of go here. You kind of play with this. You'll see like different events, are getting fired on the right side and uh one of the things I like about this about these guys is that um they fire a lot of events, they're very liberal, about firing these synthetic events.

C

So let me just refresh here, uh if I let's say just tap on the screen, what happens? Is I get a touch? um That's fire on like a like a basically a touch starter or a mouse down. Then I get a tap synthetic event. That's fire! Then I get a release of internet, it's probably a little small to see. So you can see there. Then. If I kind of um you know, there's like swipe and stuff, I guess I get something smaller.

C

I get touch drag start drag, drag in release and also this drag right event, which basically is a kind of a sugar to say that I drag to the right. So I really like this, because if you compare this to what we do with yui, all you would see is a flick event right and then the user and the developer has to go. Let's figure out. Okay, did they flick left or right? So they have to do some math with with the whatever velocity.

A

Yes, we have velocity, but.

C

You have to do some math with, like oh, like I'll, puts off. What's the offset and like basically work with the event payload. That comes back. So um that's basically what I kind of touch upon here. So one of the things that I've been what I'm proposing is industry. We start off by writing these low-level synthetic events, and I I call them start and release, but I'll probably change those names, um the name that we couldn't discuss naming in this thread, but these starter, releasing events are basically going to be.

C

um You know synthetic events which kind of abstract out the fact of are you using a mouse or a or a touch device or like an ms pointer or whatever you're using so it abstracts that out, because right now we have an abstraction there.

C

Once you have abstract that out, we can have more complicated events such as tap and fold, which can use these under the hood, and then we can have even more complicated events just flick, or you know things that are dragged me. uh We can modify some of the drag gestures, module to also use these sorts of events. So I did come up with an example. I actually already wrote event release and event hold, just as a proof of concept.

C

um So basically, a whole event is something that's fired if you hold down on the screen for a given amount of time and a release event, despite whenever you release uh wherever your mouse or your finger releases, the screen, so um it kind of doesn't have to be in a whole state first or releases happens whenever well. This happens.

G

C

Right yeah, so so that lets you do kind of things like this, where you can kind of like um hold- and um I don't know if you can see very well in there it's kind of like, but if I, if I hold it kind of and then the whole event fires, and then you change to red right.

C

So here, if I keep holding down, I got this little like animation, it's kind of very choppy there, but it's just you know like it's that ios or animation, though I click this and it kind of stops, and then, if I hold that for a given time amount of time, I get this tool deal that comes up, and so I wrote some. I wrote these and I wrote like unit tests for them.

C

They have like pretty good unit test coverage like I made musician, um just as a proof of concept and um based on the feedback that I got here.

C

I will like reiterate, and so on this thread I kind of outlined the api and uh how I follow. I think how I propose that they should.

G

Work and then I've got some feedback. You know I'll keep like.

E

Getting ready for that, but yeah, so we're interested in go ahead, so everything's like customizable you can set like, however long you want.

C

To wait for a whole event.

E

And all that exactly so, if you look at this one here, I don't know.

C

E

C

Yeah, so here it's a little blurry there, but no two on no two dot on hold um threshold 10, which specifies how much your finger mouse can move before all this cancels. So you're allowed a little bit of flexibility, duration of how long you have to hold for the identifier and then prevent default, true, um which will basically say the deal there is that a hold in in terms of browser events is touch start and then nothing and then waiting for the touch end well on touch star.

C

There are some things that happens like especially on ios, for example. I don't know if you know what, if you hold on something, you usually get some contextual.

D

C

Comes up usually, if it's on text you get the copy paste that.

D

Comes up and it's not just some buttons.

C

You get like open a new tab, open a new effect like that thing, so this prevent default actually under the hood goes in and prevents fold on the touch start to prevent that from happening. The con of doing that is that, if you prevent default touch start on something, you can't flick on it to scroll down the page. So it's a trade-off so uh based on what your use case is by default, prevent faultless faults. But you know you can say that.

A

Can you do a multi-touch or just single.

C

Touch so some events, so the we can well, for example, like camera js. They have an api from multi-touch it doesn't, it only supports you know, I think android 2.3, for example, doesn't multi-touch. I also don't know how that would work on old ie like on one mouse. You know like on desktops in general right, so I'm trying to first build up the suite which can be used across all devices and then go from there cool.

A

So, do you see this as an add-on to existing components in yy or a brand new module.

C

So I think what I would like to do is so these would be new modules and then I would also like to go into, for example, flick and flick, especially first because I think the flick module is the code is good but, like it's kind of it hasn't been like it needs some love. So.

A

C

uh What I think, what we need to do is even flick event: fires. We should fire another event which says, like flip, left, flick right, flick up, flick down to make it.

A

C

Be able to work with these events also, maybe clean up the event facade a little bit or sorry. The event a little bit uh give some more, maybe useful information, like so kind of figure, that out a bit um and then same thing with the gesture move events. I I'm hoping by adding all these events. People won't really need to listen to gesture move as much, because I feel that's. I don't like using those gesture move events because to listen to gesture move, you have to listen to just remove start first and then otherwise.

C

You can't listen to gestures from events, so it's just. I don't know it's not easy to work with, so those are some of the things we want to clean up in the api.

B

Yeah with uh having experience with the gesture stuff, I was going to ask what the difference between this.

C

B

C

Move event basically, like I think, just move events you should only use. Ideally, you should only use those if you're trying to promote some custom gesture. Okay, you know like that, otherwise it doesn't exist.

A

Like a circle or something like that,.

C

Look for if you're trying to look for a certain pattern, then yeah capture, those in some synthetic event and.

A

C

Some synthetic.

A

Useful for, like you know, with ios 7. now you have new flick events like back to delete and up to go away, and things like that. Oh.

E

So yeah I haven't like I've, read that blog post, so uh clarence just showed me some vlog, but you want to talk about it uh yeah. So, basically, what's going on, uh sencha wrote a blog post, they started testing on ios 7 and the problem with that was because of a lot of the new gesture events in ios 7. It's been screwing up a lot like javascript libraries that rely on like gestures, so.

B

It's something.

C

E

Us to keep in mind like when we're building these sort.

C

Of things yeah, one of the things I heard was, um if you are, or one thing that's in that blog post is, if you are less than like, 10 pixels away from the edge of the device or at the edge of the screen, and you flick uh in safari it'll. Take you to the previous page and like it won't let you prevent default or something right. So.

D

C

Breaks flick events very close to the edge of the screen um and there's some other issues.

A

With the screen.

E

Yeah I'll link to it on rc, but according to them, like ios 7, was like one of the worst mobile os releases since android free. So I don't think they really test much.

C

Like to see like oh, are we breaking.

A

I don't think that they don't have much of a history for doing that anything they do it's kind of like accidentally, good or yeah.

B

Well same thing with that we experienced with microsoft the or with ie updates too, is that any of the non-incremental web browser updates. um So firefox and chrome not too many issues there or webkit, not too many issues. But uh when you, when you go like years between releases on on this type of stuff, then that's when you just experience all of that breakage, but over the course of the last 18 months, as opposed to a little bit every couple months. It's funny because chrome is really good.

B

But if you look at android like the android sucks.

A

Like different teams are working on it right.

B

I think I think they finally consolidate. I could be wrong, but I feel like they've, uh because they've known that's an issue for a while, so I feel like the horse is bolted. Right, like you have to support those so they've been yes, they've been wanting to get uh chrome as the browser on android for a while. So I think that might be the case. I don't know I'm an ios user myself, ezekiel.

A

D

Know anything about ie11 is.

A

There anything.

D

Good new, with accessibility on that side of things. Well, I'm not.

A

Even on ideology, another's a new, so that's that's one thing in the news is: there's a new um beta coming out for i11 for windows, 7 that you can try out now. The.

C

Good news is, from now from i11 onwards, is auto updating.

B

Just a screenshot of that auto update defaulted to check. Yes,.

H

So yeah that'll be really nice. If only we can go back in time and like put the check box on i6 instead or just prevent ie6 from ever being released. That might be an.

A

Easier, well, you.

H

A

The I was around isaac's came out and at the time it was really an awesome browser. There are a lot of things about it that were innovative. You know, like css, was really uh progressive at the time. I mean it's easy to kick at it now, because it's what 10 years old or something, but at the time it was, it was better than anything else that came out yeah. If it wasn't for ice6, we might be testing on netscape today, so yeah.

E

So basically are with firefox.

A

Well, the thing is that, um just being from netscape um part of the reason why firefox was spun off was as a response to um just how nedscape was kind of you know, pulling things in the wrong direction, and so the the team said hey, you know we'll continue to perform escape, but we're going to build this new thing and make it open source. And now you know up.

G

D

A

Closed source yeah, that was a.

D

A

Huge release, so I was going to ask you one more also about uh your contribution, open hack, usa, stuff.

C

um Well, yeah we're going to.

A

Be talking about that.

D

C

Going to be talking about building beautiful apps using pure css, um it's a 20 minute talk with like 10 minutes for questions, and I'm going to be basically talking about everything in the pure ecosystem that I have covered today. So.

G

If you were here today, then you.

C

A

Any other talks.

H

A

Way, you kill two.

G

Breaths, just being honest,.

A

Now you can just ask targeted questions, now: watch the video and then come straight and ask your questions really yeah sure.

B

These socks are going to be recorded, correct.

D

B

Think so yeah! So even if you're not around, you can uh catch it online afterwards, yeah.

A

So in terms of other things going on, so we've covered uh the schedule for the release stuff coming out. We talked about why why conf call for speakers um if you haven't registered for y comp, yet please do get that going.

A

One benefit of happy becoming a speaker is that the yui ticket price is waived and there's lots of um the scope of your cops will be bigger this year than last year. We've got a lot more, a lot more organized and there's a lot of interesting things that will be happening during the conference that I can't really go into, but they're really cool and I've gotten a bit of a bead on the theme which is going to be really awesome too, which once we get sort of some of the like.

A

The theme has been produced and it's being worked on by like designers right now. So maybe I can show some previews of that next week. If we get a chance to see that um the only other thing I've got in terms of was there's oh I'll talk to you, a preview.

E

What you're going to talk about? Okay, so for next week, we'll be talking about a few things that juan and I have been talking about on irc, and that is um why not storage, and why not db so right now, like y dot storage, is basically just going to be like a general like abstraction over like a whole bunch of like storage systems. So you can use it with like um indexeddb local storage.

E

Everything like that and even maybe later we're going to be working on supporting like remote databases too. So that's something that we've been trying to plan out, and another thing is that I recently open sourced um a few of the some prototype code that we've made here at yahoo and that was written by satyan, and this code was originally used in like a project that was um supposed to try to create something similar.

E

It was so just to create like an abstraction layer on top of um a sqlite database for mobile devices, and so you can take a look at that code. It's called y.d.v and it's in uh one of the ubiquin trip threads. So there's a we want to try like merge. Some of those ideas take the best of both and maybe eventually pull that into decor.

E

So what did you think of that article since you bought the mobile.

A

Device like spinning.

E

Up the the wi-fi radio and back yeah, so that's pretty interesting, like using like storing uh more stuff, like just buffering like operations that you do for mobile apps and before you send out, and that's something that's going to be really useful, especially since, like the more like network um transfers, you do on a on a mobile, the faster your battery is going to run out.

E

So that's something that we're gonna take into consideration when we try to like make uh store more data uh locally and then that way like you'll make like less network calls and your app will feel more responsive. Because you don't need to like wait for that. What about things like pre-fetching, you think about as you're avoided, and that or think about that as well.

A

Yeah, that's something that we can definitely.

E

Do like you would just like so like um uh derek, like you showed me like uh one of the top spite of financial times guys before, and they basically like just download everything into their uh app store in like app cache or local storage and then from there. Their app feels like super responsive, and it's almost.

A

Like a native app.

E

Yeah, that's it.

A

Cool does this dovetail at all with like um eric's work on uh nested apps, and things like that. So.

E

Not right now, eventually wall starts to bring it.

A

All together, but these are like separate parts of that right, but you could see you can vision where, like some components of the app could be refreshed via the the data source, yeah.

E

So basically like, if you can't find something like locally, then you'll be able to get it through the network. Oh he's the only oh.

A

We just had a power shortage, pay no attention to the frantic wires cool, so that's next week and that'll, be you juan and kerry. So what's gary's role in this.

E

uh So we're thinking about like bringing some of the stuff in into mojito and um mojito next, which is based on some of the um express components that we've been open sourcing lately. So that's something that we'll consider like using um just like as part of mojito. So it's something that uh would be really interesting to have.

A

Cool and if you don't know about mojito, please go to ydn and check out what mojito is it's a framework for building mobile devices, mobile apps and other other kinds of web-based apps.

E

A

Yui is under the hood.

E

And if you check out like yahoo on github, you can see some of the more like um cutting edge stuff that we're releasing, which are more like. Even if you don't use like yui, for instance, they're, like just smaller, like express type plugins that you can just like plug into your express app and you can use like no matter what you're using on the front end. So.

F

Those are the components that we're planning on pulling back into mojito pretty soon.

A

Awesome, oh hey eugene! You mentioned last week that there might be some uh something new on the url front. Is that still in progress.

F

Yeah, that's still in progress, um we're basically being blocked by a team that moves kind of slow. So I'm thinking maybe next week.

A

Anecdotally, they were saying something about reducing the number of http requests about like 50 or something was that right.

F

Yeah, uh so basically from what we've seen, um if you're, only using core yui modules, it's more it's something like 75, shorter urls. But then, if you start using uh modules like yui modules, um application yui modules, then they start getting a little longer. Just because of the way that um they're stored on the cdn.

A

Let's go, maybe there'll be some feedback into that, so that maybe uh the future gallery components will be. You know, as as uh you know, stored in the same way, so they can take advantage of the same uh shortening yep cool.

A

Well outside of that. Let me look at this schedule.

E

Some interesting stuff that we eric and I started to talk about today, yeah good. So one of the things that we were thinking about is the idea of um starting to move yui over to use like uh es6 um style, like models, and so basically what would happen is that yui source code and any yui modules you decide to write will eventually start using like the es6 syntax and, if you're not familiar with that.

E

The idea is that you have like import statements and export statements, and so you would just like you import uh my dependency and then you would export like whatever object, that you want to export out from like that module and how we were thinking of, like approaching, that is using um an es6 transpiler, which is basically like um this program. That square has open source recently and with that you would basically just be able to turn es6 code into like a whole bunch of different formats.

E

So be it like amd common, js or whatever, and we started doing like some prototyping with that and we were able to get it to produce a yui model from es6 code and that's something that we've came across a few problems that will come up like when we start like trying to apply to the entire library like, for instance, we have a lot of like modules that directly like link things onto like the y object, and we also have like issues with some modules, like not really like having like explicit exports of the things that they're trying to like um that they're, a part of so, for instance, like um bass, requires attribute, but it doesn't actually use all the parts of attribute in it.

E

So we might need to like um do some refactoring of the yui code base. But this is the direction that we want to head towards in the future.

A

It's good. It also helps sort of delineate what what goes in core and what doesn't go in core, like you were talking about a while back about.

B

You know some things need to.

A

Stay in core and the widgets and things can be moved out that might that might.

E

Have some effect on that and the big win for this, like eventually when we get this when we plan this all out and we get it completely done- is that you'll be able to use like yui modules like just to be like completely cross-compatible, with like es6 models or amd modules, or whatever loading system you decide to use it with everything can just be used by anything else and that's what we're trying to move towards so.

A

We get um so we it's a two-way street, like other libraries, can use our components and we could use theirs as well. So it's not just us being asx compliant, also means that we get not only more usage by other libraries, but we can pull in things. If so, there's a new graphics library or something we could pull in yeah.

E

A

That's really cool.

E

What's the time frame for a lot of this at, like you know this year, still in the planning stages, we have like a very like preliminary preliminary prototype right of um uh having the es6 transpiler transpile, like esx code to your eye modules, but, um like we said, we came across like a lot of problems and there might be like some like refactoring to do in the main.

C

E

For that, so at the very least, it's going to take like at least the next few quarters, at least not until like next year, right earlier next year and plus, there's work doing done on loan or right. You talk a little bit about that. So with loader, um along with, like the es6 modules, there's also work on uh es6 loader and how the es6 loader works, it's sort of if the spec is still in progress, but from what they have right.

E

Now, it's very similar to require.js, where you basically list out all the modules you need inside of an array, and then you explicitly each of those modules has one export and for that for those exports they're. Just basically like you would just um have a specific like argument that you would assign them to, and so we kind of want to start moving loader over to that sort of system, because that's something that's closer to a spec and that's something that might be implemented by like native javascript engines in the future.

A

Oh so, eventually it'll, it can take advantage of the faster speed under the hood like.

E

For the browsers yeah, exactly so that's something that we're considering, but that's also something that's like a things that we're looking towards in the.

A

Future plus we have some some um some folks at the ecmascript and other standards bodies right, so we're able to kind of get tapped into that sooner rather than later yeah at the tc39 conference, cool yeah, I encour- I wish uh reed or eric were here to talk about that.

E

A

That's really great news, but for the future.

D

A

The future, um so in terms of like other things we had we, we usually talk about like that for grabs bugs, but we talked about that to death, I mean, if you, if.

D

You are a developer.

A

New developer and you are looking for something to try out, uh try your hand at you, contributing to the yy project, check out the up for grabs issues and on the github ios github repo.

A

We also have this kind of a query: that's um stale pull requests, and this is kind of like a chronic problem right now, where we've got pull requests that are either like eight months old, no one and their works in progress. We've got pull requests that are like a few months old and there's no.

H

A

um And I've been going through and like trying to assign owners or trying to get people that can pick them up, but if you also, if you are looking at wanting to contribute to yy other than just contributing code, if you could make comments in some of these older pull requests kind of ask questions and see if they're these things are ready to go and uh there's not anything uh new in terms of there's just um I could there are several these pull requests that I could just talk about like.

D

A

Like uh proquest 1175, look at that one, that's one that looks like it's ready to go but just needs some.

H

Owner to pick it up and.

A

A

Let's see not one.

D

One one seven.

A

Five, I don't have that many points, so just so, this is yeah. Widget medallion gets confused by clone masks. It looks like it's a pretty small change right, maybe I'm getting confused, but it has. It has tests in the head and it's got.

A

The change is ready to go. So I'm wondering if someone wants to take a look at that.

A

Or even if, uh in terms of like, if it needs uh feedback, there's something that needs clarification on.

A

The main thing is like this is this looks like a rep like a relatively new contributor, so making sure that they get feedback quickly, so they don't get discouraged. It's really important. It's been like two weeks, so you know definitely I'll take a look at it. Yeah, even just a comment like oh, this is great or you know we're still looking into it.

A

Yeah, so we need to find out like what's what's going on with that right.

A

Sometimes things will fail on travis, but it won't be. The the poor quest fault it'll, be like travis just was in a bad mood that day no.

E

No, so another thing that I kind of want to call out is um eric's been doing a lot of work on router recently, and um one of the things that we've been coming across is that there's been some problems with how the new router changes affect people using multiple routers on the same page. So if you use multiple routers on the same page, um we definitely like want your feedback on like if these new changes are going to like break anything in your app.

E

So that would be something that we would love to have comments on.

A

So do you hear that internet, if you're, still watching by now and you better, be then pay attention? So that's everything I've got um this. The like. I said the stale pull requests are just kind of a chronic problem and basically, if you're a pull request and you're like you know several months old, it seems like it's a lot less likely for your stuff to get in.

A

So if you do have a pull request and it's been a while and you have everything ping uh go on irc and and ping, one of the reviewers or ping uh being like go on there and see if you can get some feedback and get uh some response going. The committers.

H

A

Is in the repo, so you can also see a list of people who can can pull that in for you, you can ping someone directly and ask them. Take a look.

A

So thanks everybody for coming today and I look forward to clarence's talk next week and we're getting that much closer to you. I cough so and also thanks tilo for coming in flying all the way from canada just to talk to open.

H

G

B

G

B