Supper Club × What's New With Astro with Fred Schott

Welcome to Syntax.

In today's supper club, we have the first ever Three time guest on Syntax.

This is Fred Schott. He's talking about Astro once again. In fact, Last time we had Freda on, it was for the release of Astro 2.

And now, as of, you know, very, very recently here, Astro 3 is hitting the, hitting the ground. By the time you're listening to this, Astro 3 will be, ready to go for you. So we're gonna be talking all about Astro 3 along with some of the new cool stuff coming down the line from the Astro team as well during their launch week. My name is Scott Taliska. I'm a developer from Denver. With me, as always, is Wes.

And, this episode is brought to you by [email protected], the perfect place to track all of your errors and exceptions. Maybe you're picking up a new platform Like Astro. You listen to this. You say, I wanna write some Astro, but you, aren't in aren't experienced with it quite yet. So you make a couple of mistakes here and there, ship them to production.

Well, Sentry has your back. Use the coupon code tasty treat, all lowercase, all one word, and you'll get 2 months for free.

Okay. After that extended intro, Welcome, Fred. How's it going, my man? It's going beautifully.

I didn't realize I'm a 3 time I'm the 1st ever. That's never happened before.

You know what? Has Courtney been on 3 times? My wife may have been on 3 times, but I think this you are definitely the first,

to develop important than your wife. Incredible. On camera, on recording this cut.

I will have to confirm or I'll have to confirm that before I I set that in stone here.

But, yeah, the 1st time we talked to you was even before Astro was a thing. We talked about Pika and Snowpack and then Astro 2, and then here we are now.

And I do want to just I do want to to clear this up. Courtney has been on the show twice. Okay. So So confirmed.

I'm I'm the most important person in your life. Got it. Awesome.

That is, it is it's not confirmed in that regard, but thank you, Fred.

Not trying to get you into trouble.

No. No. No. She's actually scheduled to come on the show, and she was scheduled to come on before this episode. So that was a it was a Funny, little tight call here. Gotta get work on that stretcher.

Yeah. Right.

You will also have a a new baby on the way. Is that right? Yeah. Yeah. That's true. That's true. I,

our 2nd, which is a fairly scary thought. It is. Yeah.

How was number 2 for you, Wes? 1 to 2, we had them really close together, like Like, 14, 15 months apart, so it was a bit of a whirlwind initially. We we thought we were being smart and Just kinda grouping them together.

But, yeah, definitely, 1 to 2 was pretty tricky. 2 to 3 was was not as tough. Although, it all is it's all a A sleepy blur at this point.

Yeah. Yeah. Yeah. The hard part about 2 is I don't I don't know what your your distance between the first two fright was, but for us, it was like you were just Starting to feel like you're catching up on some of that sleep, and then you get hit with the the no sleep training. Yeah. It's like, oh, no. You don't know what not being Well, rest it actually is until you have. Yeah. Yeah. And it's definitely a lot harder to deal with, you know, challenges of 1 while the other one is, Has the needs of a baby. Yeah. We've avoided

the we've avoided the iPad for the most part, but I'm sure that's gonna be making a Grand entrance when this 2 baby comes home.

We avoided, like, the Ipads or or phones when going out to restaurants for for so long. And then at some point, you're just like, man, we gotta bust out the phone right now. Put on some Bluey at the restaurant so that we can get through this meal.

Yep. Yeah. That's that's a 100%

me. Fast forward a year too, for sure. Yeah. Yeah. Well, good luck. I I'm really excited for you. That's a special time. Right? So let's talk about another baby that is coming into this world, which is Astro 3 point o.

Do you wanna give A quick rundown about when Astro 3 is being released here to the audience depending on when they're listening to this, and then maybe we can dive into some of what makes Astro 3 a full major version release.

Yeah. For sure. Yeah. No idea when you, the listener, are listening, but for us, it is Currently launch week, so we're getting ready to launch a bunch of cool stuff this week. Wednesday, Astro three point o, that's August 30th. So by now, it's already out, but we just announced.

Yeah. It's a big new release with a couple of really interesting, major this is one of those ones where it's like

Some big stuff's going out that's not, like, huge, and they're like, oh my god. You have to relearn astro, but, like, it really kinda changes the game in what's possible. Maybe give us, like a a 1 minute rundown of what Astro is as well, because I know not everyone had listened to the last episode, and people are are tuning in right now being like, should I listen to this one? Good podcast hosting.

Yes.

That is, thank you. Yes. Astro is a web framework.

It's a way to build websites, similar to your Next JS's and your SALT kits and your remixes and all the good JavaScript frameworks, but what makes Astra unique is our focus on, content sites.

So working with a CMS or local markdown, whether it's a marketing site or a blog, basically, our whole focus up to this point has been Getting you, content and websites that are fast, really easy to build, without all the complexity of, like, a JavaScript app architecture.

It's much more HTML first. It's really lightweight, and it's, really fun to use. And you still get to bring your favorite UI framework, so you still get to bring React or Salt review. Beautiful. Yeah. And also one of the things I think that draws a lot of people initially to Astro is that,

you know, that's the hard part about any of these other Frameworks is that they're typically they're choosing the view layer for you or the UI framework for you. Okay. You wanna build a site Next. Js. You are using React. Right? And Astra doesn't. And I I think that's one of the things that really sets you apart is that you can write it in Svelte if you wanna write it in or or view or anything. Right? And that's just that's such a, when you when you all first announced that that was, like, Something that could happen in Astro. I remember thinking that is extremely ambitious and, going to be very hard to pull off, but here Here we are, and you you all have really, really nailed it. So,

yeah. It's been a journey. I mean, it's it's it's easier than it looks Harder than it looks in a lot of ways. Like, building a web framework in general is difficult. I I wouldn't wish that on anyone, but, it turns out that, like, this is kind of, what all those frameworks were built Or so, React felt view, they all started with, like, can I just put this on the page and, like, my old PHP server? That's what's gonna generate the page. This is just for this button or for this comment.

This is like in the DNA of everyone of these UI frameworks. So the idea of revisiting that, doing server rendering, but then using your components in more isolated kind of islands, it's it's very old school, and it's kind of

this has been supported for forever. We're just the 1st ones to kinda connect it to a modern dev experience. So okay. That's that's Astro, yeah, that's Astro now. Okay. So we got that out. Let's let's talk about some of these Astro 3 goodies.

We should circle back to that because it's actually something that I was really curious to hear your all's thoughts on about where that exact thing about, our architecture takes us. I'll I'll get to in a little bit, but I really wanted to hear your thoughts because it's a a fairly new, surprise for us of how you can connect Astra in an interesting way. So Send some steaks. I'm I'm putting some ground breadcrumbs out there. Yeah. So the big launches, for this, view transitions are this really cool new feature that's actually a browser feature. We don't get any of the credit. We're Just the 1st web framework to really lean into them.

The idea being that the browser can actually control your navigation from page to page, which lets you Morph elements from 1 page to another, fade them outside them in, persist like a video player across pages, and doing all that without having to ship a JavaScript app to To power your site. So it's still browser navigation. It's still server rendered HTML, but the browser has this new API which lets you hook into that Seamless transition versus just a hard cut to the next page you click. That's by far the biggest one in terms of what that means for how you can build and what you're actually able to build in Astro. That's so Cool. So, like, how does how how have you implemented that? So, like, anyone who's,

like, listening to this, if you transition as you got 1 HTML page You've got a sidebar, and you got another HTML page and that sidebar gets bigger.

The browser will know how to animate the sidebar from its smaller state to its larger state when you go from literally an HTML page to another HTML page, which is is wild for me.

So how have you implemented that and and sort of made it easy for people to use this new API? Yeah. It's it is a fascinating,

like, kind of superpower that just yeah. Every everything up to this point has been alright. You're going to a new page. Let's just, like, Shoot you out, like, slingshot you off the old page into the new one, but this is letting you kinda bridge it in a much more interesting way. Yeah. The The big thing is that the browser has kinda provided these APIs, and they did it in a really interesting way. So, they've actually kinda slowly been implementing this with a JavaScript API first With the goal of a full HTML CSS API second. So you can actually control this fully without JavaScript. Zero JavaScript page transition.

0 JavaScript page transitions is the goal, and where all the browsers are heading. That's still a little ways off.

But what they did thankfully was starting to ship a JavaScript API controlling this. So what we're able to do is basically give you an API that's gonna feel like what this all feels like in the platform API.

Gives you the ability to, like, name an element On on both pages so I can transition and kinda more for, you know, say, this should fade, this should slide. So we kind of build our API in In a way that kinda wraps that JavaScript API. So it's actually supported in way more browsers today than the 0 JavaScript version is because we leverage that lower level API even if what we're shipping is Trying to feel a lot more like HTML c and CSS.

And then we also do a good job with polyfilling on even older browsers too. So we actually kinda work all the way down the stack even if Your browser doesn't support

view transitions today. It still runs. Oh, really? I was gonna ask that because it it you know, Firefox and Safari support are still, like, out For view transitions. And and it's one of those things I I added to the the syntax beta site in Chrome just to Just because it's like it degrades gracefully, and it's a great API to get used to. But, yeah, you you said you are polyfilling, so Safari and Firefox Firefox will be using the same API, and then it's just going to just work for users? Yeah. It's kind of the beauty of the fact that the JavaScript API is there.

It means that you are still controlling this behind the scenes with a basically, like, a 3 kilobyte JavaScript layer, until that full HTML and CSS layer exists, but this is for now powering these transitions, pretty much across astro.

And It's cool, but, like, it also can be recreated in JavaScript. It's not as nice. You kinda wish you had the browser API, but it's essentially what these spas, these full JavaScript apps have been doing since, you know, 5 years ago where you have the whole thing in the browser available to kind of animate and transform.

So we bring a much, much lighter weight version of that to the, browsers that need it so that it's controlled by JavaScript happening on the page. It's essentially like turbo lengths if you've used that, but totally built into Astro and and designed for Astro, if the browser doesn't support these new APIs. That's nice.

In fact, like, I you know, it's one of those things I think people have been using, like, native in native app development. Right? Yeah. Those things are just built in, or they're really easy to do, like a card transition or a a swipe up or anything. Any type of, like, page route In a a mobile context, can be easily animated using, you know, on hardware support. And, you know, here we are on the web Getting the full page alone between things like we're 2nd class citizens here for the for, serving data. So yeah. No. It's, I think it's really awesome that you guys are are taking the approach where it's like, we're we're not only gonna focus on the API, the browser API, but we're also going to make it easy

for anybody to access that. I I think that this is really cool that you are focusing in on this, and I didn't even think that this would be, like, a next Point what frameworks will start to be implementing, but it makes a lot of sense because in JavaScript, we got the history and push state APIs, which is How you can change the URL bar without actually reloading the page. Before that, we had to literally change the hash in your URL bar because There's nothing else you could change in the URL bar without, actually triggering a full page reload. So when we got pushed in the history API, That's when we started to get like, we had it for a while, but it it didn't catch on until frameworks started to implement it. And then it's as simple as writing a link tag Or even just an anchor link in some frameworks, and you click on it, and it will just do the magic for you. And I think that this is the next do the magic for me thing, which is I'm gonna just render out my HTML.

You figure out how to make it swoopy between the 2.

Yeah. A 100%.

It's always hard to know with specs, like, how much there's a narrative or a a longer term vision, but you can kind of tell this is all part of the story of making the web Easier to build these things natively without frameworks having to do so much heavy lifting. If we as frameworks can tap into the net the native browser API, It just means a lot less work, a lot more consistency, and it lets frameworks then move on to tackling the next problem. So there always feels like this is kind of, like, Wait for the platform to catch up. You go out ahead and explore, and that's like the 1st proving ground on a lot of these ideas. If they're good ideas, they'll make it back into the spec. If they're not, then that framework just did that weird thing, and we're all just gonna pretend it didn't happen, and move on. So what else you got? Oh, yeah. I mean, that's That's the headline. There's a ton of just behind the scenes improvements that have been happening. Image optimization built in Astro is something we've been working towards for, like, 2 years. It Turns out images are hard to work with, but, this API we finally settled on is really, really nice and solves a lot of the,

kinda leap frogs over a lot early problems that we've seen other frameworks struggle with. Like, how how does it work? Let let's let's dive into that. Because if somebody with a Gatsby website that prerenders 40,000 images or something silly like that. I'm I'm, I'm I cry when I need to clear my cache, and, obviously, that's been fixed.

I just haven't updated my site yet, but, like, that was the wrong way to go.

Oh, god. It's, I mean, doing anything with images at scale gets really tricky. I I think our the way we walk the line is trying to do enough that for small sites, it's essentially invisible. For large sites, you do wanna think through that kind of stuff Or use a hosting platform that supports, built in image optimization, and you can just skip it all during build. So even static sites, and serverless sites on Vercel, for example, can tap into their build API and, get that all basically handled by Vercel, not as a part of your build, not as a part of your, your runtime.

It's it works a lot like other frameworks. We didn't really wanna reinvent the wheel. You import your images if they're local, using like an SMM plus like anything else, any other asset in your, code base, and then you place it in a image component, a special image component that we provide to you. And we do the work to then make sure that that gets optimized based on the height and width of that component.

So we try to smooth out all these edges, you know, image priority, things that are just kind of not you know, all text is required. We're trying to mainly,

focus on where we can help and, and building it for you where That makes sense. Yeah. And people are probably a lot of people stumble upon images and they they say, like, why can't I just use image tag and put the source to d c I m underscore 5,002 dot JPEG. I took it right off my digital camera, and it's 19 megs.

Why can't we just have that as an API? Oh, just using the image I mean, this uses the image tag. It's it's really, like, Similar to how we're using view transitions and kinda letting the browser do all the work, we're not doing a lot of work on top of the image tag. What we're trying to really like, 90% of our complexity is Taking the source image that you just dropped into your repo and making sure that that's gonna render as optimally as possible. So it's really about the asset

And less and less about the actual markup. And, also, you you have issues with like, not you, but I I've known in the past that there's been issues with, like, Do you want it to be fixed within height image, or do you want it to be fluid? Do you still have to specify or do you need to do that at all in Astro, specify

Which type you'd like it to be? Yeah. That was a big so that's a big change if you use the old image integration. So we built it separate from Astro at first, experimented Now we've merged it into Astro, so it's in core in v three.

That early integration though, yeah, that was much more we kinda tried to do everything. We're kinda curious what people wanted out of images. We did Cropping. We did aspect ratio fixing. We did background fill if you've cropped it in some way. Like, there was a lot going on there.

And it turns out that made it a kind of difficult. Users didn't really expect a lot of that cropping. Like, I give you an image why is it different on the page. I wanna do CSS to kind of restyle this, but now it's actually cropped based on the defaults.

That's not super responsive.

And also just like, I wanted to connect this to Cloudinary or ImageX or Vercel, and, like, they don't support all this crazy stuff. Like, why did you why did you do this? You've made it very difficult. So, we tried to have some pretty basic support without going too far. And, again, that's kind of the nice thing is if you really wanna connect this to some other Provider, you still can. So we're we're basically trying to be lowest common denominator.

If you wanna go above and beyond with some provider that supports more or you wanna optimize your images Beforehand in a custom flow, that's all still supported.

We're just trying to do the good default. Beautiful. Nice. Let let's keep going here on features. Any anything else we add?

There's some really nice speed up improvements.

The HTML is now all minified and optimized. The rendering performance has been sped up. There's a really interesting thing about HMR then. This is kinda getting into the thing I wanted to get your thoughts on, about what you can use Astro for.

As, Wes, do you have an AstroSight? I know you've used it before.

I don't have an AstroSight, but, yes, I certainly have split it up a couple times. One of the things we started with with Astro was just that content focus. And even before that, like, static content, there's no serverless, no server anything.

And it was just that was a a really nice message for us. It helped us stay focused. I think a lot of our early adoption, early success was really based on our our focus there.

And a big thing that we've seen people now take Astrone kind of take away from us a bit is, well, I actually wanna have, A, like, web app. I wanna actually do something more complicated.

You mentioned, like, the frameworks can all kind of exist, happily within Astro. You can use react, you can use view.

The weird trick that we've seen people start to do as well, an island is just a piece of UI on the page, so why don't I put an entire application on the page? It's 1 big island, and essentially treat Astro like a create react app or a a Veed almost.

So I have my marketing site that's super fast. I have my, like, SaaS app that's behind the login page, and that's just like a Classic SPA, and my APIs are powering either both or 1 or all or none.

All in the same code base Kind of mixing and matching based on your use case. That's something that we don't really talk about a lot in this, release post, But we've been making a lot of improvements to the react flow, and now we're starting to see that more and more. So it's this weird, like, feature and then also like a New use case that's unlocked by the feature that we're still feeling out, but it's really weird and really interesting to see people starting to use it for this. Nice.

What so okay. So it's it's launch launch week for you guys. And the Monday announcement, and Guillermo was kind of teasing this a little bit, but the Monday announcement was a partnership with Vercel as a hosting partnership.

Do you wanna go into some details about, like, how that happened and and maybe what are the benefits now of having a hosting partner for Astro? Yeah. Definitely. Yeah. That was our our Monday release,

which yeah. The result team kinda teased, very sneakily in a way that I think everyone was, like, oh my god, for sale is Astro.

AstroX.

Yeah. That's, like, certainly what I thought. Yeah. Yeah. I Wes, did you see that tweet?

No. No. I I surprisingly didn't. I I I saw that there was something coming on Friday. I got so many DMs of, like, oh my god. Is this real? Like, yeah, it was I, Guillermo tweeted out something like, we have some astronomically good news, and everybody's like, oh, okay. Oh. It's like Wow. It's a I saw that. I didn't catch That at all. That's funny. Did it? I saw that. I was like, that's not a very good teaser. That's, like Provester Versacell. They Versacell really has the, like, The hype on on dial with that type of stuff. People get really excited anytime. It was such a specific, like, 7:30 p it was like it Sound like a press release was going out. It really, kudos to them. They do a great job at that. They're trying to bury something on Friday night. You know? Like, what's going on here?

Incredible.

Yeah. No. They've been awesome partners over the last couple of months. We actually have had, so Netlify was our previous, hosting sponsor. Sir? Mhmm. But over the last couple of months, we've been really working closely with Vercel, especially with their build output API, has started to unlock a couple of really interesting features that we just haven't been able to see anywhere So I mentioned that image built in optimization, code splitting of your routes. A lot of these are focused if you're if you're serving serverless like an actual server. Static sites are are Fairly easy no matter where you host.

But in addition to the performance, there's also just this really cool features that they unlock to make your site faster.

Yeah. Code splitting image optimization.

Oh, edge middleware, of course.

Lets you actually run your middleware, which is an astro concept we already have. They give us the API, and we kind of Work with that API to deploy it to the edge closer to your users if that's what you want.

So there's some really cool features that really I've only ever seen on Vercel. They do a really good job of connecting it all. So when we started talking about this with them, it just kinda felt like a no brainer.

Yeah. Been really, really nice working with them and collaborating on them, and They're basically letting helping letting them help us, test all these new APIs.

Kinda dogfooding them for us has been really helpful. So when a deal come like that comes around, is that like a, do they come to you and say, hey. We really like what you're doing with Astra. We would love to Work with these types of features to improve the platform, or or, like, how does that even come about? Yeah. We're fairly unique in that we take our open source,

model fairly seriously.

You know, I think there's a couple ways to do open source when, so the background very quickly is that Astro was formed, and we, early last year, created a company to basically help support the long term maintenance and sustainability of it. So Raise money. A couple of the core community, members got together, and that's what helps us work on it full time.

Really hard to do that That level of investment, with just open source sponsorship alone. We after you came on last time, like, I'll I'll be straight up because this is a straight up show.

We had a we had a couple of people wait. Is that our new, like, our new angle, Scott? Yes. Straight up. Joe. We're we're out of the skies.

I don't know if anyone can see there's a light being shown in my face right now. It's very dark all of a sudden.

Straight up with Wes and Scott. But, A couple of people were like, I don't know how, like, Astro can survive. Right. Like, there was some talk of maybe themes. Like, we specifically asked you last time. We Like, how does Astro make money? And it's it's such a tricky thing when, like, you build an open source project that isn't directly tied. And some of the options are, you can sell themes, you can get bought by Shopify, or you can like like Yours, you have a partnership with Vercel. Right? Exactly. Or you have a hosting company. That's that's a very popular one as well. So it's interesting to see, and I'm I'm glad to see it, that, somebody who does make money like Vercel can then take open source projects under their wing. And I'm assuming that's how it works. Right? Like, they pay you money, and you say use Vercel? Yeah. So there so oh god. There's so much here. I I I think something that's really interesting about this moment too is what's going on with Terraform

and Gatsby getting acquired by Netlify, and I I tweeted about this Realizing Mhmm. They've kind of laxed on their commits to that repo. It seems a little dead. I think they picked it back up since, but I think there's a lot of, like, Maybe not fear, but, like, I think anxiety about, like, what is sustainable open source? Does it have to be a company? Can it exist? And we're, like everyone else, just trying to figure that out.

Yeah. I I think so. Vercel is sponsoring the project. They're fantastic. It is definitely not an amount that lets us build, you know, a 5 10 person team around this project.

The nice thing about our open source sponsorship, it's all public, so you can go and check this out. It's $5,000 a month that they support. I think our total sponsorship is now run up to a 150 a year a 150,000 a year.

Maybe I'm over exaggerating that. It's something that, like, means we now could and can from open source funding sustain this project, and that's a huge milestone. I think that's something that kinda gets Lost in the conversation is that we are choosing to build more with this company for astro users, but the framework itself is fairly self sustaining. These sponsorships are a huge part of making it So that we can go build stuff for Astro developers and try to sustain even more, resources with the company. But for the most part, this project is self sustaining.

Nice. And I'm going to say, you know, you don't get a massive wall of contributing sponsors like you do, for just any old project, if you if you go to the Astrodox and you look at the, contributions page, you can just see How many people, like, truly believe

in this work because of how many individual sponsors there are. It's it's pretty incredible. Oh, and we can go super deep on It's like the whole idea of how we run our community is that it is an open covenant. So anyone can come in, contribute, become a part of the community, and eventually become a core maintainer, which is Literally, like, higher than like, there's no extra rights or privileges in our community you get by being hired by the company versus, coming through. You're actually a lower class, like, Governing citizen in this world, if you come through the company versus the community. So it's it's a really interesting model. I don't know if I've seen anyone do it. We're we're still figuring it out, but it's been fairly I I've seen a ton of reward just in seeing you feel that I think if you go through our Discord, if you go through our Twitter, like, it is such a community vibe, which I think really fits well with a framework which is Touching so many people's work, you know, it's the directing you interface. But I think that's why we're able to collect sponsorship where an MPM package might not be able to our JavaScript library. Like, It's a very personal experience working with a framework, and I think, users, understand that. We definitely understand that, and I think it's something we, Yeah. Don't think lightly. How many how many people are currently working on Astro on, like, a day to day basis? So I have to check. I think our core maintainer team is about Fourteen people now, and that's a mix of of company and just, like, community. So, yeah, it seem maybe, like, half of that is, company. I I'd have to go back and check, to be honest. It all again, it's the thing. Like, it all kinda blurs into just core,

open source. But, yeah, we're super lucky to have the contributors we have. Yeah. Was gonna ask, like, how do you ship so much? But I I feel like that's probably you know, community first is probably one of the reasons how you do ship so much. Has it, like, been difficult, to grow from, you know, being a smaller

org to an org where people are, You know, working on a bunch of different things, or has that process been smooth? Has it been, like, a big learning experience for you? For me, personally, definitely. I think for the team, it's felt fairly organic. Just, Yeah. Like, it's like any open source project or software project I've ever worked on. You kinda constantly wanna do more.

I think that kind of scope grows naturally as the people can kinda come in to fill it. So saying no is still something you have to do, but, like, one of the things we're actually getting ready to announce, so our time tomorrow, this will be Tuesday of the launch week, is actually announcing launching our docs template.

It's the thing that we use for our own docs called Starlight.

We built it for that reason, and now Sharing it on Product Hunt, getting out there, like, that's our docs team has just been, like, crushing that, and that's something that we couldn't have done, back when we were, you know, 3 people, 4 people, 5 people.

So I think you just naturally kind of fill the space with with stuff to do. But, yeah, it's been a while since I've coded anything myself on the on the project. I'm fairly now Helping everyone stay, you know, focused, prioritized. I'm I'm the steward of the project is now the kind of open governance rolling. That's a a full time job for sure. I was gonna it's it's so funny. I was on your docs, and I was gonna ask you, like, what what is the docs built in? And, you know, I was gonna ask you some interesting questions about that. So,

It's very nice of you to bring that up for me. The the docs

being part of your launch thing. What what is this built in? I mean, you said it's built in the asteroid's a template, but, like, what what UI framework is this Yeah. I actually realized I just I misspoke. It is actually a custom framework. The relationship goes the other way. We basically, after 2 years of working on the site, realized, like, there's a lot of good. We actually had a docs theme. It was one of, like, the first 5 we launched. Like, when you use Astra, check out the docs theme. And what happened was that that Stagnated, and our doc site just went, like, internationalization, dark mode, side barn, site nav, typography, code block search. Yeah. All these things getting built on.

And now what Starlight is this project is basically trying to take the best of that and bring it into the community. So instead of having Crappy old open, theme and our beautiful dockside. Trying to flip that where there's now beautiful theme, and then we're just in the play implementation of that.

We actually haven't jumped it, so it's still the custom build. But essentially, everything in Starlight is inspired and and for the most part, an improvement on our current doc site.

Yeah. Custom custom spot and custom built. Yeah. Okay. So since we are listening or, you know, the viewers will be listening to this in the future after your launch week here, Let's let's go step by step. So we got Wednesday is gonna be the 3 point o launch. What what's going on Thursday, or what has happened already on Thursday?

If you're if you're listening to this, what has happened? Thursday is our community day. So there's a couple of, small All our announcements are really cool blog post about how we do internationalization, a swag drop.

Just trying to kind of give a day to celebrate, again, everything we just talked about, that community that supports us. So a couple of, kinda swaggy kinda swaggy dropping. Oh, I wonder if I can get you a hat in time. A hat. We're dropping some hats. Oh.

Oh, yeah. I would love that. I'm I'm I'm a hat guy. I'm a I can tell you. We just we started this podcast talking about it. Love would love an Astro hat. So, Thursday community day.

Yeah. Nice. And what about Friday? Friday is still a secret. I actually, I'm not even gonna tell. I'll tell you guys off air, but, I got even for our listener. It's it's our kind of one more thing that's secretly a really big deal.

But yeah. Okay. More more on that. I have a reason to go back and check out our our Twitter. We've been watching all these things? I think we made the mistake. I think we should have Courtney on,

for her 3rd episode before, we had Friday, if we're gonna be delegated here. It it it feels like really great. Is is Launch Week you something you do every year? Is that like a annual initiative? Yeah. It's actually every 3 months. So we, if there's any part of, like,

company thinking that's, that's found its way into our open source community, it's actually just like the focus of, having a release date and sticking to it, it's all fairly, like, you know, that's the nice thing about a launch week. And, actually, I'll give a shout at super bases, the first people I've seen really Talk about this and blog about it.

The idea is every 3 months, we come together. What do we wanna do? You come up with, like, the best possible launch week you can imagine. Like, what are the 5 coolest things we could do? You then go that 3 month period, you go and work through that, and then you would celebrate it and announce it.

The things we love about it is it's not like, alright, like, Here's how we need to break it up exactly. Here's the super complicated, like, everything has to go out. Like, it's really lightweight. It's basically just everyone at that moment deciding what they wanna do and then committing to it.

But you can announce it 1 week into that, 3 months, like, whatever it is, It can kinda go out. That's the interesting thing about how communities work. Like, the image optimization people, work has been visible for the last. Actually, funny enough, the Vercel announcement blog post was sitting in our repo open and accessible for, like, 3 weeks while we work through the migration.

And then Realizing, oh, wait. This is a launching. We should it was actually when Guillermo tweeted it. I was like, oh, no. People are gonna find this. Like, delete the PR, Clear out the branch. But, like, it's a weird thing about how communities work is you can kinda do stuff in public without people really carrying, and launch week is the celebration of the work. The, the announcement part is is kind of a a side effect. Nice. Yeah. Oh, that's great. Although when we did our century announcement,

we we usually have, I think between, like, 5 10 people sneak ahead episodes, meaning that we'll put the episode in the repo With a link to the m p three, but, obviously, it doesn't go live until the the release date. But we have a couple people who who go literally download the m p three and listen to it early.

So when we're doing the swag drop and the the century announcement that we joined century, we had to hold it back.

And everybody's just like, where where's the Where's them going on? You guys are trying to Yeah. You guys alright. Something's up. Yeah. Something's going on.

This this is our 1st time having some sleuths. So Yeah. I am curious about your thoughts on

seemingly that a lot of open source needs to be, like, good at marketing.

Recently, like a lot of open source projects have have swag. They have really good branding. They have a design team. They obviously have launch weeks and all that type of stuff. And, we we felt this maybe 5, 10 years ago where people thought, like, can I just be a developer? Do I have to tweet and blog and Go to conferences and do all this hula to have to be a developer. I don't want to do that. I want to write code.

And and now it seems like like open source is somewhat heading the same way. Do you have any thoughts on that? Yeah. I mean, I think there's a couple things there. The first thing I'd say is I don't think this is new. I think this has always been true that the people with the popular projects were good at marketing. I think Before Twitter, it was probably conferences. Before conferences, probably email list. Like, I don't think it's an accident that some projects, have lots of users and some don't have, many at all. And, also, I don't think that's always the goal. Like, I love just building software, and it doesn't have to go out and get a lot of users. In fact, after running after, like, I'd rather not have any users on any of the things I have built, like, building for yourself is still a super valid, wait. So I don't know. I think sometimes maybe the judgment gets a little messy of, like, oh, if I don't have users, if I don't have a Twitter account or or swag, is is my project still good? I I think it's it's it always depends on what your goals are. For us with Astra, we wanted the most people to be using it, so everything has kinda come from that. We really wanted to actually Change the way people saw web development. That was very much an explicit goal, and everything came from that.

By no means do you need to be a great marketer to work on something open source, and Plenty of projects have millions of users Yeah. And you've never even heard of them. It's a it's a certain type of project, I think, that benefits from it, and the others, it's probably a waste of time. It's true. Does curl even have a logo? Oh, it does. Okay. Exactly. It's got a logo. Wow.

HTML still doesn't really have a logo, does it? I mean, it's got that the or CSS doesn't have a logo. CSS had, like, the CSS 3 had the 3 inside of a badge. Let's see if it fills up. Are kinda tired now. Even h t h the HTML five logo is kinda tired now. We're Super tired. Yeah. Later. We get it. It's fine. We were doing our redesign for level up tutorials. We had all the different badges of all the different logos for the tech. And then when we hit CSS, we were just like, what do we even what do we even do for that? Do we make our own? We, like, invent something? Like, it was such a a weird thing to look at some of the older ones. We just made our own, but it was, like, Very frustrating because it was like, I guess, we're definitely not using the CSS three one. Speaking of threes here, a nice little transition with Astro three. You guys are doing a big version number change, like a major version release. Like, I I wanted to ask you about, like, your thoughts on Releasing a major version considering, you know, last time we talked to you is version 2, version 3. It's it's not like it was no time, but it you know, it's pretty quick. Do you do you get concerned that that major version will, cause some people to be concerned about the upgrades and maybe like, what do what are you guys doing to do major version upgrades To make them less, scary?

Oh, yeah. That's a great call. I mean, it's it's kinda There's no perfect answer there. Right? Like, either you do 1 once every 5 years and it's a huge ordeal that, like, breaks your community, or you do them very quickly and then it's Lots of little changes, which is still super frustrating. Like, quite every month, I have to come back in here and make some breaking changes and be really nervous about this PR getting merged and going out to users.

We try and try and kinda walk the line between both. So I'd say every 6 months is seems to be the cadence. That's not anything other than just, like, that seems to be what's happening.

Again, we go into one of those launch week planning. It's like, you know what? It feels like we're kinda getting close to that experimental thing you worked on last week.

We We'd wanna keep testing it, but it's probably gonna be ready to go out. It's gonna be this breaking change. Okay. That's kinda how it seems to be going for us.

And yeah. And then I think really good docs and really good tooling, whatever you can do to kinda help people as the other side of that coin. So, 6 months means nothing's really huge. We haven't made giant breaking changes.

A lot of it's just, I'd say it's kinda more the annoying if anything, like, this API changed or this, import changed, but it's not like our router changed or anything that drastic. So we have a really great upgrade guide. I'd recommend anyone with an ASTRO site looking at ASTRO 3 to definitely check that out. Our docs team did an incredible job. It's Item by item, here's what had changed. Here's I can't remember if it says why, but, like, here's if it affects you or not, and if it does affect you, here's what to do. And you can just kinda breeze through it, try Trying to keep it as informative as possible. No guesswork. No digging through change logs to find out what broke. It's it's fairly,

thorough. Have you thought about, like, code mods? Is that a a thing that's on, anybody's

radar for upgrade paths? Yeah. I we were just talking about that in Discord. I would love To see. And I don't even think I think it has to be Astra specific, like, the idea of our upgrade guide by going step by step could be a tool where it's just like, alright. The next thing I'm gonna check for is this thing. Good luck. It doesn't affect you. On to the next one or, like, oh, it does affect you. We found 3 places to go change. We're gonna wait here. Like, I actually would love. It's like a personal we have a hackathon Coming up that we might be running, and I might go and build that.

Won't get in for this launch, but, like, there's so many cool ideas here, and I don't think anyone,

I haven't seen anyone do it really right yet. Yeah. You know what? One of the things I think, I I don't know if you were, Into that the world pre one version 1 of SvelteKit, they had made a major change to, data loading in SvelteKit, And they did a code mod for that. Yeah. And it was probably one of the better code mods I've ever used because if it couldn't fix Something it would leave a comment in telling you, why it couldn't fix it or whatever. But any anytime it would you know, the the comments that it left And the work that it did were both very clear, and it worked. So if you ran the code mod and your whole site wasn't just working immediately,

You knew exactly where to go and what to do. It wasn't just like, you're you're screwed. Yeah. I I remember that was a huge change at the time and, like, fairly close to their one of the whole. World. But they did that. I like, I never hear anyone talk about that anymore. I think they did a really good job with that. Yeah. Clearly, in retrospect.

That's always the worst thing now that 3.0 is 3 that it was 2 days away, and we're just like, oh, what about this breaking chance? Like, no. No. No. If it's existed this long, it can wait till the next one. Yeah.

Are you still using Vite as your tooling under the hood, or was that what you were using? Yeah. We are powered by Vite.

I think they actually have a new major version coming up soon as well, so we're gonna try and coordinate with them on that. Yeah. Beats great. If anyone doesn't know it, it's like a Builder bundler similar like webpack or or roll up, and it's what, like, all the new all the cool new build tools are built on it. So Astro, SvelteKit, Nuxt, I think.

Remix and Next. Js are the only 2 I'm aware of that aren't still haven't made the switch yet. Yeah. I I really personally like working in Vite. And, Wes, you just said Vite 5, which is is wild to think because when you all said that there's a new major version of Vite, I was thinking, oh, Vite 4. And I even googled Vite 4, and it was like Vite 4 blog post. 4 is out. And I was like, what? When did that happen? I man, it's a it's a while. I work in feed every day, and I didn't know it was really worth They do such a good job at

Making it not

A pain?

Yeah. It's not a pain. Like, I had I had a whole course on Vite. Like, Like, not, of course, on Vee, but, like, I had the whole thing built with Vee. I had a couple plug ins, and I literally just updated Vee, and that was it. Like, no no breaking changes. There was. I don't know if there was any breaking changes at all or they were, like, flagging things, but it was

it just just worked. And then all of a sudden, you had a couple new features, it was a bit faster. Yeah. Veed, in my experience, is one of those things where, like, it's a zero effort change usually because what it takes your code and it bundles it. Like, there's not really an API to Veed.

But the way it bundles things changes, and, like, sometimes you gotta be it's one of those things where it's, like, 0 effort to upgrade, but then, like, you really wanna go through the page and make sure nothing broke, in my experience, it's I'll take that over, like, you know, a huge break and change where you have to, like, spend a week refactoring your code base, but,

definitely not, not painless either. Yeah. Yeah. Because it's always a pain if you do a big refactor, then you gotta, you know, push it. You you you're it's like it feels like much more of a commitment than, oh, just roll it back if It doesn't work type of deal. Yeah. Yeah. Snowpack always had that trouble as well. How do you feel about working on a a framework Like, Astro compared to working on a tool like Snowpack.

Does it feel like the work is, More fun, I guess, so to say.

Fun fun is definitely a word. It's definitely, I'd say more rewarding maybe. Yeah. I I think it is more fun. The, I mean, snowpack, we wouldn't have stopped working on snowpack if we weren't confident that Vete was, like, clearly got taken off. Like, it was only because that existed. We said, hold on. We can actually, go pursue this other idea that we had that we were, like, really excited about, but, like, snowpack needed to exist to power it. Like, I don't think there was ever, like, a huge master plan there, but I do think that that was, saying that on Astro only exists because we came along, I'd say.

So, yeah, it's it's just different parts of the stack. Like, we love being able to build on top of it, couldn't do without them. But we also then end up kinda being this, like, layer on top.

Different set of challenges and different set of, user expectations.

Oh, I have a question. Last time we talked, a tree fell on your freaking house. That's What happened to the whole time? No wonder oh, god. That sucks.

I don't wanna talk about it.

Oh, no. No. No. We're as you can tell, we're happily in a in a new spot, but, man, it's been a tough year. Man, a tree. We just had a crazy storm last night, Wes, and I was driving through the neighborhood. And I saw, like, 3 or 4 trees down, and I was just, oh, shoot. I hope nobody's,

nobody's house got cranked here because that, Man. We need to go back and pick up some mail, and it's like the house is still just like they basically, like, tore it down to its, like, studs, and that's it. And they're like, how to re because It was during the middle of a rainstorm, so then it just tree fell on house and then rained for 2 more weeks. Like, the whole Oh. I think they just I don't know. I don't know. It's a the the nice part about renting for all of its flaws is we just to get out of there, but still

Yeah. Yeah. Scary.

What what are your thoughts on, like, text editors in the cloud first locally? I know that's, like, kind of a big push now. A lot of different tools are coming out in cloud text editor. Do you think that's something that we're finally gonna start To see being a a big push or do you think it's a little ways off still? Oh, that's a good question. I, I don't know.

I think it's all I think enterprises, yes. I think enterprises want it really bad. It's like less to configure, like I don't know if they just, like, Specifically want, like, all local in the browser versus, like, a server that's running somewhere in the enterprise cloud.

But I feel like that's something that I've seen enterprise talk about where it's like, You know what? We have a 1,000 developers, and we don't wanna have to manage everyone's machines. Like, let's just automate this and and wash our hands a bit.

For us on open source, it's really nice having these hosted tools that let us repro and share code really well. I think for, like, code sharing, it's, like, just a Huge superpower for, like, the world. Like, so much more efficient to be able to, like, here's some actual code. You can open this link and reproduce it.

I don't know what the Businesses there, but it's, like, definitely something we use every day in our own repo. Yeah. Those tools are really cool. That's they they're all popping up lately. We're having

Someone on from Google next week Nice. Talk about, Google's IDX.

Built marketing site built with Astro.

Is it really?

Yeah. Oh, there you go. Wow. It tells us so much better than anything we could build. It's, like, kind of painful to see. Like, damn it. This is That's a beautiful site.

That's a really nice so it's idx.dev if anyone wants to to go visit it.

That's beautiful. So what's the it's all built in Astro. Did they what do they use for their

Templates. Oh, I think the product itself is maybe Angular, I wanna say.

Something googly. But, Oh, yeah. Not sorry. Not not the actual product. The the product itself is, like, Versus Code just hosted, but, like, their marketing side, I mean, is it is it

Using rack components. What you're saying? You know what? I didn't even think to look. I just saw it after. I got excited, and I closed the browser and did a dance. Let me know. I know what you can say. Celebrated.

Yeah. I'm looking here. Here's my trusty,

Wappalyzer.

Svelte, it looks like. I see Svelte showing up on the screen.

Yeah. Spell tailwind and Astro. Wait. There's there's gotta be an acronym for that.

Smelt, tailwind, and Astro. Sat. The fat stack.

You can yeah. That's yeah. A a s s is definitely an acronym you could too difficult to, put together. Honestly, I really think in in 2023, we are ready for the Aztec.

I think everybody is ready for the Aztec. And, honestly, that That used to be talk about open source marketing. That used to be it. Yeah. All you needed was a That's all you needed. Acronym for your stack.

You know, like, we had lamp stack, and then we had the MERN stack, and we had, what else?

None of those are as catchy as the ass stack. None of them. What what what's the Aztec? Astro, Svelte, and

Oh, serverless styles. Yeah. Serverless.

Sir, that's right. Sure. We can come up with something. We'll we'll make it work. Right? What where do you go to keep up to date With stuff because it you know, especially, like, browser APIs, it seems like that's something you gotta be, like, really up on, running a project like this. Like so for you, How how are you staying up with these browser APIs, this modern stuff? I'm yeah. I I find it really difficult with specs

Specifically, I think unless you're, like, fully living in that world, it's really hard to stay on top of, oh, this repo's out of date. We actually moved over to this, like, it's This mailing list post links to this repo. It's it's a big world out there that I do not know how to navigate.

Twitter, a lot of it comes through. So Spent yeah. What? Like, a decade now on Twitter. I feel like I've honed the algorithm enough to get me good information.

And I do keep an eye on there. There's some mailing lists like JavaScript weekly and, invites.

I check out both of those pretty regularly.

A few others that I a little less regular on. But you just kinda try and, like, you miss stuff too. It's, I'm definitely not perfect. There's something about also information kind of competing for your attention where, like, if it's something really important, it does kinda have a way. Like, the view transit yeah. The view transitions API is a really good example of that. Like, it came to Through people experimenting and showing us, like, it just happened that they were using Astro for it. So, oh, like, this API exists.

It's awesome, and Astro users want it. Like, it's come to us almost like with a bow on it. It feels like like you have to be open to seeing that, but then once it comes by your kind of your eyesight, I don't know. Yeah. Yeah. I think the good idea is make it through. I figured it out. I asked Chad GPT. It's the vast spat.

The Vercel, Oh, Astro, Svelte, and Tailwind stack. Nice. Ship it. I love it. Alright. Who's gonna tweet that 1st? Who gets the points?

Yeah. Oh, that's great.

The Astro code base, do you write that in In TypeScript? And if so, what do you think about the Svelte move

to JS doc? Oh, right. Yeah. They moved their internal code base Back to JavaScript, but using JS dot comments to give you typing anyway? Exactly. Yeah.

Yeah. That's Good for them. That's super cool. I I think we like TypeScript too much, but I totally respect why they do. I think Rich has always been, fairly pro JavaScript and Maybe a bit more skeptical of TypeScript, and it comes with its costs, I think, certainly.

I and I always saw that as more like That's a philosophical thing that Rich believes, not, like, necessarily, you know, that the whole team got together and voted, and this is what we want. So we haven't seen I haven't seen any real need for it, and I think we're pretty happy with TypeScript. So, yeah, we're probably sticking around. Oh, that's good to hear. Okay. One of the major needs that they pointed out of why they did it is that if you are

working in Svelte, And you're writing in TypeScript, and then the the code base itself is written in TypeScript. There's, like, this, like, double TypeScript Transpilation that needs to happen and trying to debug it is slow and confusing. So if you're If you're dev ing on it and there's no build step there, then it's it takes that out of it. So I I've never built a framework like That's why I don't know. Yeah. That's a good point. Like, Svelte runs on the front end

and is, like, way like, everything we do is almost, like, part of your build. So if it breaks, The build breaks, it's not like if it breaks your code base breaks. Yeah. It's kind of maybe the the difference there. So, yeah, there's like this kind of golden, like, Especially with the SM now working in the browser. If you can get your source code as close as possible with its chip, absolutely. Like, yeah, debugging, Bundled code or TypeScript code is just such a pain. Like, yeah, there's definitely something there, to just bring that as close to source as possible. Yeah. It'll be really interesting once this,

types of comments proposal continues to see some updates. I'm I'm seeing this. It looks like there was And update it in March 2023, so I might wanna dive into this thing to see. This is the, like, make JavaScript add types to JavaScript keywords? Yeah.

Yeah. Yeah. I feel like that will change the conversation considerably.

Right? One thing that I'd never wanna go back to is worrying about the code I write not running in certain browsers, so I will say that's, like, a really nice thing that TypeScript does for us. I'm I'm sure Babble and other things do, but, like, Compile it all down to whatever the target is you set means you can write modern code.

Your users run the code that they need.

Yeah. That's something I never wanna go back to. It's like, oh, wait. Does this feature actually run? Let me check out open up my browser support matrix, and, oh, they don't have that. So I have to

Go search Stack Overflow for the polystyrene. Oh my god. Never again. Never again. Yeah. That was brutal. Never again. Yeah. I know. If I have to write another browser prefix for CSS. That'll put me in a happy place in life.

Okay. Well, let's talk about sick picks and shameless plugs. Did you come with a sick pick For us today, anything that you think is sick or cool? I always I think I made this mistake last time. It's not a picture. It's a sick pic. It is not a picture. It could be a picture. Yeah. It could be if you have a sick thing. Wait. This is gonna work out of audio.

What am I doing? I was literally opening up my phone. Like, I wonder if I do have any Oh, an audio porta pot. Yeah.

Alright. I got a yeah. I got a picture of a, a crab that's holding a cigarette. That's pretty cool.

As a part of the launch I've been super busy, so I don't have any time for, like, video games or or movies or fun. But one thing that I just got very excited about was, Factorio just announced a new, expansion For space age travel, so once you get off the planet, where do you go next? Factorio is this awesome game. It's it's very much like programming the video game where you, like, Build a factory and, you know, you throw it together because you just need 1 or 2 parts, but then once you need, like, a 1000000 parts, you start to automate it with Conveyor belts, and it's basically like refactoring different levels of abstraction as you go. It's so much fun. It sounds so lame. It's so much fun.

No.

No. I I played to death a game on the iPhone that was like it was like builder or something that was In fact, it was an iPhone factory, a clone. And I was the whole time, I was thinking, I'm just refactoring code right now. That's what I got. But yeah. Exactly. It's, really addicting. I probably sunk, yeah,

more hours into that than I care to admit. But, yeah, very excited about that expansion. There It's based off of this one that I think takes 600 hours to compete. There was complete, there was a mod, and they've kind of, I think, stripped it down to a 100 hours to complete, which is still insane, but,

I at least we'll start it. I don't know if I'll finish it. Yeah. I'm looking at the blog post announcement for it, and there's a really amazing animation,

that I could just look at for hours. This thing is so cool. That is that game is also, like, the open source of video games where, like, they, I think, have just been working on that for, like, what, maybe 7 years now, And just consistent updates. You buy it once and that's it, and, like, there's this whole community energy behind it of mods and people. Like, I think that post was at the top of Hacker News. Like, the The love for that game is so real. It's cool. Yeah. I I,

I've avoided really diving into it because I know how much I played the 1 on my iPhone that I will. Yeah. Oh, avoid that. Avoided that one last night. On this thing.

What about shameless plugs? Anything you wanna plug? Any anything that we didn't hit with Astro that you wanted to to drop?

No. I mean, there's so much going out this week. Just, yeah. Check our Twitter if you wanna stay, up to date on all the latest. That's, Astro dot. I'm sure it'll be in the show notes. Astro dot build is the full spelled out, handle. And, our Discord community is awesome if you wanna get involved with Astro in any way. Contributing and just hanging out with us on Discord.

That's aster.build as the website slash chat as the URL.

Sick. Well, thanks so much for coming on, and, congratulations on being the 1st 3rd time guest.

We should get some, like, We should get some sweat. I'll trade you a hat for a hat. Yeah. Three x hat or something. You could be the, only one. I'm just gonna make my own hat with, like, a Sharpie and,

I don't know. Piece of paper. That kid. Alright. Well, thanks so much for coming on. Appreciate it, and we'll catch everyone later.

Ace.

Head on over to syntax.fm for a full archive of all of our shows.

And don't forget to subscribe in your podcast player or drop a review if you like this show.