I've only learned about Svelte from another react vs angular comparison video about a month ago. Got addicted to everything Svelte does and the way it works. It makes the development so much better experience compared to the other overwhelming stuff out there. Decided to use Svelte in large scale production apps. Very happy about my decision to use Svelte and being able to stay away from the mainstream Angular and React BS.
Not sure if Rich reads these comments, but an even more comprehensive framework built on top of this new tech would be so welcome! I'm imagining something like Django/Flask or Rails but with Svelte underlying everything. I would use something like that in a heartbeat - currently a lot of web frameworks have significant problems, especially in bridging the front/backend gap and it makes doing web development tiresome.
I've been using sapper to migrate (mostly) static-content websites that used a PHP CMS from my main web hosting server to netlify for the past year and the results are just fantastic. The sapper sites run like knife on butter, my server resources are back available to larger e-commerce projects and all I had to invest was a few weekends of my time. I spent a good sum of time in late 2019 researching how to accompish what I did and I went through Gatsby and a sh|tload of other frameworks and libraries that either had a steep learning curve or you needed 10 files and 50 lines of code just to add 1 navigation route (at least efficiently). The work you've done with svelte is absolutely brilliant and moves like this (svelte-kit) prove you are here to stay. My only complaint is about popularity. In my country (Greece) the js usage goes like: 40% Bootstrap & jQuery (iknr) 50% React 8% Angular 2% All others (vue etc) 0% Svelte (all colleagues I talk to hear it from me first time)
oO που χρησιμοποιείται το React; Όλα τα projects τα οποία έχω δει να παίζουν εκεί που δουλεύω είναι όλα σε Angular (ανάθεμα τα κάτεργα). Όποτε τους αναφέρω το React τρελλαίνονται γιατί σώνει και ντε το Angular είναι framework ενώ το React θέλει βιβλιοθήκες για τα πάντα κλπ, δεν έχει πάρα πολλά components (και καλά) και όλα τα γνωστά. Σοβαρά, ποιές εταιρείες (με target group enterprise πελάτες και χρήστες, όχι site-άδες και αρπαχτούληδες) έχουν επενδύσει στο React εν Ελλάδι;
Rich Harris is some kind of Saint or at least the only person who appear sane, showing up in this total mess of Angular/NextJS etc, bringing us clarity... Thank God, he's a true Hero !
Likewise, even though I kinda new to it; I am loving it and already looking into redoing some projects in it. Plus all new development will be in Svelte
I saw a lot of react killers come and go in the last 5 years, but looking at this video, it made me believe this technical approach is superior than what's the market is offering, same way react did when others were doing two way bindings and scope variables. I am a svelte believer nkw
@Rich Harris: Great decision moving forward. As a new developer to Svelte, I ran into the exact issue you described: deciding whether to make a Sapper or Svelte app. Having just one recommended way to build (and only one framework to learn) is major improvement!
Svelte incorporating key aspects of Sapper, and using Snowpack are both dreams come true. I'm am extremely happy to have chosen this framework as it's not only super fast but intuitive. Thanks so much folks! I even chipped in $5 a month at the opencollective site!
I’m the founder of a small startup and our original tech stack was Elixir, Phoenix and Elm. Prior to building out our first iteration we spent about a month reviewing tools. We looked at Rust, Swift, Python, Haskell, C, Go, Julia, Lua, Elixir, Elm and Dart taking into consideration ecosystem and relevant frameworks and tooling. We always came back to elixir & elm. Elm seemed like and obvious choice for the front end. However it’s proven to be a challenge for some devs and recruiting talent isn’t easy because elm devs are a precious commodity. Typescript sort of change my opinions on JS. So Svelte was on a short list of passing JS curiosities. The past few months my team has been struggling to integrate certain features of the UI in elm. Startups have to move quick because time to market is a real thing. I discovered svelte and knew immediately it would be the perfect solution to solve all of our problems I sold on components, typescript integration and the compiler and the work to support Deno.
I'm a React developer and for the love of god if I had enough free time and I wasn't heavily invested in React I would have switched to Svelte in a heartbeat.
I am a React developer too. After trying so hard to avoid Svelte as much as I could, I finally made the hard choice to stop everything I was doing to take an entire week to learn Svelte. And it was totally worth it in the end!
Been using Svelte for a few months now, loved Sapper - This seems even better! Can't wait for the final form of Svelte Kit, can't wait to get rid of them bundlers! Good luck with development, looking forward to more news
I've built my SaaS platform for a client with just pure Svelte (no Sapper), main reason being almost everything is dynamic. But all this looks like something even I would benefit from. Will be waiting for the release!
This all looks truly awesome. My only wish is that, with this route defined, backwards compatibility starts becoming a priority. I think this is key for Svelte to keep growing. Having to move out of Sapper, though apparently not complicated, is still work that production projects don't appreciate making. Svelte is awesome, and I applaud them for pushing web development forward and embracing new technologies like Snowpack... But I think it's important to keep in mind that if you keep changing too much, it might scare off adoption from the developer community... And community is key.
In my opinion this is not actually an issue. If you only need dynamic content, you can still work with the previous bundlers. Since Sapper will not come out from beta, people that used it in production were aware (or at least should be) that things like that could happen. I think that the most of the work will be to migrate Sapper apps to the Svelte-kit.
@@lennarttenwolde9407 I don't know if you're trolling as you didn't elaborate, but seriously If you ever worked in any company with at least some coworkers then you can not say this or you just didn't tested it out. Doing javascript is fine but not doing typescript when you have other developers you work with is like saying "oh no it's find to use silex to light a fire when you could have a lighter"
yeah it's a pitty meanwhile quasar.dev getting ready for v2. I'm an Angular guy but angular is moving too slow. And I fear like everything Google it will be dropped at one point to chase the next goose around town with no regard to the people who depended on it. And yet... svelte while being really the way to go TM is lagging behind. React and Vue seem to be the only reasonable choices right now.
Richard, I'm blown away by your vision, and looking forward to that goooood stuff 😍 Thank youuuu 🙏🏻 This will, literally change our lives and the world.
Same here !! ... but I realize we can continue with the current versions and when the next one is ready just migrate couple of things here and there, looking forward for it
I gave my all-in vote to Svelte early this year but got confused by Sapper & Svelte's movement (also too busy to learn both). I love this reimagined modern workflow 😍. So eager to try my next project with the new Svelte 😎!
Tried the npm init svelte@next and it was pretty sweet! Only thing that was broken at the time of writing is the node build command after npm run build as it is giving error about not finding "prerendered" directory. Other than that, I'm eagerly waiting for this!
I've been learning Svelte for a while and started building my companies sites with Sapper, but in the back of my head, I was hoping Svelte would have routing, etc built-in eventually and now this! I had to go through Bash for Windows to get building to work right (dev was fine, but build is using unix style commands only it seems), but you have a day one (ish) tester for Svelete@Next. I'm looking forward to the public repo and official release! I hate React. Vue is okay. Angular... But, Svelte makes me actually enjoy front-end development.
Some thing i think you should work on is the experience with images, sass, less. Another thing i have wonder about when you build the the app is way do not create a sitemap, you now the structer of the map so why not give it to the user instenad of user most create it. But i love that you are taking the best of sapper and make it to one framework so good job!
I think you might’ve used “pages” folder name instead of “routes”. NextJS didn’t patent this word :) Thanks for introing svelte-kit, and sharing it via twitter. I personally a bit too coupled to NextJS, - they already have such a good ecosystem, like next-pwa, letting you setup a good defaults for service worker (with ability to extend its functionality ofc), and many other good things too (like next-mdx, or new builtin next-image). But I can’t wait to see some “examples” folder in the new repository, with various libs and demo apps... I definitely wanna see something like NextJS but for Svelte, - that’s for sure :) That’s cool Svelte is evolving! I’m also glad you’re getting some sensible donations so soon after starting the campaign!
I really really like this new concept and if it was stable enough i'd be using it right away. I am in front of the choice to learn React, Vue, Preact and now Svelte. Svelte@next seems to be the best of all I've tried.
bye Gatsby. I enjoy writing Svelte code a lot more but needed a well-developed static site generator for a bigger project and went with Gatsby. even with around 500 pages my build times are 4-5min, so if this ability to tag routes as either static or dynamic puts even a dent in that I will make a migration work.
Gatsby's a bitter development experience. Should've went with Next.js. Server-side rendering? Static site generation? Client-side rendering? No configuration needed.
I don't get how this video is from October and on the Snowpack website it says: "Update (April 20, 2022): Snowpack is no longer actively maintained and is not recommended for new projects."
I currently have a semi-large application written in Vue 2. I'd like to evaluate this vs Vue 3. Does this have a name or is it tied to a version number? When will we know when it's ready for prime time and not in beta?
Really cool to see Rich embrace Snowpack. In some respects, Snowpack competes with Rollup. If Rich had NIH syndrome, he'd likely bristle to including Snowpack in the workflow.
Well, as a naive (I have to admit) full-stack-ey-sometimes dev this looks interesting and perhaps the rationale behind the whole implementation would be worth a discussion / presentation of its own, esp. if it targets people who work for totally corporate / enterprise-level, you know, customers who want total ownership of data and can't trust stuff like Azure or AWS or whatnot. But can we take a moment to acknowledge the fact that what we have here is the Twelfth Doctor going through a "young" perception filter when he finally decided to dabble w/ web development?