Sticking to web tech (and reposting from twitter) my main prediction is a move away from “Rehydration” or “Resumability”, and a better understanding of holding state either on the server, or in the browser, (mostly) not both.
HTMX becomes the fastest growing frontend toolkit in 2023 as devs feeling stack fatigue migrate from React and other JS toolkits towards a simpler architecture. A draft “HTML 6” spec is written incorporating ideas from HTMX and HATEOAS. Smaller front end JS toolkits like AlpineJS, "petite-vue", and HyperScript will explode in popularity, accelerated by the move towards basic server side html rendering.
At the same time “offline first” and PWAs make giant leaps forward, off the back of WASM based SQLite and CRDT toolkits Yjs and Automerge. Offline first webapps will have their RoR/Django moment, a new BYOFE (bring your own frontend) framework based on SQLite in the browser, eventual consistency, and automatic schema migrations. Probably using Postgresql as the source of truth.
There will be a move away from React Native back towards WebView based toolkits such as
Capacitor and Tauri (I think Tauri will win over Electron in the long run), as people realise that a web view is just as performant. Complimented by strategic use of NativeScript.
Ultimately there will be a split, if you need a fast initial reader and want simpler development, “HTML over the wire” and HATEOAS will become standard. No more “Rehydration” or “Resumability”. No more JSON APIs that are only ever used to render your html template.
Alternatively if you are aiming for an app-like experience you will go “offline first” caching as much data on the client, implementing traditional server side functionality in the browser. Again no more “Rehydration” or “Resumability” as (almost) everything is client side. No more splitting state between client and server, just pick which side your particular product sits best in.
> There will be a move away from React Native back towards WebView based toolkits such as Capacitor and Tauri (I think Tauri will win over Electron in the long run), as people realise that a web view is just as performant. Complimented by strategic use of NativeScript.
> HTMX becomes the fastest growing frontend toolkit in 2023 as devs feeling stack fatigue migrate from React and other JS toolkits towards a simpler architecture.
Haha, thanks for the laugh. Don't underestimate the network effects of a behemoth like React. Even React frameworks are starting to tend towards server side, like NextJS or Remix.
I would like to see the offline first philosophy adopted more thoroughly though.
So, it does seem funny, and I'm being a little OTT. But I do think it could genuinely be the fastest growing framework this year, it's absolutely not going to overtake React in total market share any time soon. But momentum wise it's going to be like a rocket.
Very different things. Native script is a bridge between the iOS/Android apis and JS/TS. So in a single code base JS or TS app you can call out to native apis such a the camera or taptic engine.
HTMX becomes the fastest growing frontend toolkit in 2023 as devs feeling stack fatigue migrate from React and other JS toolkits towards a simpler architecture. A draft “HTML 6” spec is written incorporating ideas from HTMX and HATEOAS. Smaller front end JS toolkits like AlpineJS, "petite-vue", and HyperScript will explode in popularity, accelerated by the move towards basic server side html rendering.
At the same time “offline first” and PWAs make giant leaps forward, off the back of WASM based SQLite and CRDT toolkits Yjs and Automerge. Offline first webapps will have their RoR/Django moment, a new BYOFE (bring your own frontend) framework based on SQLite in the browser, eventual consistency, and automatic schema migrations. Probably using Postgresql as the source of truth.
There will be a move away from React Native back towards WebView based toolkits such as Capacitor and Tauri (I think Tauri will win over Electron in the long run), as people realise that a web view is just as performant. Complimented by strategic use of NativeScript.
Ultimately there will be a split, if you need a fast initial reader and want simpler development, “HTML over the wire” and HATEOAS will become standard. No more “Rehydration” or “Resumability”. No more JSON APIs that are only ever used to render your html template.
Alternatively if you are aiming for an app-like experience you will go “offline first” caching as much data on the client, implementing traditional server side functionality in the browser. Again no more “Rehydration” or “Resumability” as (almost) everything is client side. No more splitting state between client and server, just pick which side your particular product sits best in.
(Originally posted in extended form here: https://mobile.twitter.com/samwillis/status/1608542221166587...)