typehtml v0.0.0

TypeHtml

I am just dreaming of a better world here 🌹

Imagine the following API for creating component using functions

const Heading = ({text}) => <h1>{text}</h1>

or using classes

class Heading {
  constructor(
    private props: {text:string}
  ) {}

  render() {
    return <h1 onClick={this.handleClick}>{this.props.text}</h1>
  }

  handleClick = () => console.log('hello')
}

• All state is powered by mobx.
• All code is written in and for TypeScript.

Like to see this dream become a reality? Go ahead and ⭐ it!

Why?

I absolutely love react and recently shared my love as well. That said:

• Every single time I need to run npm install react react-dom @type/react @types/react-dom a part of me dies a little.
• Having external type definitions for a library is not enough. You need to think TypeScript first when designing the API.
• It also gives an opportunity for people to complain about TypeScript whereas in reality the alternative is really an non-analysed mess (if the compiler can't analyse it, neither can new dev to the code, have empathy).
• React has stuff I no longer need e.g. setState and abstractions to support react native. Learn once and rewrite wastes human resources.
• Also the React codebase would benefit from TypeScript / More type annotations

Key Differentiators

Beyond being in / for TypeScript, the following are a few other differentiators.

It is nearly impossible to know yet if these ideas are good or bad. In a magical world where this idea performs well, it definitely results in less developer code.

• Components are pure.
  • They are always a side effect of a mobx observed change.
  • You wouldn't want a reaction to happen unless an observable changed, similarly you don't want a component re-render unless a prop (or local observable for classes) changes.
• Components using functions and classes have the same features (results in simpler code base + types + learning curve).
• Context is just a fancy global (and an untyped one at that). And there are better ways to handle globals, so removed it.