I can’t say enough about how ReactJS stands alone from all the other KVO (Key Value Observation) frameworks. @floydophone is absolutely right in his two videos on ReactJS. 100%!
I always do an Ajax call from a Flux store. It works great. That being said, server side rendering would be great for content that should be SEO indexed. I created a github issue to add server side rendering to the example repo.
This video explains design philosophy of React and why MVC is not the right model for building UIs.
This video compares ReactJs vs. Key Value Observation(EmberJs) and Dirty Checking (AngularJS).
I’m also a fan of React with Rails. I’ve been using the ‘Reflux’ implementation of flux - just great.
My test suite is: Teaspoon / Jasmine 1.3.1 / Guard / PhantomJS. Code and specs written in coffeescript. No JSX. My experience is: 100% bulletproof - no intermittent errors.
I tend to populate the flux stores via JSON objects rendered server-side into a ‘script’ tag, then use AJAX for updates.
I dropped Angular in favor of this stack, and would never go back. Because simplicity, lower conceptual overhead, easier to train new people.
JSX, IMHO, is one of the best parts of ReactJs. And if you’re using CoffeeScript + browserify, then I can see why you wouldn’t use JSX. Wepack with ES6 + JSX + npm packages is
I’m also a huge fan of JSX and really don’t like CoffeeScript. The whole, “Write code in one language and get error messages in another” is just way too big a turn-off for me…