Hacker News Reader

Comment by bluejekyll | Hacker News Reader

After working with react, IMO angular 2 is probably going to come to the game too late. If it's anywhere as complex as 1, then I see no point in using it. React takes the cake in simplicity and expressiveness. It's also a truly component oriented system.

View on HN

From looking at the Angular 2 alpha, design documents, etc., it looks to end up with considerably less accidental complexity (from an application developer point of view) than version 1, and it has a lot of ideas under the hood which are reminiscent of React.

I think the most visible difference will remain that React is code centric, with templates as a sort of special case of code. Angular (1, 2) is more explicitly divided between code and template.

I believe the main driver in choosing between them will not be technical, but cultural. There are many organizations where considerable value is ascribed to the notion of code versus template, such organizations might lean toward Angular (especially in version 2) over React and friends.

Also if you've got an Angular 1.x app already, moving to 2, while probably just as hard as moving to React or something else, let's you remain soley in the "Angular community" where people in similar positions of straddling both versions can share knowledge.

Honestly, there are plenty of other options out there besides Angular and even React, despite the tremendous hype around it at the moment.

Aurelia seems to be the best ideas from Angular + the best ideas from the creator's previous framework, Durandal, in one framework with much cleaner syntax and much less complexity than both Angular & Angular 2.0.

Ember has made some big updates recently, rewriting their rendering engine to use a virtual DOM.

A lot of people still seem to really like Meteor.

Mercury, Mithril, Ractive, JSBlocks, etc. are all smaller, lesser known projects with a lot of potential.

> A lot of people still seem to really like Meteor.

Meteor officially supports React and Angular. Meteor is a web and mobile development platform, not a front-end framework.

True, but by default it doesn't use React or Angular and provides Blaze for creating UIs.

It supports React and Angular as of recently, I suspect, simply to maintain itself as a competitor to other stacks, due to the extreme popularity of React and Angular.

Not true, I still get recruiting emails every week asking for angular experience, which means there is probably a huge group of people locked into the angular framework, and since 2.0 will be backwards compatible, I'm sure a decent amount of people will be switching their angular 1.0 apps to 2.0. It will be easier then converting a whole angular stack to React and figuring out all of the other React libraries and frameworks you need to use.

Learning all of the libraries and frameworks for React is quite complex and not less complex then learning Angular 2.0. with its backwards compatible support.

Angular 2.0 is backwards compatible the same way that React is compatible with Angular 1.0... basically it's mainly just the new Router that makes it possible to use other libraries easier with 1.0.

React never solved the architecture issue. With Angular 1.x and dependency injection , it solved that problem for you. React is just a view layer.

With Reflux or similar, you have everything you need. Simple to build UI, simple to hookup to backend, simple to dynamically update backend and front end based on change or events on data.

When you say it's not a "framework" I don't even understand what you mean by "framework". If you mean a system that locks you into one way of doing things and is hard to integrate with other tools, then by all means, it's not a framework, and then I don't want it anyway.

> When you say it's not a "framework"

I never said that , don't put words in my mouth. The rest of your comment is irrelevant since you're making stuff up.

The architecture problem was solved separately in Flux: https://facebook.github.io/flux/docs/overview.html

...and Flux is a philosophy or recommendation, not an implementaton. There are various independent Flux implementations out there, but none that have been battle-tested at the enterprise level AFAIK. Plus, from recent rumblings, Facebook seems to be moving "beyond Flux".

React is not a framework...

Yes, the future will be web components, which would mean Polymer / x-tag until the spec is in all modern browsers. React will die when web components don't need polyfills.

All major web frameworks have settled on a pattern of using components to manage not only private DOM, but also to manage data scoping. The web components standard is a long, long way from doing anything related to data-passing.

For anything larger than a simple widget, I highly doubt you will see a move to use native web components for a long time. There are application architecture issues they have not even begun to tackle.

Reply on news.ycombinator.com