EmberCamp 2013 Recap

Feb 16, 2013

By Jamison Dance

A recap of EmberCamp 2013, the first conference for Ember.js.

Despite having neither actual fires or actual camping, EmberCamp was rad. I had a few thoughts, and then used some electrons and wafers of silicon to transform those thoughts into words.

Ember is ready for production

Ember has been in alpha and beta for a while, with constant breaking API changes and lots of rough edges. The highlight of the conference was the 1.0 release candidate announcement. No more breaking changes! Ember apps that work with 1.0 will work without modification until 2.0, which basically means you don't have to be a masochist/early adopter/hipster to use Ember in production.

Ember Data is not ready for production

If you use Rails, and use the ActiveModel::Serializers gem, and don't do anything crazy with your models, and are willing to spend lots of time reading the source, then Ember Data might be a good fit for you. The promise of having a single canonical representation for server-side data on the client is seductive. Unfortunately it is seductive in a "sirens trying to murder you by making you crash into rocks" way, not in a "delicious jelly doughnut that you finally give in to devouring" way. The API is not stable or locked down, and there are major conceptual problems that still need to be thought through and implemented to make Ember Data robust enough for general use.

The Ember team is aware of this, and I imagine Ember Data will receive a lot more love after 1.0 is released, so expect it to improve.

Ember cares about the web

Yehuda Katz spent part of his talk speaking about the unique strengths of the web. The Ember team wants to make it easy to build apps that are better than desktop quality by combining the speed and interactivity of native apps with the accessibility and openness of the web. This means embracing URLs for linking to each state of your app. The Ember router is designed to give this to you by default. It also means working with upcoming technologies like Web Components and upcoming JavaScript language features like ES6 Proxies and ES6 Object.observe to make development easier and faster.

Also check out Trek Glowacki's talk about genuinely reusable code and UI elements. This is still a tough problem to solve, and it is encouraging to see Ember core people concerned about making the web better.

The Ember Inspector is amazing and I want to subscribe to its newsletter

Yehuda showed off the Ember Inspector, a plugin for the Chrome developer tools that makes debugging Ember and its abstractions much easier. You can inspect controllers by clicking on UI elements that are rendered in their context, easily check out models and their properties, and avoid the frustration of endless chains of digging deep through the object graph of Ember internals. It looks really useful and very impressive.

It is not scary to contribute to Ember

There are at least two core team members (Trek and Peter) who are responsible for making Ember easier to understand and contribute to. Peter manages GitHub issues and Trek is the documentation czar. If you find a bug, there are smart people willing to help you track it down and fix it. If you find an area that is hard to understand or poorly documented, there are also smart people willing to help you make it better.

Concerns

Or, "Now, with the powers of constructive criticism, I summon people who know lots about Ember to tell me why I am wrong!"

Ember <3 Rails

The RESTAdapter, the only production adapter that Ember Data ships with, is optimized for Rails. One of the talks about Ember Data had about as much Ruby code as JavaScript. Another talk was entitled "Client Side Validations for Ember", but the talk was about automatically turning Rails ActiveModel validations in to JavaScript code.

edit: Tom Dale pointed out that Rails is an easily-targeted platform since it can emit JSON in a standard format, while other platforms may not have common conventions on data formatting.

Ember already has the reputation as the "Rails of the client MVC frameworks," and there is a real danger that Ember will become the "Rails client MVC framework." Why is this concerning? After all, the client has to get data from somewhere, and Rails is a fine back end to get it from.

The web is amazing partly because it is a meta-platform. HTML, CSS and JS are general enough that there are dozens or hundreds of different toolchains for creating apps. This freedom fosters rapid innovation. Tying a client-side framework like Ember to a server-side framework makes everyone who doesn't use that framework a second-class citizen and hurt its adoption.

I understand that by targeting some Ember Data features specifically at Rails, Ember can appeal to a large exisiting community of developers who follow a common set of conventions. I would love to see (and am working to produce) more information on writing an Ember Data adapter if the RESTAdapter doesn't fit.

I would also love to see more examples of Ember applications built by tools other than the Rails Asset Pipeline or Rack::Pipeline. Trek's example app is a good start, as are efforts like Ryan Florence's ember-tools. I hope things continue to improve in this area.

Summary

Despite these concerns, EmberCamp was a great experience and I hope to be at the next one. If you are on the fence about Ember, now is a great time to check it out.

More Ember notes

Carlos Cardona took detailed notes about each session if you want to relive the experience before the videos go up.

Acknowledgements

Thanks to Brandon Hays, Tom Dale and Stanley Stuart for reviewing this.

Jamison Dance enjoys kittens, making computers smarter and the magic of children's laughter. He tweets at @jergason and GitHubs as jergason. He also podcasts on JavaScript Jabber and Soft Skills Engineering.