Ember Data Is Pre-Alpha Software

Posted on May 28, 2013 - Subscribe - Home

I decided to commit to Ember Data because of it's value promise. I think that ember data is going to change the way we write web applications. However we're not there yet. Ember Data is still under rapid, heavy, and usually with breaking changes. ED's github project page has a warning banner. It states this. I think many people are unsure of how strong the warning is. This post is for those of you are are (for whatever reason) considering using or getting involved with Ember Data.

I think I have a pretty high debugging tolerance. Most of the time don't mind having to debug other peoples code. It's fun in its own weird way to open their code in an editor and poke around. It's also fun to solve problems in foreign code. If you choose to use ED today you will spent a fair amount of time debugging ED internals. This is simply a fact of life. There is going to be weird errors. My coworker and I spent two days debugging a series of issues all related to a missing error handler on the end of promise chains. I immediately opened a PR once we knew the root cause. This PR has been merged. This will prevent people from wasting their time like we did. The PR made it into the first official release of ember data announced today. Now you guys won't get burned like I did.

We were lucky we were able to debug our problems. I'm knowledgeable about ED's internals. I know about the different parts. I know where problems are likely to originate. But you probably don't. Maybe you just want to try it out. That's fine, just don't expect anything to work--at least not without some work on your part. If you want things to work you'll have to take some time to learn and understand ED's architecture.

ED is a very ambitious project. It's attempting to solve something that's never been done in the browser before. I think it's going to fundamentally change the way we write applications. There are many of use out there who believe in the project. At this point it requires a very special type of person to get involved. You have to accept the risk of using pre-alpha software. These are the decisions I had to make:

  1. You are comfortable building form source (How many of you would only install something if it's in homebrew?)
  2. You are willing to actually spend time understanding ED's internals to the point where it's like your own code.
  3. You are willing to give constructive feedback and fix bugs you find
  4. You understand the previous 3 points and accept that will be much harder and painful than you expect.

If you can commit to those points and the inherent risks of using early stage software then I recommend you check out ED. If you are not ok with any of them then simply wait. Don't get involved with it if you're not willing to accept the responsibilities and risks.

Luckily the Ember Data core team and other contributors are aware of the issues facing the project. They have refocused their effors over the past few weeks. Stefan Penner has done some great work implementing promises inside ED. I'm hoping to work with him in the coming weeks to implement promises in other parts of the architecture. I'm primarily focused on improving the greenfield experience. I think many users have heavily customized setups. They are far removed from the out of the box experience. I'd like to improve the developer friendliness so newcomers don't give up in frustration. Paul Chavard is always hard at work improving ED's feature set. He's been working on adding real server side validation handling to ED for some time. Igor Terzic is also helping out a bunch. Of course Tom and Yehuda are involved as well. The team is meeting biweekly to discuss what needs to be done. Hopefully this and quick public releases will really move the project forward. None of us can wait for the day when we can release Ember Data 1.0--hell even 1.0-beta1. We need all the help we can get to make it though. We need people, willing to accept the risks and responsiblities, to help us find bugs and improve the core concepts. We will get there. Give us time and we'll show you.

Discuss on Hacker News.

— Adam Hawkins