Is anyone here using Ember on big projects? How's it working out?
I tried ember, maybe 6 or 7 years ago, and was turned off by some weird magical behaviour, but I'm curious about how ember is doing these days, and particularly the glimmer engine.
But Ember gets a bad wrap from the internet because (admittedly) their marketing materials with the gooofy chipmunks make it look like a toy while React's make it look like you're learning quantum physics. And then there's the timeless "convention over configuration" divide where Ember people are generally on the left side, and React people just love the thought of fiddling around and hand rolling their own custom ungoogleable bespoke artisan framework before actually getting to the meat of the work.
I highly recommend using Ember for any size project if you like getting shit done.
This is a good talk kinda about this: https://www.youtube.com/watch?v=rWH3UNIKFeg
More importantly, I think that frontend development still is influenced by its long history of "duct tape" coding practices resulting from JavaScript and CSS being very crippled for much of their existence. Thus, many JavaScript developers seem to prefer simple libraries that can do a lot with a little investment in configuration and composition.
Of course there's also the fact that Tilde is the most notable supporter of Ember, whereas Angular has Google and React has Facebook.
But some people, such as myself, like an opinionated toolchain such as Ember.
I worry because I know they’ve hired some of the most talented folks from the Ember community (people who care about performance and scrolling jankyness). This is not a knock on them, but it is worrying that even the best of the best may struggle with larger, more complex apps.
Ember.js is best technology that I've ever dropped, but let me assure you, the Tomster had nothing to do with my decision.
Ember got into the spotlight when Atwood and Co. have launched the Discourse. Back then it looked super impressive - first open source forum software and arguably one of first open source apps following "JS fronted and API backend" architecture. And it was all done with Ember.js, js FW marketing itself as the modern way to build JS apps!
A lot of people wanted to try it out (myself included), some moving from Angular, other from Backbone, and others coming from backend development. If Ember ever had the cool factor, it was that time. Ember.js was the next big thing, a framework for creating ambitious web applications, and all the people came to it to build their ambitious applications, looking up to Discourse as example of such application.
And then those people have hit the roadblock. Ember.js's pitch was building applications, but it's documentation has never moved past explaining individual parts of technology, and into the field of how those should be put together in final application. Angular had millions of easily googleable tutorials like "this is best practice for setting page title when user enters the route" or "this is how you can configure framework to authenticate with your backend technology".
I also clearly remember the "calling set on destroyed object" error that happened when you tried to update the state of object that was already destroyed that was impossible to debug, because the error was literally "Error: calling set on destroyed object" and it put world to stop. For comparision, in React.js that was "Warning: calling setState with X, Y and Z on unmounted component ComponentName". That one error was absolute productivity killer when debugging Ember.js apps from day one, and it was years before somebody from Ember.js core actually went to Ember's source code and changed "assert !obj._is_destroyed 'calling set on destroyed object'" to "assert !obj._is_destroyed 'calling set on destroyed object'". There was always bigger fish to fry for Ember.js core team, who were so bussy making blogpost after blogpost of new great features or API changes happening, while being oblivious to comments below those same blogposts asking "can I now do partial updates on models because they are super costful to update?" or "is calling set on destroyed object fixed yet?". And then guys from their ambassador product, Discourse, started writing blog posts following the pattern of "Here's how we've replaced Ember.js's feature with our one because framework's didnt work", and those involved things like `Ember-Data` or Ember views.
I always felt like EmberJS meant well but chose rigidness where it should have been flexible, and flexibility where it should have been rigid. It's probably changed and solved many of these early problems. I want to go back and re-evaluate, but nowadays so many different technologies, frameworks, languages, and tools, are pulling at my time that I don't think I will be able to.
FWIW my current stack of choice (and at work) is Golang and Vue (w/Typescript) - which might identify where some of my biases and predilections are.
Umm... those are the same?
Ember.assert(`calling set('${inspect(keyName)}') on a destroyed object ${inspect(obj)}`, !obj.isDestroyed);
This is not at all correct
We've grown from 10 to 150 engineers in that time and Ember's strong conventions has helped us continue our tradition of enabling new engineers shipping to production on their first day and shipping a feature to production in their first week.
If you don't mind elaborating, which version of ember did you start with? Follow on: can you comment on how ember has improved since then?
I'm particularly interested in improvements to performance and error reporting, which I'm lead to believe were pretty bad in early versions.
Performance is great now. The rendering engine has been rebuilt (without major changes to the templating syntax) and the new glimmer-vm architecture yielded huge performance and payload size wins and continues to deliver incremental performance gains every 6 weeks.
I've never considered error reporting an issue in the past, but I may just not remember the pain. It's not an issue for us now though.
If you (or anyone else reading) would ever like to chat in person, feel free to reach out:
The things that have improved:
1. Performance, performance, performance. It’s not fair, but anything that existed before react really was slow, because all JS dom updating was basically .innerHtml. Once we (as a JS community left that), everything has gotten faster (especially ember). Seehttps://madhatted.com/2016/11/30/5-things-to-know-about-embe...
2. Composability of components: being able to compose components together dynamically based and have a well defined boundary for colmunicating with the outside world is great. There is still some papercuts with the actual component API, but those are getting smoothed out.
3. Build-tooling/ecosystem. The amount and quality of tools keeps going up to accomplish tree shaking, AST parsing, etc. I just updated an ember app from 2.3 -> 3.0 and it took less than 2 hours to upgrade, test, deploy.
And fastboot.
And glimmer.
It's really nice environment to work in when building a full featured web-app.