How to use RequireJS in AngularJS Application?

Javascript gurus convince me that I should use javascript over compile to js languages like Dart or Typescript?

• Hello I am about to start developing massive frontend application and I researching what is my best option for client side ecosystem. I am a guy with Java background, doing Spring applications for some time and during my spare working on my private project that most of it's logic parts reside in browser. It's a web based image processing and diagramming application that requires lot of graphic processing, advanced user interactions like drag and drop and I expect codebase will be large enough. I had dealt with poorly written large js codebase. It's quite headache to me. Then I have developed bad reputation for javascript. Initially I was looking at Google web toolkit. But i found it's not suitable to my application, cost of developing in GWT was very expensive because of it's complex abstraction. I finished research and planning phase. Also recently I switching to Ruby for my day to day work. I'm tracking http://www.dartlang.org/ development since it's inception and very impressed with it. I love the features like classes, mixin, built in approach of library(For me it's the huge benefit over Javasript), Stream/Promise and Web Components. I like the it's nature of it's strictness. Optional static typing of dart language minimizes misbehavior of the application and help me write more effective/optimized code. Importantly I can apply familiar design patterns and programming techniques used in Java world. Right now I'm decided with go with it or gave more points to it. More recently I spent a lot of time in discovering more advanced usage of Javascript. Just finished reading John Resig's "Secrets of the JavaScript Ninja" and quite opened my eyes. I can see using RequireJS and Grunt can manage my large codebase. I read the arguments of people saying long time risks of using Dart. Concerns are: Compiled javascript source code can have unexpected bug and it can possibly to affect application behavior. Javascript is evolving. Ecma.next promises nice features like modules, maps and classes. And even right we can write Ecma.next code using traceur. If Google shipped native dart vm in chrome. Performance of my application will differ in other browsers. It's not good. It's not widely adopted so when your application grow it's hard to find good people who code in Dart. For me these are not convincing enough to give up on Dart. So if you have some interesting point on both Dart and Javascript side please tell me. This may sound like I'm confusing but it's just a part of my research. Of course there's other parts of development like app architecture/design and patterns in addition to particular programming language or technology.

• Answer:

  As a famous quote in Computer science goes. All problems in computer science can be solved by another level of indirection Except problems caused by too many levels of indirections. Let me elaborate on a previous experience that I hope will be useful in this instance. I worked on a project that used Hibernate 3 when it initially came out. The team members were Java veterans. Hibernate enabled us to Avoid the syntactic mess that was JDBC. i.e. avoid all the prepared statements and setters of JDBC and use annotations instead. The resulting code was concise and modeled our application accurately. Allowed us to deliver a complex app in a short timeframe. Ensured (in part) that our code was maintainable. Functional and load testing went well and was deployed to production. As soon as we hit approximately 10% the expected load the application slowed down to a crawl. Turns out the load tests were not representative of read and write operations. As you may have guessed we got so lost in the abstract Hibernate layer and how it solved all our problems that we forgot that it was backed by a relational database with locks, keys and indexes. When we turned on SQL logging it became apparent that the SQL being generated had too many joins and our (hibernate generated) DB schema was a detailed guide of what not to do when designing a performant Database schema. Each of us understood SQL enough to never have come up with the schema that resulted. The reason that schema came to be was our reliance on our ORM to do a job that we should have done. Here are a few questions I would ask when adding a layer of abstraction to any real world projects I work on now. Does it have a large enough support and adoption to easily troubleshoot complex problems which I do not anticipate? For the features that this abstraction promises are there enough developers in the team who truly understand those features? And what happens when we bring in a new team member? How easy would it be to ramp up a fresh college grad to an acceptable level of proficiency such that their code does not bring the whole system down? This is a measure of how easy is it to mess up when using this abstraction. Given the above can I afford that risk to the project at hand? If I can answer those questions satisfactorily,only then will I use the abstraction :-)