I’m not a framework or library person, but they exist for a reason. There’s a need of solving things quickly, a clear growth path and fast adoption for new team members.
Frameworks are exactly that, they are an inversion of control for your developing product, they will tell you how to do something and how to scale it.
Frameworks and libraries exist for a reason
The problem is that they come at a cost that’s usually paid by the user. Big bundles to download, long waiting times to actually see content and not the best out-of-the-box performance in mobile devices.
But again, they are here, they are a thing. So I shaked the denial off and decided to write a simple web application, the exact same one again and again, but with different frameworks and libraries.
Amazed by its increasing popularity I decided to go with Vue first and I’ve found out why a lot of developers are happy with it.
Vue is really well documented and it comes with a set of features and tools that make the experience great, like single files components.
Performance is not a trade off at all, some case studies show it is as good or even faster than other popular libraries.
VueJS strongest points are documentation and ecosystem
I think that its strongest drawback is its self-defined field of action. Initially Vue borrowed notation and approaches from Angular and React plus its own interface and features. Those blurry lines allow you to fall into anti-patterns easily, closing trails some times pushes the developer into the right ones.
Next, probably the one which made component an every day topic in the front end development sphere. React leitmotif is to provide a declarative way to build and render your application.
It could be said that it just handles virtual elements and their internal state that later translates into real DOM.
Facebook team built React with architecture in mind, it pushes you to separate concerns and build reusable small pieces which communicate in a one way flow.
After understanding its concept, building stuff with React is great
It isn’t slow, but it isn’t fast either unless you apply certain patterns in specific parts of your project, shouldComponentUpdate is an example.
To make a performant React app you need knowledge about how it works, knowledge that most of the time a team doesn’t have time to absorb, and though documentation is complete, it definitely needs some gardening.
Though it comes with similar features and goodies, it is hard to find products built with Polymer in the wild.
Polymer extends already existing web features instead of redefining concepts or creating extra layers between your code and what ends up rendered in the browser.
Without that extra layer and taking advantage of the native platform under the hood, it achieves really good performance. Then why isn’t it more popular?
Well, while performance is something almost everyone cares, scaling a big application through HTML imports and loosely declared components is a hard choice to make.
Polymer relays on native features, and native is hard to beat
As someone who prefers vanilla solutions over framework-based ones, I really hope Polymer takes some shape with time and helps you structure and scale your project in a similar way others do.
The first version dominated the framework scene for years, but a lot of years went by since it was released and today the needs are very different.
That’s why the Angular team is betting on a complete rewrite to keep pace.
It won’t be hard to adopt if you are coming from Angular 1.x, but if you are not prepare for a big learning curve. Angular 2 requires a lot of configuration to bootstrap a simple application, but scaling over it is really easy.
When using Angular 2 in your project, you are choosing a slow start for a rapid and organized growth
This second version introduces interesting concepts and improvements under the hood, but they will be hard to sell if documentation and initial configuration aren’t friendly for begginers.
As I mentioned at the beggining, working with frameworks or libraries with heavy influence in the project structure is hard for me.
This was an experiment for me to jump in the game and learn, the result is a set of articles describing my findings.
The intention wasn’t to provide a how to article for each, but to quickly describe and compare how these four work and my opinion about my experience.
I still think frameworks focus more in making life easier for us the developers than for the final users.
Right now, I don’t feel there is one that clearly beats the rest, same results can be achieved with either of them.
To give the user a reliable experience you’ll have to read and learn a lot about best practices and concepts for each of them, no matter what your choice is.