Ask HN: Why is ExtJS not getting any traction? | Hacker News Reader

Ask HN: Why is ExtJS not getting any traction?

24 pointsporker27 comments

It's 2014. I'm looking for a rich-widget JS framework (datagrids, forms, widgets - all the stuff you need for building desktop-equivalent admin focused apps) and all I find is Sencha's ExtJS.

Yet everywhere I turn (esp on HN) people are talking about AngularJS, EmberJS and Backbone-based solutions. While no doubt great, none seem to be focused on back-office tasks. With the rise of building mobile apps in HTML5/JS I'm surprised - I can't be the only person building data-focused apps (and associated CRUD).

So, why do you use these, do they have a sweet spot (custom apps, games, some "I'm a website but I don't want to be" malarkey?) and why not ExtJS? ExtJS sucks, I know that - but it looks the most RAD way to develop admin interfaces, data input etc.

View on HN

Let me preface this with the fact that I am a Flex dev who ventured into the HTML/JS/CSS stack during the "Flexodus". Cleary, ExtJS seemed to suit my fancy more than the other libraries in the JS world.

As far as why it doesn't have traction, I would attribute this to 2 things. The first being that the JS community isn't really into app development. What I mean by that, is when you look at JS, and you look at languages like Flex and Silverlight, you notice a massive disconnect in ideas. In the JS world, visual components are rarely if ever first class citizens, simple layout tasks are far more difficult than they need to be, and SPA paradigms are also second class citizens. After noticing this and poking around, you find out that the main JS community either doesn't want the solution to begin with, or they seem to be re-inventing the wheel.

The second strike against ExtJS is the cost. I got a license for it a few years ago when I was delusional and thought I could use it to replace Flex. Dropping $1,000 on something when virtually every other tech out there is free is hard to choke down.

Also, it seemed like a major step down for those of us coming from the Flex community. Like I said, with JS app paradigms seem like afterthoughts and each peculiarity wore me down further and further. For instance, if you have a button and you declare a listener function but misspell it, you don't get an error complaining that the listener doesn't exist, you get a blank white screen. Another issue was marshalling data into objects. Whenever I brought back a Classroom JSON representation, the Student objects in the students collection remained unmarshalled. I had to use ANOTHER store to get those to behave as objects. All of these "why on earth is this acceptable" moments led me to believe such features weren't really in demand in the JS community if not resented altogether.

Thank you, that's a great answer! I hadn't considered the different community mindset.

I'm curious what you settled on instead of Flex?

You've got me thinking wider than HTML/JS, for something which could be more satisfying (for me): mobile apps could be the way but I need desktop/browser/cross-platform support; Silverlight I've heard good things about but again isn't widely used I believe.

I went back to Flex ='(. Like I said, the lack of basic app principles and the fact the JS community wants to fist-fight you for saying there is a better way was kind of off-putting. If you want desktop support, you might want to look at Adobe AIR. There is also JavaFX, but it appears very, very, very half-baked to me right now.

You should really try out AngularJS. The declarative syntax of it reminds me of Flex quite a bit. I haven't used AS3 or Flex in years and really miss that environment, and when I started using AngularJS I felt like I got a taste of it again.

To me Web Components is in early stages but seems like the closest we are going to get to that perspective/approach. Makes it easy to define your own HTML tags.

Having you checked out KendoUI? They seem to be somewhere between the ExtJS/AngularJS spectrum.

I've used Sencha's frameworks on 3 different projects now, and I still hate it. For many different reasons:

1. The learning curve: This is probably something you already know. But you may not be aware of is the other effects this can cause. If your team consists of members of different level of experience (from 0+), the code quality is going to suffer a lot from the lack of knowledge.

2. ExtJS and Sencha Touch are 2 different products. Today's web app demands to be runnable on many different platforms. With ExtJS you get point and click. With Sencha Touch you get swipe and pinch. But what if you're running it on a touch laptop that needs both point and click and swipe and pinch? I'm working on a product right now that integrated the 2. While we succeeded at reaching the goal, the result is not something I'm proud of.

3. Sencha is a walled garden that makes it hard to integrate other libraries into it. For the most part, Sencha already have a lot of things that you needed (like data grid, combo box, etc). But what if you want to do complex data visualization with d3? Complex interactive behavior with Rxjs? Or realtime updates with socket.io?

Eventually, I found out that no matter which framework you choose to use, you'll end up needing to read its source code to understand what's underneath. If you have to do so anyway, why not pick a framework that's simple to understand?

Worth mentioning here that KendoUI Web is dual licensed (Commercial and GPLv3). KendoUI DataViz, Mobile, and the various server wrappers are Commercial license only.

PS. I use a combination of ExtJs and KendoUI at work.

Thanks, I've looked at KendoUI from their 'Preview' days onwards, when the license was so confused (and no one could answer) that I gave up. I've been back since and read through the demos, but to be honest I'm not a fan. KendoUI doesn't feel cohesive, and while I'm pleased they advertise framework bindings, having their own MVVM implementation seems crazy.

So, I'm biased against KendoUI as my gut feeling says something's not right with it, but have no actual experience to back that up ;)

Completely agree re the learning curve and hating ExtJS - I've 'used' a bad implementation of ExtJS 3.2 in the MODX CMF [0] where it powers the CMS's admin and it's hideous.

The backend integration is completely optional. You can put any REST controller (given that they follow Kendo's simple protocol) behind the widgets and everything can be bound using Javascript.

It integrates well with other Javascript libraries. But not so much with Angular, which is more because Angular doesn't play well with anything than a flaw in Kendo.

Most ExtJS apps are probably internal admin apps, but I imagine it has decent usage throughout the corporate world. 5 Years ago if you were looking to create a SPA business oriented web-app extjs was the only option really, and at the time it also looked damn nice (against java generated web-apps).

It seems the ExtJS community has fallen off a bit as the company has (in my opinion) put it in the rear-view as they focus on Sencha and mobile. I also never really felt that I have seen what a true ExtJS app could be. The examples are the same as they were 5 years ago. Only the theme has been changed. You get a lot for free with the datagrid and charts but they aren't the only game in town for this.

No one wanted to build customer facing ExtJS apps because of performance issues, hard to style so that it doesn't look like some dull enterprise thing and at the time Rails was just getting popular. I havn't touched ExtJS in a while but building an admin CRUD in Rails would be much easier. The extra features of ExtJS can be gotten elsewhere and easily integrated.

Angular just seems like a breath of fresh air when comparing it to ExtJS, and again most of the components can be found elsewhere.

I think this is exactly correct. I've used ExtJS exactly once and it was for building an internal, administrative app.

It's also correct that building an administrative interface with Rails or Django is much easier. In the case of the app I was building we had to interface with some Django code, so I wrote this Python meta-class that would inspect Django models and produce classes that would then produce ExtJS models, widgets, etc. that could then interact RESTful interfaces provided by the Django app. It worked great but was slow due to javascript blocking issues. So like you said performance issues were a barrier, but I doubt in that case it was ExtJS's fault. I've thought about posting that code somewhere but I'm kind of embarrassed that I actually tried it. After that little experiment we scrapped ExtJS and built a strait Django application that did what we needed.

Cost. In my day job we use extJS so I know it pretty well. I would have been the first one to use it in my side project. But because pretty much every other javascript frameworks is free and the extJS licensing is terrible they weren't even in the decision.

TL;DR: I use Angular (or * framework) simply because I can rapidly develop the features and adhere to ALL the requirements. ExtJS doesn't suck, but it typically doesn't fit into scope or doesn't fit all the requirements.

ExtJS is a great way to facilitate development when working with large data sets that involve (as you mentioned) CRUD utilities along with business validation constraints. With that being said, these kind of tools (And KendoUI too IMO) are being stuck inside intranet systems where the custom widgets and tools remain proprietary. No visibility == no hype.

I work exclusively on FE stacks and if the choice is Angular or any other top TODOMC* frameworks, typically that means the requirements don't fit well with ExtJS. Most JS-engine driven apps typically require a responsive implementation (cross platform is a big business trend), ability to incorporate a comprehensive design, and being able to do it within a reasonable amount of time (which is associated with budget).

ExtJS has a ton of overhead (especially browser performance and scalability in SPA sites), a large learning curve (IMO), and requires a ton of time to implement custom designs. This doesn't mean there aren't great tutorials or start kits, but its not something you can master within a week. Responsive support OOTB isn't there (KendoUI has solutions for this on their framework) and the only choice for reusable code for portable devices is an adaptive approach using Sencha Touch. This doesn't mean ExtJS is a bad choice, but typically ExtJS excels in desktop apps that are data driven with designs that were made for ExtJS (its rare to not recognize an Ext app if its built on Ext).

There's my two cents. I'm sure Sencha is going to destroy me if they ever read this.

Absolutely correct. ExtJS definitely excels at Desktop applications (pushing and pulling data sets from departmental DB's). That's we use it for and it's fine for that. However, the learning curve is fairly steep.. 4.x is much better than 3.x, more structured with some nice MVC hooks. Unfortunately, I would never use this for anything BUT an internal app.

TL;DR: because ExtJS sucks.

Long answer:

1. The learning curve is really steep. I mean, really really steep. There's a critical shortage of (good) books, video, and tutorials on the web. I guess ExtJS' learning resources should be 0.1% of jQuery's. The examples on Sencha site were still using the 3 version last time I checked;

2. If you need to customize ExtJS, or need something slightly more sophisticated than CRUD screen, then you are pretty much on your own. You'll spend a large amount of time trying to "fight" the framework and your chances you succeeding are very low;

3. It's expensive and its open source license has restrictions;

4. Sencha seems to hide and obscure the access to the learning resources so that you are pushed towards their paid support.

5. It's slow;

6. The error messages (when they show up!) are cryptic.

7. ExtJS was re-designed for version 4, but some things like the Model objects are cumbersome and suffer many usage limitations;

8. It will never be adopted as a general purpose solution on Internet facing webapps for the reasons exposed above;

Last time I tried ExtJS, it struck me as a framework made by people who suffer from an extreme case of NIH syndrome. Personally, I prefer small, light frameworks that play with other frameworks well. I don't like big "uber" frameworks that re-implement everything. That's not value, in fact, it's negative value in my mind.

If I want to manipulate the DOM or do AJAX I'll use JQuery, if I want to do data binding I'll use knockout.js, etc.

But these are minor annoyances compared to the huge ExtJS no-no for me which is the fact that ExtJS seem to actually hate Javascript. Or, they targeted the framework towards people who'd rather write C++ in the browser - giant "class" hierarchies everywhere. Ugh.

Could the problem not be the lack of desktop-looking ui-frameworks, but a lack of willingness to design something that doesn't look like a java swing app from the 90's. (Yes, ExtJS is butt-ugly).

Could the app you want to write maybe lend itself well to the UI-stylings of Bootstrap? Or Pure? Or Foundation?

Or maybe you could combine something yourself... Googling Angular CRUD gives me this: http://angularjs-webapi-crud-grid.azurewebsites.net/

I mostly use Extjs for admins and such. I would say that it offers little control over the design, structure and HTML compared to something like angular which allows full control over the presentation. Creativity room is important.

The licensing is pretty restrictive also and the prices are big.

I wouldn't say ext has no traction. I see sencha going pretty well and I've worked with extjs since the early days. It just looks more "business oriented"

The licensing is bad, the control over the design is terrible, and the experience developing and learning it is painful.

Never ever use it for a client who has any ambitions on GUI design. Figuring out how to put an extra pixel here or there can take hours...

http://qooxdoo.org/

I came across this just yesterday. Would love to hear experiences with this. Thanks.

A guy I know built a product with it that went nowhere:

http://normforge.org

He indicated the framework was not something he would use again.

I have no experience with ExtJS, so I can't really compare the two. I've only used the Desktop variant of Qooxdoo.

It's easy to learn, but hard to master. Though there is quite some extensive documentation, you will have to take glances at the qooxdoo sources to fully understand some more advanced stuff. And to track down if something doesn't behave as expected. The source is quite clean and readable though, e.g. the button class: http://goo.gl/Zw6zNI

There's also a simple playground http://goo.gl/QAht5O and an API viewer implemented in qooxdoo itself http://goo.gl/xV5tvn

From the top of my head some random facts:

* Strictly object oriented. Every widget is a class with properties, methods, inheritance, mixins etc. http://goo.gl/AaPb6w

* It makes (or rather you should make) heavy use of data binding. Mostly to link widgets to data stores (being populated/transmitted by some sort of rpc call, the app itself or hardcoded etc), but also to bind widgets to other widgets. http://goo.gl/DxvNWC

* Backend communication comes in several flavors, but most prominent is the JSON-RPC interface. Qooxdoo provides backends for several languages, but it's also rather easy to implement your own. http://goo.gl/IjCT8n

* Communication between objects is done via a simple event chain, i.e. you will use addListener and fireEvent/fireDataEvent a lot. http://goo.gl/Yxfmot

* You can mix your project and have some parts compiled and others loaded from source at runtime. We've been using this to create a metalanguage for forms, which streamlined our application a lot. Basically we don't have to write qooxdoo code anymore to create new forms, but just a xml description of it.

All in all qooxdoo helped us more than it stood in our way. I would use it again.

It's pretty restrictively licensed.

To expand on that, their interpretation of the GPL is that your server code must also be GPL'd if you use Ext.

Switching from LGPL to GPL (ostensibly to clear up some license issues with the graphics) killed their mindshare.

Sencha Touch plays rather well with CoffeeScript

anybody has experience moving from Flex to dojo toolkit ?

Reply on news.ycombinator.com