We’ve just started using Sencha’s ExtJS 4 in our project and have been following the documentation to build an MVC architecture on the client. As time progresses, I’m inclined to think that it’s more trouble than it’s worth, because:
- We render to different views (ExtJS being one), so we keep all the logic in controllers on the server to minimize duplication
- Because of this, there isn’t much to unit test since there is very little logic. We do functional tests with Selenium
What are the benefits to keeping an MVC architecture on such a thin client? Any that aren’t evident now but might be down the road when we start developing components for it?
In my experience with Ext JS 4, using MVC is in general more trouble than it’s worth.
If you’re building an entire site in Ext JS then you might get some benefit out of it, but nothing that you couldn’t get from creating/overriding components. If you’re developing a thin client with little to no client-side logic then you’ll end up introducing more overhead.
The main problem is when you start needing the same component on many different pages (such as a combobox with a list of employees). You’ll probably want different listeners on it depending on what page it’s on, but for the most part it’ll be the same everywhere. It’s simple enough to create a Store that loads data on initialization, then define a subclass of whatever component you need that has the default values set. Then all you need to do is give it the storeId and whatever listeners you want, and there you go.