By definition, a hybrid application is simply an application that can be installed on a device the same way as a native application but contains what is called a webview that renders HTML code either via a web request or via static HTML that is loaded as a resource in the application. This definition covers a variety of different types of applications that could be hybrid apps. Because of this, it makes more sense to breakout hybrid apps into three different categories; Thin Hybrid, Full Web Hybrid, and Full Device Hybrid.
Full Web Hybrid:
A full web hybrid application is an application that can be installed as a native application. The entire native application is a webview. The webview is a view that can render HTML code the same way the device’s browser can. The webview will then point to an HTML5 web application that resides on a web server. The benefits of a full web hybrid application is that once installed, it is possible to update the code base in one place, the web server where the application resides, and instantly, all of the users will have the updated software. The downside to this type of application is much the same as a web application in that speed is a big issue. The users device has to download the entire interface and data each time it needs to render anything.
Full Device Hybrid:
As you can imagine, there is a trade off for each type of application. These trade offs have been listed in these three blog posts. I hope this clears up any speculation you may have had on the subject. If you would like a more comprehensive look at these different technologies including examples and how they affect the business check out a white paper I have on the subject.
I look forward to your comments and questions!