I had an excellent discussion with my coworker Susan today about refining certain internal processes, and one of the tangents of the conversation went off on what the actual definition of a Rich Internet Application actually was. As we know, anything from a banner ad to a product configurator can be considered an RIA, and the only common element seemed to be that an RIA retains its functionality within the context of what the user is interacting with. In other words, if you click on button in an RIA, the resulting action does not significantly change the page or window the user is interacting with; Clicking to go to a new page loses context, using an animated accordion to display different content does not.
The similarities, though, end there. Implementation varies, technology varies, scope and location and functionality varies, and all in all it ends up being a pretty difficult convoluted mess to describe. At best you can group them via complexity, and after a brief exercise of that nature we realized that a new breed of networked application was emerging. Well, alright, perhaps not emerging, but instead gaining momentum and acceptance in the mainstream. Here’s the scale, see if you agree with our reasoning.
Level 1: The Widget
At this level of RIA you are attempting to display information in context of a particular page, however you don’t care about anything but the most basic user interaction. These could be things like drop-down menus, product detail pop ups, rotation views, buttons that reveal and/or expand text content (like reviews) and so forth. They are almost always implemented in JavaScript, because to use Flash or another plug in technology would be quite a bit of overkill.
Level 2: The Functional/Interactive Widget
This level of RIA’s describes widgets that allow a user to complete a particular functional task. No longer content with simply displaying information, we’ve now added functionality or a experience that responds to user input. This could be as simple as a DHTML login form or as complex as a Flash-based page takeover, but it necessarily remains restricted to a specific, easily definable task. "Log In", "Rate This Product", "Check Convention Schedule" and so forth are good examples, as they add a richer experience that remains in context with the page itself.
Level 3: The Rich Internet Application (RIA)
The next level of complexity takes the task mentioned above and strings them together into a flow, or objective, thus defining an actual application. While previously you would have perhaps a few simple form fields to fill out, an actual RIA causes the context of the page to change dramatically via user input. User interaction is no longer restricted to a single action, but instead is intended to enable an activity, such as "Tracking your time", "Editing a Photo", or "Managing a Color Palette". This is where the bread and butter of RIA’s exist, as well as the holy grail of Web 2.0: A fully interactive and functional application contained entirely within one browser page.
Level 4: The Rich Networked Application (RNA)
The Rich Networked Application (RNA, I’m trying to coin a term here, help me out) ceases to be bound by the browser, and instead has become an experience that bridges and is uniform across all digital touch points. The service is available not only from a browser, but may also be accessed from a desktop, a mobile device, a vehicle dashboard, a kiosk, a gaming console, or any other networked or partially networked device you can imagine. The RNA reaches out to many delivery channels, and while it may provide a different experience for each it nevertheless remains connected in context across them all. Excellent examples of this are Twitter (and all its clients), Google Maps (available on Mobile, Internet, etc), Kuler (Integrated to the desktop and the entire Adobe Suite), as well as upcoming games like Spore (Share creatures across platforms). Implementation… well, lets be honest, it’s a nightmare if you go into it unarmed. You have to support many different platforms, frameworks, systems and limitations, yet even so we’re starting to see toolsets emerge that address them all (Most notably Adobe’s Flex & AIR, Microsoft’s DLR via WMF & Silverlight, and Javascript libraries like SproutCore, MooTools and Prototype).
Did that make sense to you? It does to me, and I’m really excited to see how what we have today is going to start bridging the Device Divide.