This is part 2 of 2 in my coverage of Jakob Neilson's recent Alertbox column entitled "Web 2.0 Can Be Dangerous". Part 1 was called: "There's a Forklift in my Toolbox".

Jakob Neilson spends a lot of time in his Alertbox column, making the point that "rich interfaces" should really be reserved for "actual applications" and not "websites". Very much like the Flash 99% Bad article, I am sure the following pull quote has the potential to create just as much confusion and misconception out there amongst people that are non-technical, but make the big decisions:

"Such interfaces can work well, especially for actual applications that offer true functionality and thus require a full GUI. But if you're just designing a website, the more advanced UIs often confuse users more than they help. Why? Because users engage less with websites than with apps."

This quote drives me nuts, because "actual application" is a very malleable reference. As a matter of fact, I bet most developers out there could spin things so that any project sounds like it is an application. If anything can be an application, then that means we can use rich user interfaces for anything! If the criteria is being sure that the application offers "true functionality" then this is a mess, it's like asking a waiter that depends on tips if "the food is good here?".

Data visualization dashboards, embeddable widgets, toolbars,forms, wikis, maps, social networking, social inventory, audio / video players, shopping carts, and on and on and on. These are just a fraction of things people develop on the web that are often referred to as applications.

What about the term "website" which is used as the opposite of application in this column. In today's day and age, does website really mean the same thing to everyone? I wonder how many people out there in the world that are not web professionals would consider Gmail, Hotmail, Google, or Buzzword a website?, likely, the majority of people.

The term website really means: a destination that I can visit using a web browser. It may be a bit dangerous to use it to describe something that is not an application, since rich internet applications meet all the criteria needed to be considered websites.

Beyond that even, rich internet applications can exist in micro from, manifested as widgets embedded in traditional web content. This is totally confusing, nothing makes sense here!

The reason I say "Whose Coat is this Jacket" is because if everyone uses the term application or website to describe everything, then it really means nothing at all. That gives us no basis on which to form judgment on whether or not deploying rich user interfaces makes sense. Without a hard definition of "true functionality", we are unequipped to create any kind of real distinction between the two terms to try and make sense of it all.

So, here we go. Let's try and make some sense where none is present.

Rich user interfaces have nothing to do with engagement if you consider engagement to be the user making a cognitive decision and then taking physical action. I get the feeling that Jakob is trying to suggest that this is the case when explaining that people engage less with websites. I believe that this is a false statement.

Some of the most valuable capabilities of rich user interfaces involve notification and visualization. These types of interactions are passive for the most part, involve next to no physical action be taken, but require rich interfaces to realize. Video Players, audio players and real-time graphs and charts are examples of applications that fall into this category.

The above point alone, is enough to dismantle Jakob's claim. So let's shift the focus away from him, and on to the real question, which is:

"When is the right time to implement a rich interface?"

The question is not something you can answer with a metric, and apply as a standard across the board to any project based on a calculation. It instead comes from a clear strategic vision and understanding of the mental models you are trying to portray to the user in an interface.

All life forms that have any type of sensory perception use that perception to create a model of their world. That model is what is used to make decisions as well as predict the future. When humans encounter user interfaces these two factors come into play.

Creating a model of their world involves coming to an understanding of what is going on as fast as possible and with as little energy spent. Rich user interfaces have always been used to do this, when it is clear that the "out of the box" mental model of the World Wide Web does not do the job. This may in fact be measurable, but not in terms of the amount of buttons clicked. It could be measured potentially by the amount of time it takes to realize what is going on.

The WWW model is one that it is document centric. It gives the user the feeling they are observing a piece of paper that is linked to other pieces of paper. It is astounding at how dynamic this model has been, and how easy it has been for people to understand.

The reason rich user interface technologies have emerged with such vigor, is because we have come to a time on the web, where the information and processes we are required to model for people, feel as if they are crippled by the traditional document mental model. These technologies free us to explore other mental models that are more appropriate while remaining compatible with the traditional model.

The answer to the question "when to use a rich user interface on the web" is partly found, when it is clear that you are unable to present the end user with the mental model they require to get value out of the content/process quickly ,and with as little energy as possible, using the basic UI components of the web.

This does not mean that you should design an entire web-site in Flash just because you are having trouble providing an appropriate mental model for a piece of sub-content. Widgets are wonderful, and when used appropriately, envigorate the web experience and enhance mental models greatly. Embeddable Google Maps, or YouTube Video players are brilliant examples of micro rich interfaces at work.

When the document centric mental model is clearly the most effective model to use to give your users the understanding they need to get value out of the content quickly, then you should surely use it. Betraying this, is just bad design and strategy, not the fault of the technology.

I mentioned that mental models and rich interfaces are also about predicting the future. This is a key factor because it is really the heart of usability. Usability is measured by how many times that what the user expects to have happen, does or doesn't happen . Good user interfaces present a process in steps where the user can effectively predict what the next step will be. Great user interfaces take this a step further by condensing steps or eliminating steps that are tedious, creating a sense of surprise and relief. Poor user interfaces introduce steps that are unexpected or unwieldy.

Rich internet application technologies allow us to take command of the client-side and it's potential processing power. This power is most effectively utilized to eliminate or condense steps in a process, and that is the second reason why you would choose to use them.

I have come across web applications in the past that have introduced steps into a process simply because the interface was deployed as HTML. This happens, because as we all know, without rich UI, you must refresh a page to invoke change of state. Rich UIs allow us to change state as the user is interacting with the interface, which ultimately means that we can adapt to their every decision without making them wait. The most incredible rich UIs remove the feeling of steps completely. This is very much like where desktops apps are now. You operate directly on the data as it is, instead of in the form of a paged "Wizard".

Think for a moment of Facebook, as it is somewhat of an AJAX champion. Imagine what it would be like to re-arrange the layout of your profile page without the AJAX drag and drop system they have implemented. If it was not present, I would not even know what to expect from a process that would allow me to rearrange my page. Because I cannot fathom an alternative mental model to achieve such a thing, I am unable to predict the future. When I am unable to predict the future, the chances of me getting lost in an interface and abandoning it go way up.

The point I am making with all of this is that deploying a rich interface is very situational, and there is no save the day metric that you can use to calculate whether or not it is a good idea. Like my article previous, I come to the conclusion that great strategists and designers produce great solutions, and technology is not a factor. Technology is merely a line item within the great or poor solution that was conceived.

Whose coat is this jacket? Well, we have surpassed the point of websites and applications being something deemed as separate, they both exist together. There is no one way to do things. Applications can be components of websites,websites can be components of applications, applications can be websites, and websites can be applications.

As I said in my previous article, when it comes to Jakob Neilson, you have to understand who he is writing for. He is not writing for you and me, instead, he is writing for all those people that really don't have the time to dedicate to being up to date with the state of experience design. He tries to simplify things for these people and give them safe parameters to work within. Safe parameters can often mean disregard for the value of experience design. If this mentality is left too long and it reaches the point where these parameters become out of sync with expectation of experience of your users, then you are opening up the doors for being easily surpassed by your competitors. That catch up game is hard to play.

Recently, I have come across some great discussion on why people put up with crappy interfaces and experiences. I am going to cover this in my next article:

"My Greyhounds Enjoy Eating Poo"

(can even say poo on this website?, ummmm application?.... oh whatever!!!! Deal with it!!!)

[Edit] - Updated all the "Who's" to "Whose".

Read more from Tony MacDonell.