i know, everybody hates Singletons, and i do, too. But i find it an entertaining exercise to come up with new ways of circumventing the lack of private constructors for Singletons. Today this new one came to my mind:

First, we define an Interface, which specifies, how our Singleton should look like:

IMySingelton.as:

package {
	public interface IMySingleton {
		function doodle(something:String):void;
	}
}

Ok, next we define our Singleton.

myInstance.as:

package {
	public var myInstance:IMySingleton = new MySingleton();
}

class MySingleton implements IMySingleton {
	public function MySingleton() {
		trace("MySingleton is now there!");
	}
	public function doodle(something:String):void {
		trace("yippee: " + something);
	}
}

Very simple, isn’t it? We could optionally also go without the Interface, but then we loose code completion in our editors, because the class MySingleton will be ignored by most editors, although it would compile and run just fine. I do agree, that this is not really exactly the idea of a Singleton, because it has no getInstance() method, but in the end, the result is the same, we only have one global instance of MySingleton.

And since the variable is already global, we don’t even have to declare it in our class, we can directly use it:

Main.as:

package {
	import flash.display.Sprite;

	public class Main extends Sprite {
		public function Main() {
			myInstance.doodle("loodle");
		}
	}
}

now, that this flash / iphone thing is closed, Adobe, can you please make sure, that Flash Player 10.1 on Android will be nothing else but mind-blowing? You know, not like “it works OK”, but more like “wtf, is this flash? you’re kidding. this is flash? Can’t believe it, this is sooo fast!”. It doesn’t matter matter, if it takes another month, but please get this right.

i just saw the post from Colin Moock, where he was kind enough to post the different links to buy the new Flash Builder 4 in the US, UK, Germany and so on. And what do i see, Germany and the US share almost the same price, actually the german price is even a bit lower. That’s good. Let’s see, what happens, when Flash CS5 comes along. Would be cool if we could see a trend there.

just read this article from Ryan Stewart (Evangelist for Adobe). He is complaining about the hype, that is generated around Google Wave and the fact, that he thinks, Wave could have been built with Flash/Flex years ago. And why Flash/Flex apps haven’t managed in the past to generate such momentum. Also read the comments, as the discussion there is a great read. I made my own thoughts about it.

One thing is, in most cases, where new stuff on the internet is getting hyped, it is about concepts, not technology. Only if the technology supports an innovative concept, then also the technology participates from the hype. So with Google Wave, sure, perhaps this could have been built with Flash, but it wasn’t, because no one had the idea and/or the will and/or possibilities to build it. When Flash had its hype years ago, it was because of vector graphics and animations running in the browser. It wasn’t the technology, it was the concept of having the possibility to show rich graphics and animations on the Internet and build them easily. Now Flash is an established technology, so it won’t see another hype itself so easily. Guess, we have to live with it: Flash isn’t cool, innovative apps are cool, no matter, what the technology behind it is (that’s not to say, Flash wouldn’t be a good starting point to build great apps).

In the comments, Wave is compared to Aviary in terms of “hype-potential”. Well, i don’t think, you can compare these two apps. Google Wave is a tool, that is targeted to mostly every Internet user. Aviary is a tool for a very sharply defined user group. So, if Google Wave was featured on CNN, you cannot expect Aviary to have that, simply because of the target group.

Also, a very important point is, that Google simply has the monetary power to build platforms like Wave. It is the same thing with Google Maps or Google Earth and so on. You need a heck of a lot of money and resources to build and maintain these applications. That limits the number of companies / teams, that can build them.

My last point would be, that Flash in terms of a technical platform is still pretty young. It is hard to make the cut, but i would consider the introduction of ActionScript3 as a point in time, where Flash seriously started to turn into a technical platform, where it became interesting for classical developers to work with it. And still, considering the developer tools, that we have for flash compared to stuff, that Java- or C#-developers have, there is a long way to go. If i am not mistaken, a major part of Google Wave was written in Java and then ported with their Web-kit to JavaScript, so they probably could use those nice tools that exist for java development.

So, what does that tell us now? Actually nothing new, because it’s always been that way, that great ideas at the right time, that come together with enough money and resources have a big chance to excite the crowd. Let’s see, how Adobe will improve the flash platform not primarily in terms of featureset in the player, but more in terms of developer tools (some fresh ideas for the Flash IDE perhaps? We need to get those designers back on board).

it’s been quiet here, lately. I was working hard on my upcoming book Flash Engineering (it’s in german, sorry for the english speaking people). Now i’m almost done. The script is currently in the editorial office and being checked. Can’t wait to see it in reality. By the way, you can already pre-order it:

The application in question is one, that has a very rich user interface with animations, video, sophisticated controls and so on. The domain model that lies behind all that is rather simple though. It is one of those applications in the advertising industry, that is more about the show than about utility.

I would roughly estimate, that 60% of the code that my colleagues will write is related to the UI and the rest goes for the model and the general application logic. Actually most of the applications we develop are partitioned that way. That’s nothing to get excited about, after all we’re building thin clients and of course a thin client is mostly made up of ui code.

But knowing that, i come to ask myself, why we are actually so excited about MVC, Domain Driven Design, architectrual frameworks, IoC and stuff. For the particular type of flash applications that we build, we should be more concerned about how to structure the V in MVC than about the whole MVC in general. That’s not to say, we should not care about general architectural matters, but being a flash developer, who is building rich user interfaces, we should rather think about structuring our views than about structuring our model, since in many cases, our model is not the core of our problem, but the view is.

Screenshot of the eco zoo campaign.

However, i do see too little articles covering this topic in particular. I also do not see enough design frameworks trying to deliver support in that area. Of course, we have frameworks, that aim at simplifying building UIs, such as the Flex framework. But for me, these frameworks rather try to look at the UI to be something, that should be made easy, so you don’t have to think about it to deeply, so that you can concentrate on the application itself. But what, if the presentation is the thing, you want to concentrate on? That’s why i’m so sceptical about tools like Adobe Catalyst, because i have the feeling, this tool tries to hide away the problem of developing complex UIs, suggesting that the act of architecting and developing a rich UI is something, that can be somehow automated and hidden behind a clickable tool. I strongly believe, this is not possible. It takes experienced UI developers to build stunning presentation layers and these experienced developers should start thinking about effective ways of building those UIs.

Screenshot of the ikea wardrobe solutions campaign.

I think it is time, that we think more about the underlying patterns in the development of rich and animated UIs. A discussion about those should not only be about components, because really rich UIs are not solely made out of classic components. Of course, in a sense you can define everything to be some sort of button, but i feel that doesn’t live up with what we really need. Especially as we are heading to UIs, that are try to go away from classic form style in favor of more integrated, chromeless UIs. Also we need to incorporate timing in a more fluent way. We already have some very good state patterns yet, but my feeling is, that todays UIs go beyond simple state, our concern is not so much the state, but the transitions. And sometimes UIs are constantly transitioning, so that one might start to doubt, if it is even useful to talk about states anymore.

To make a long story short, i think with the advent of rich, interactive and animated UIs we need a deeper discussion about how we can build those in an effective way, that allows for the basic principles like extendability, maintainability and reusability.