The Problem with Google Wave: User Experience

Google Wave is no more. For those who remember the sound of the jaws of the tech mainstream dropping when Google showed the demo at Google I/O, that may come as a shock. For those who tried to use it, it’s probably less of a shock. I kind of liked ReadWriteWeb’s take:

Why did Wave fail? Maybe because if you don’t call it an “email-killer” (and you shouldn’t) then you’d have to call it a “product, platform and protocol for distributed, real time, app-augmented collaboration.” That’s daunting and proved accessible to too few people.

To say that people don’t get collaboration or that Wave was ahead of its time is a cop out. Wave IS an awesome product. Real-time collaboration IS changing how the world works together. On the Flash side that is one of the reason I’m so excited about Collaboration Services; real-time collaboration is fantastic.

But this was a case of Google’s user experience coming to bite them. Some people love the minimalistic experience of Gmail. And it worked a few years ago when it was first introduced, but the iPhone has shown how critical a great user experience is to user adoption. And frankly, Google’s user experience hasn’t changed much since the Gmail days and the applications are starting to feel dated. That’s not a big issue when you’re doing something as straight forward as email, but when you’re trying to completely change how people communicate, you need to provide a user experience that abstracts the technology and just makes it easy. Have we seen that done before?


Exactly. Wave was a great technology showcase but it was not a great product. Google had the chance to fundamentally change communication on the web but they didn’t have the design chops to put it in a package that was useful to people and instantly easy for them to dive into. You can’t do an 80 minute demo for something that’s this big of a shift in thinking.

Hopefully Google takes this to heart and realizes that technology isn’t good enough. When you’re being revolutionary you have to design a user experience that makes the technology feel second nature.

New Project: Flex Collaboration Library

I’ve been digging deeply into the dark crevices of Adobe Flash Collaboration Services (AFCS) my session on creating collaborative components at MAX this week. One of my examples is a collaborative map so I relied heavily on the Yahoo Maps example that ships with the AFCS SDK. One of the things I like about the Yahoo Maps example is that it includes a class that extends ArrayCollection and includes some hooks for collaboration. It’s essentially a collaborative ArrayCollection and I rewrote parts of it for my own example and have found it to be very useful across a lot of my projects.

I realized that it could lower the barrier to entry on AFCS if some of the Flex classes that people rely on a lot had “automatic” support for collaboration via AFCS. So I took the class I created based on the Yahoo Maps example and tried to make it as “drag-and-droppable” as can be for someone who wants to start using AFCS. It merges two concepts, the ArrayCollection concept, and the CollectionNode concept from AFCS. CollectionNodes let you store any piece of data on the server and lets multiple people add/change/remove information from the collection. Events are fired so that any time someone changes a piece of data it can be updated across all of the connected clients.

What I’ve done with my SharedArrayCollection class is put all of that logic into a single class that extends ArrayCollection. Developers can use the SharedArrayCollection just like they would use an ArrayCollection but the difference is that the SharedArrayCollection is automatically enabled for multi-user collaboration with AFCS. The only major difference is that instead of listening for a collectionChange Event you listen for a CollectionNodeEvent.

I’m still hacking out the basics and after MAX I’ll try to provide an example use case so people can see exactly how it works. I’ve put everything up on GitHub under the Flex Collaboration Library project. I’m hoping to make the SharedArrayCollection more bullet proof and then create more AFCS-enabled Flex classes and components. Let me know if you find a bug or if you find this at all useful.

The Web Way vs the Wave Way vs the Flash Collaboration Services Way

afcs_logoAnil Dash has a post worth reading up on his blog that talks about the “Web Way” versus the “Wave Way” and why Google Wave won’t succeed because it doesn’t really fit the same pattern of successful web technologies. He makes 4 general points to define the “Web Way”:

  • Upgrades to the web are incremental. Instead of requiring a complete overhaul of your technical infrastructure, or radical changes to existing behaviors, the web tech that wins is usually the sort of thing that can be adopted piecemeal, integrated as needed or as a normal part of updating one’s websites or applications.
  • Understanding new tech needs to be a weekend-sized problem. For a lot of web developers, long before they start integrating a new protocol or platform into their work, they hack together a rough demo over a long weekend to make sure they truly grasp how it works. And a weekend-scale implementation on a personal site usually translates roughly into a 90-day implementation cycle in a business context, which is a reasonably approachable project size. (In tech, three days in personal effort often translates to three months of corporate effort.)
  • There has to be value before everybody has upgraded. This is basically a corollary to Metcalfe’s Law. While we know networks increase in value as they add more nodes, the nature of web tech is that, in order to be worthwhile, it has to provide value even if the people on the other end haven’t upgraded their software or web browsers or clients or servers. Otherwise you’re shouting into an empty room.
  • You have to be able to understand and explain it. Duh.

The entire post goes on to explain details of where Wave fits and where Wave fails. But as I read it I couldn’t help see it as a ringing endorsement of Flash and especially Adobe Flash Collaboration Services (AFCS). Before I dive in, I understand that Anil’s “The Web Way” has an inherent requirement that everything be “open”. Flash and Flash Collaboration Services probably won’t fit in most people’s definition of “open” as it relates to the web. In this case, I think that’s part of the benefit. One thing Anil does is looks at Wave from the developer perspective and he provides a list of technologies required to use Google Wave and add real-time collaboration to your web application:

  • Federation (XMPP)
  • The robot protocol (JSONRPC)
  • The gadget API (OpenSocial)
  • The wave embed API (Javascript)
  • The client-server protocol (As defined by GWT)

That’s a lot of stuff for a developer to know and understand if they want to start building something that interoperates with and leverages the technology behind Google Wave. Now think about a Flash developer who wants to add real-time collaboration to their web application. They’ve got no real new protocols to learn (RTMP behind the scenes but not necessarily exposed in such a way that developers need to understand it), no new languages to learn, no new client-server protocol, it’s just ActionScript and(/or) Flex, and some new APIs. Then your application is real-time enabled. So lets look at the four “Web Ways” and see how they apply to AFCS.

  • Upgrades to the web are incremental. With the pods and APIs for AFCS, it’s pretty damn easy to just integrate it with your current application. There is no rewriting from scratch and you can literally just add an AFCS component and enable collaboration for your application. As you dig deeper, the service gets more complex and you can do more with it, but to start, it’s dead simple.
  • Understanding new tech needs to be a weekend-sized problem. If you’re a Flash developer already, all you’re learning are a few new APIs. You still have to understand the fundamental issues behind real-time collaboration if you want to create complex components, but you’ve got the core development skills to create those applications so you can focus on learning the theoretical stuff and not the code stuff.
  • There has to be value before everybody has upgraded. This is my favorite, because it’s one of the benefits of Flash. AFCS has 2 versions, a Flash Player 9 version and a Flash Player 10 that adds some more audio support. If you’re targeting Flash Player 9 then 98.8% of the web can see your application and with Flash Player 10 it’s 86.7%. No one has to upgrade anything to see your new real-time enabled application.
  • You have to be able to understand and explain it. With AFCS you can easily add real-time collaboration features like video chat, whiteboarding, and shared data into your Flash-based application. I think that works.

Now again, I understand that openness is a pretty core part of what the web is. But there has always been a trade off between openness and innovation when it comes to the web. And even in cases where “open” can be innovative, like with Google Wave, everyone else has to catch up. With AFCS, even though it may not fit with the wider definition of the “Web Way” you can take advantage of the cutting edge technology that everyone is excited about and ensure that it’s 1) easy to build and 2) easy for your customers and users to view.

Thanks to Sachin for tweeting the link to Anil’s post. I didn’t take the time to read it until I saw he did.

Week of AFCS Revisited: SharedObject

After my trainwreck post on using CollectionNode in AFCS Nigel suggested that I try to recreate my example using the higher level APIs instead of trying to build everything from scratch with NodeCollection. As I mentioned in the previous post, the AFCS team did a great job of creating very basic APIs like CollectionNode and then building on top of those to make APIs that are easier to use. SharedObject is a perfect example of that.

Here’s what I wanted to accomplish: Create a basic multiplayer game where players try to match colors with each other using a ColorPicker. Using the SharedObject class it’s really pretty easy. The first thing I did was set up a room where guests are automatically promoted and no username/password is required. Then I used a combination of AFCS components and Flex components to create a user interface for entering a display name, picking the color, and showing how many players there are:

<rtc:AdobeHSAuthenticator id="auth" />
<rtc:ConnectSessionContainer id="cSession" 
     <mx:Panel id="panelColor" title="Pick a Color!"
          horizontalCenter="0" verticalCenter="0" alpha=".3" 
          width="200" height="200" enabled="false"
          <mx:ColorPicker id="color" change="color_changeHandler(event)" width="100%" height="100%" />
<mx:Label text="Playing with {cSession.userManager.userCollection.length - 1} other people" />
<mx:TitleWindow id="tw" horizontalCenter="0" verticalCenter="0"
          <mx:TextInput id="username" width="150" text="Enter A Username" focusIn="{username.text = ''}" />
<mx:Button id="btnLogin" label="Login" click="btnLogin_clickHandler(event)" />

Nothing major there, notice we’re using the userManager class to get the number of other users in the room (and that it’s bindable). The big thing is the synchronizationChange event handler. In that function we’re going to configure our node and create the SharedObject.

public var sharedColor:com.adobe.rtc.sharedModel.SharedObject;
protected function cSession_synchronizationChangeHandler(event:SessionEvent):void
if( event.type == SessionEvent.SYNCHRONIZATION_CHANGE )
var config:NodeConfiguration = new NodeConfiguration();
config.userDependentItems = true;
sharedColor = new com.adobe.rtc.sharedModel.SharedObject();
sharedColor.sharedID = "color";
sharedColor.addEventListener(SharedObjectEvent.PROPERTY_ADD, onPropertyChange);
sharedColor.addEventListener(SharedObjectEvent.PROPERTY_CHANGE, onPropertyChange);
sharedColor.addEventListener(SharedObjectEvent.PROPERTY_REMOVE, onPropertyRemove);

We have to use the full namespace when we reference the SharedObject in AFCS because it conflicts with the normal Flash SharedObject. We first do a check go make sure we’re in a SYNCHRONIZATION_CHANGE event and then start setting things up. The only change I make to the default NodeConfiguration is to make it so that when a user leaves the room, they take their items with them so we don’t have “ghost” matches. That’s done by setting the userDependentItems property to false.

Next we create our SharedObject. I give it a sharedID of “color” and then make sure the application is subscribed to the SharedObject so that it sees any changes. Finally I set up event handlers for whenever a property is changed, added and removed. The add and change properties use the same function, onPropertyChange.

Let’s first take a look at how we create the SharedObject; something that happens whenever we change our color selection.

protected function color_changeHandler(event:ColorPickerEvent):void

We use the setProperty() method on the SharedObject to give it information. A SharedObject is just a name/value pair, or in this case, a propertyName/value pair. I set the propertyName to the userID so we can know where the color came from and then set the value to our selected color. Whenever we set that property it will create a property change event and call our event handler.

protected function onPropertyChange(event:SharedObjectEvent):void
if( event.propertyName != cSession.userManager.myUserID)
if(event.value == color.selectedColor)
var user:UserDescriptor = cSession.userManager.getUserDescriptor(event.propertyName);
sharedColor.removeProperty(cSession.userManager.myUserID);"You matched with " + user.displayName + "!");

This is where the game starts to happen. The way AFCS works is that when you change a something locally inside of your application it is sent to the server and then you get a change event when it comes back. This is so you can be sure the change was successful. In our game we want to make sure that we’re not dealing with the event we just sent so I use the propertyName property and compare it to myUserID to see if it’s an event that the user sent him or herself. If it isn’t, we look for a match using the value property and our selectedColor. If we have a match I use the UserManager class to get the display name of the matched user and pop up an alert box to show that we found a match.

The other thing I do is to call the removeProperty() method the user’s SharedObject. I ran into an issue where when a user selected a color that matched another one, the other user would get the popup but there was no easy way to notify the user who selected the color that it was a match. I solved this by removing the property which would then trigger a PROPERTY_REMOVE event and fire our event handler.

protected function onPropertyRemove(event:SharedObjectEvent):void
var user:UserDescriptor = cSession.userManager.getUserDescriptor(event.propertyName);
if( event.propertyName != cSession.userManager.myUserID && user != null)
{"You matched with " + user.displayName + "!");

It looks similar to our change event handler. I first get the user information for the popup and then I check a couple of things. First, I make sure that the event isn’t coming from the current user, but the user I have the match for. I also want to make sure that the SharedObject wasn’t removed by someone just logging out (remember our NodeConfiguration settings). If my user is null then that means the user is gone and the event fired because the user left the room and not because I specifically removed it.

That’s pretty much all there is to it. You can grab the project here or I’ve embedded the game below (and here’s a direct link), so feel free to play and see if you can match colors!