Rich Internet Applications (RIA) Landscape/Flex January 8, 2008

Courtsey– Jon Rose’s Blog

When it comes to Rich Internet Applications (RIA), I am way behind to some, and way too far out there for others… Oh well!

It is time to completely embrace the RIA model and put to bed the old Request/Response type model of building web applications. My conclusion is simple, no longer do we have to train our customers to be disappointed with the final product when purchasing online software. The tools are finally good enough to give our business stakeholders and users the ‘rich’ applications they have been asking our industry to build for all these years.

Until recently, I have avoided the ‘Rich Internet Applications’/’Web 2.0’ craze, mostly because I saw Ajax as the primary method available to Java developers, like myself, for enhancing web based applications. I am finally ready to get onboard, but I am still uncomfortable with using the Ajax techniques on a large scale.

In this post, I am asserting the underlying assumption that the majority of the online user interface work going forward will be done in a Client Server like model over the Internet (e.g. Ajax and the like). As far as I am concerned, debate around the question of continuing to use JSF and Struts like frameworks is pretty much over, and any further discussion is for someone else’s blog. We should all be excited that we finally have better ways to build online applications. For me, it has been about finding a method that will enhance the user’s experience, but not be a complete pain to me as a developer.

Going forward, my personal platform of choice for building online user interfaces will be Adobe Flex. Adobe’s recent 2.0 release of Flex finally gives developers a robust solution for building ‘real’ applications and delivering them over HTTP through a browser.

 

I think Flex is cool and you can clearly build ‘sexy’ applications using it, but really who cares? I see way better reasons to utilize the Flex 2 framework as a Java developer:

1. Of most importance, Flex provides the ability to dramatically improve the end user’s experience, build complex features with ease, and improve productivity (of both the programmer and end user).
2. Tools:
   1. It builds on the tools Java developers already know.
      1. Eclipse IDE Plug-in.
      2. Ant build tasks for command line and automated builds.
      3. Flex and/or Granite Data Services used for easy Java integration, which deploy to a J2EE Servlet container.

1. Runtime:
   1. Everyone computer already has the Flash Player and users are used to interacting with Flash based content.
      1. This does matter!!! Overtime it will matter less, but it takes time for users to adopt a new runtime, and as we all saw with Applets it can end up being difficult to get a consistent runtime installed to the masses.
      2. Flash movies run inside a browser and allow for integration with other web technologies (e.g. HTML, CSS, Javascript).

1. Framework:
   1. A framework 100% geared toward user interfaces.
   2. ActionScript 3.0 is Object Oriented.
   3. The language’s heavy Java influences are obvious.
2. Seeing is believing… Spend some time looking at some of the sample applications at http://www.flex.org/showcase/.

 

I have been getting a bit of push back when I suggest Flex/Flash as a viable option for building user interfaces to other Java developers. Some of the more common questions/comments are below:

• ‘Doesn’t Flex cost money?’
  • Adobe does charge for the Flex Builder IDE ($599) and Flex Data Services (FDS) deployed beyond a single CPU. However, you can build and deploy an entire Flex application for FREE. The SDK is free and there are Ant Tasks for creating a build. For the applications I am building, I have concluded that the only paid tool I need is the IDE. Granite Data Services (http://www.graniteds.org/) and Open Source alternative to FDS has met my needs to this point. In addition, Flex also provides methods for integrating with Web Services.
• ‘I tried developing applications in Flash before, and it just doesn’t seem geared toward developers.’
  • I share this experience, and had the same apprehension about Flex. The key thing about Flex is that it is a framework geared toward application developers, not designers. Although the applications are deployed as Flash movies, little-to-no knowledge is needed of traditional frame-by-frame like Flash development.
• ‘Wow. Flash, really? That sounds like a really heavyweight method of building applications.’
  • Not to oversimplify, but I have come to view the Flash Player as just another browser when building and deploying Flex applications, except the Flash Player displays consistently across all supported Operating Systems.

 

If you are willing to embrace the Ajax-like model for building web applications there are a number of options to consider in addition to Adobe Flex: Microsoft SilverLight, Open Laszlo, Java FX, and the many Ajax frameworks. Below is a brief review of each of these from my perspective as a Java developer.

 

Ajax

My depth of knowledge on Ajax is limited, as I have avoided using it on any large scale to this point. I do know there are many developers out there with successful deployments of online applications that incorporate significant amounts of Ajax functionality. For me, it just seems to push the standard browser functionality beyond its natural capabilities. Regardless, by its nature, it is always going to be painful for the developer, as they have to support deployment to multiple runtimes and the many cross-platform issues that arise from that.

 

I do believe Ajax, along with HTML, CSS, JavaScript and server side technologies, will always be some part of any comprehensive online application. The only question that still intrigues me is, what parts should they make up?

 

Microsoft SilverLight

I would anticipate SilverLight being the strongest competitor to Flex in the near-and-long-term. As much as I love to hate the Evil Empire, we all know they will commit to making this a world class development platform that integrates with all of their other tools. It appears to already have stronger video support than Flex, and will surely catch Flex in other areas very quickly.

 

My biggest reservation with SilverLight comes strictly as a Java developer. It doesn’t seem to make a lot of sense to learn a new tool set (e.g. Microsoft Visual Studio) and be outside the mainstream trying to integrate Java and SilverLight when there are comparable technologies that are much closer to the Java space. That’s not to say a .NET developer shouldn’t consider Flex, but hey, I don’t know how things work on that side of the software planet.

 

OpenLazslo

OpenLazslo is similar to Flex, as it is an Open Source Framework for building applications that deploy as Flash movies. However, it is my understanding that the current release of the OpenLazslo Framework does NOT target the Flash Player 9 virtual machine (and its many performance benefits). Flash Player 9 was a virtual rewrite with many improvements including support for ActionScript 3.0 and much improved performance (up-to 10 times faster than previous versions according to Adobe). The Flash Player 9 actually ships with two virtual machines for backwards capability (Flash 8 & 9).

 

Regardless of the performance trade-offs, it is still attractive to some because it is an Open Source project. However, Abode committed to Open Sourcing the Flex framework in recent weeks. So, I am not sure that is still a compelling motivation to choice Lazslo over Flex.

 

Java FX

Java FX is a complete joke to me! It’s just another grand Java One announcement with NO release date… and foggy details on what it will look like when it is released. The most promising descriptions describe an Applet like deployment (over the samples that run in Java Web Start outside the browser), which just isn’t going to cut it. Thanks Sun… Thanks for showing us the raw work going on in your R&D labs (more Jon Rose skepticism run amok).

 

The only interesting thing about Java FX is that it is essentially Sun’s acknowledgement that the industry has already shifted away from server side user interface frameworks (e.g. Struts, JSF), in favor of richer client side runtimes.

 

Summary

Each of us comes to our own conclusions on how to best build online applications. For me, Flex is clearly the way to go. As a few wise developers have said to me, ‘Flex is just another tool.’ To a large extent I agree, but the difference to me is that the end product produced from Flex is such a large leap forward when compared with the current landscape.

 

 

References:

http://www.adobe.com/aboutadobe/pressroom/pressreleases/pdfs/200606/062806Flash9.pdf

http://msdn2.microsoft.com/en-us/library/bb428859.aspx

http://www.openlaszlo.org/front_page

http://www.jamesward.org