The Fusebox conference this year was yet another excellent event hosted by Teratech. One thing that was noticeable was the smaller turn out than last year's event, although this was easily attributable to the massive turnout from CFUNITED 2005 and the close proximity of that event to this one.

The usual suspects were present, in addition to a few new additions such as Isaac Dealey and Jon Rowny. As for the attendees, the congregation was comprised mainly of developers and CFUG managers (which is to be expected for such a code-centric event).

Jeff Peters and Hal Helms kicked off the event with an interesting diversion from your everyday keynote. Rather than covering prepared topics lecture style, they staged a performance of a would-be radio drama. The drama revolved around a psychologist specializing in "framework psychology", and the conversations between the doctor and his patients, who ranged from good ol' Fusebox, to up-and-comer model-glue, to a surfer-dude-esque individual known only as "NMM", or the "No Methodology Methodology." Considering that Jeff has a background in theatre and Hal is more creative than most techies, the keynote was oddly amusing and fitted nicely with the laid-back theme of the conference.

Following the drama, Hal and Jeff had an interesting discussion regarding the emergence and penetration of the various ColdFusion frameworks, as well as the staunch advocates that still exist (although in much smaller numbers) who are not interested in a standardized framework and prefer the ad-hoc approach. Overall, the consensus seemed to be that (a) CF now has plenty of frameworks, and (b) the ones that are out there are becoming mature and flexible enough to warrant a likely slow down in the appearance of new alternatives. If you ask me, that's a good thing.

From my perception, much of the excitement at the conference was centered around model-glue, and rightfully so. Joe Rinehart has done an admirable job striking a firm balance between the somewhat purist OO principles of mach-ii and the procedural simplicity of Fusebox. Joe held two sessions throughout the conference, which allowed everybody to get a chance to check out the framework and ask questions. I attended Joe's session and saw familiar aspects from both mach-ii and Fusebox, not to mention a number of nice features that Joe had decided to bake in (such as the IoC principles from ChiliBeans, another of Joe's creations).

As usual, John Ashenfelter and myself spent most of the conference discussing Java technologies with the masses. John decided to venture even further out of the proverbial box and provide a presentation on Ruby on Rails. It's nice to see Teratech open to non-CF frameworks and languages, since it can't hurt to look outside the CF world occasionally and add principles from elsewhere to the mix.

Throughout the conference, the item that was most prominent in my mind was how rapidly the CF community had adopted OO principles in the last twelve months, and applied those principles to their work. I remember at Fusebox 2004, I was one of a handful of individuals promoting the integration of Fusebox and CFCs (using CFCs as the model layer in an MVC style), and presented a case study on an application we started developing in late 2003 that used this approach. For the most part, the developers at the 2004 conference were mainly the procedural Fusebox type and were widely unfamiliar with a fully independent, pure-OO model (unlike the handful of mach-ii developers present, who were of course much more familiar with this concept).

In contrast, I found myself surrounded by OO discussions at this year's conference. I saw large turnouts to all the OO-based sessions, from the mach-ii BoF session, to Joe's model-glue presentations, and even as far as my own Spring and J2EE/ColdFusion session. When I asked my attendees how many were familiar with patterns such as DAO, and principles like IoC, I found that most of the developers at least understood these patterns, and many had applied them to their work. This was a radical departure from the expectations that I had as I began this conference with my memories of discussions from last year's event.

I guess the bottom line is this: the CF community is as alive, strong, and open as ever. The now-widespread adoption of OO principles in what was primarily a procedural developer base just twelve months ago speaks volumes to the openness and spirit of cooperation present in the ColdFusion community. It's encouraging to see so much support in the conferences from Teratech and the blogs and forums of the CF thought leaders, without whom the adoption of OO would most certainly have been slower and more painful.

More interesting to me still was the interest in applying technology and practices from the Java and J2EE communities to CF development. Items that cropped up in the conference included AOP, IoC containers, service locator-based frameworks, and discussions on ORM for CF. As I stated earlier, the number of attendees at my J2EE-centric presentation was at least three times my expectations, since I had believed only a handful of developers would be interested in such a demonstration. Not to mention, Merrimack saw the light of day as the conference began, allowing CFCs to claim their rights as first-class Java objects.

I'm interested to see how this story plays out over time as more CF developers discover and adopt OO, and frameworks such as model-glue, ColdSpring, and Tartan start to take a stronger hold on the community. I believe that many CF developers will find themselves taking the time to look over the fence, and see what the Java side of this J2EE product (known as ColdFusion) has to offer.