An Architect's View

I have to say that today went much more smoothly in terms of my session plan than yesterday. I started the day with Joe Rinehart's "MVC for you and me" session in which Joe explained MVC and several other patterns (MVC is a composite of Strategy, Observer and Composite) and then walked us through his "Pig-Latin" Model-Glue example and his, now famous, "Blog-in-nine-minutes" demo. He covered a lot of ground with a lot of clarity and no fluff. Even as a seasoned MVC application developer, I felt I got something out of this - confirmation of my understanding of MVC and more insight into the impact it actually has on my style of development. I particularly liked his comment that if Elvis was a Design Pattern, he'd be MVC because it's The King! I took a break to work on my duck typing presentation (which has proved to be the bane of my life, to be honest - Hal you owe me a large beer for this one, buddy!), coming back down for lunch and then a solid afternoon of sessions. First up after lunch, Simon standing in for Hal (several of us got to share the pain of Hal's project deadline) doing "Practical Design Patterns". Simon chose to use a module from his one day domain modeling course, going through about ten design patterns and a handful of anti-patterns (common mistakes). It was fairly comprehensive but it was a bit dry and a bit overwhelming - I would have preferred to see fewer patterns explained in more depth, with some code examples. Next was Dave Ross on "Inversion of Control and ColdFusion: Using ColdSpring". If anyone ever tells you that CFUNITED sessions don't cater for advanced developers, point them at this talk and prove them wrong! Dave covered a phenomenal amount of material, starting with an explanation of what ColdSpring is and what problem it is intended to solve (rigid models and coupling) and then moving into the techniques it applies (injection) accompanied by some examples. Then he talked about typical usage scenarios for ColdSpring and how it fits into a tiered application architecture. Finally, he covered some of the additional features of the framework: autowiring (how ColdSpring can figure out dependencies just by looking at method names), factories (ColdSpring can serve up beans from other factory objects as if they were just regular ColdSpring beans) and AOP. He said this could be a full day course in an of itself but somehow managed to condense all of the key concepts and terminology and a full demo into the last 10-15 minutes of his talk. I was exhausted at the end of it! Amazing presentation! Then it was my turn, with "Objects & Persistence". Since Joe had already shown Model-Glue: Unity, many people had a fair idea of what Reactor can do which made my job both easier and harder. Easier because some of the concepts were already at least partially familiar. Harder because, given the basic exposure, the questions from the audience were much more detailed than when I have given this talk in the past! At that point, it was back to writing the duck typing presentation. I swear I've written over a 100 slides for this talk by now! I took a break at 8pm, hoping to find food at the "ColdFusion Celebration" event but it was all gone so I ate alone in the restaurant. This proved to be a good choice with an incredible walnut-crusted seabass with wild herb risotto! A brief mingle and then back to my presentation. By about 11:30pm, I was finally happy with the 40 slides remaining and decided to treat myself to a few beers. Not surprisingly, the bar was still in full swing with many heated discussions in progress and a lot of happy CFers. Doug Hughes had popped in to hang out with folks - great to see him! - although he couldn't attend the conference itself. I spent a fair bit of time chatting to Dion and Adam from Cincinnati - enjoyed the company, thanx guys! Folks began to retreat to the patio with more beer once the bar closed so I decided to come back and write some code examples for my preso and then write up the day. It's 3am. I won't make the first session. I'm on at 9:45am.