CF Conversation

I listened to this podcast today. This interview highlighted a couple things that i have been into and out of in other blogs, podcasts about the CF community, perception of CF as a light weigh tool, frameworks etc.

CF has come along way, especially with the last couple of releases that make CF much more object like. I can see how CF has grown up over 10 years, from its initial release which really was alot like HTML with its tag based format. It was easy to use and easy to create dynamic content. CF excels at hiding much of the complexity of what it does in the tags. This results in users being able to *just do their work. In the past couple years, i have been working more in the architectural side of things than in the front, so i appreciate the emphasis on CFCs and OO techniques.

There is a perception that many of the CF developers, because of it ease of use, do not have the traditional CS type background and consequently are sloppy programmers who rely too much on frameworks. An interesting point. CF's own efficiency has spawned lots of programmers who otherwise may have never been programmers. Folks who only dealt with static content, layout and design had access to programmatic tools that were once exclusive territory of programmers.

There is the perception among programmers that the within the CF community, there is not enough emphasis or development integrating CF with other more mature back end solutions, like Java or PHP, for example. Andrew Powell is a guy who has been working to bridge that gap. This is a good ADOBE sponsored article around this type of integration. If CF is to continue its acceptance into the *big players sphere, it must play better with other technologies. Matt Woodward posts a helpful article around Java integration.

The discussion also talks about the proliferation of CF frameworks and despite the backing of the CF *elite, are overused and are a crutch to many CF programmers. Is there a perception in the CF community that if you are not using a framework, you should be? Hal Helms and Simon Horwith have this debate about frameworks.

The frameworks include, machii, coldBox, model glue, fusebox, coldSpring. Each of these were mentioned. There was a good discussion of trade offs and what that means.

For me, i started CFing without using any frameworks. In a perfect world, CFers would have a change to program without any *overhead of a framework. This aids in the understanding of variables, scopes, and other application wide settings and how the whole bootstrapping process occurs. The frameworks hide many of these concepts and could create a programmer who does not really get some of the crucial concepts behind the code. This could result in programmers who cannot program outside a given framework. I think the frameworks are very useful in some applications and perhaps a bit overkill in others. The size of the app is a big factor. I like the idea of a framework because it creates predictability in the application. This is important for the future of the application and as new people come into it to support it.

There was also some good discussion around the software life cycle. A good reminder that we should do our heavy lifting before we start programming and that we should not start the development in the database. FLIP was mentioned, since there was discussion of frameworks. I think i heard these numbers:

50% time gathering requirements and prototyping
30% implementing the prototype
20% testing and documenting

Was a mention that Flex has really become a popular prototyping tool.