I started a new job last week working for a local company as a senior CF developer.  The job I signed up for was supposed to be a high-vis project, with multiple applications written in CFMX 6.1/Fusebox 4.1, with plans to move to Fusebox 5.  I was looking forward to this, as I have been interested in taking a peek at Sean Corfield's work in FB5, which has gotten really good reviews. Plus it was supposed to give me an opportunity to work with a small team for the first time in a long...... time.  I was stoked.

However, as sometimes happens, things have stalled. My position is not 'turned on', and I don't mean that in the way you would want it to.  In the interim, I have been assigned to do some work on some other projects that need some attention.  For how long, I do not know.

Oh my heck.  

My worst fears have been realized. I have now inherited 3 projects left by my predecessor, all using different frameworks/versions and no documentation.  He left several weeks prior to my arrival, so there was no formal turnover, although I did get him on the phone for about 30 minutes.

The one I am working on now is a Fusebox 3 application without the Flip and Fusedocs.  The database is Access and there is a printed ER diagram, but no CASE tool, and no descriptions of what the tables are for, or what the fields mean. I know about what the application is supposed to do, but I don't understand the business domain at all, and I was forced to meet with the clients to discuss changes *before* I ever laid eyes on the application.  Yowza.

My next one is one that is 95% complete according to the former developer, and it is written in Model-Glue Unity.  Again, I have a rough idea of what it is supposed to do, but there is no formal documentation.

This company has a lot of CF developers, but they are spread out at client facilities, and there clearly has not been an effort to standardize anything where development is concerned.  Maybe I can help them with that.

On a positive note, I should get some good experience working with several flavors of Fusebox and Model-Glue Unity.  However, since the projects are already way down the road, the clients are starting to complain about late deliverables.

This is going to be painful for a while, I just know it.

Development frameworks are a hot topic in the ColdFusion (or any development) community right now. Many consider them essential to developing maintainable, scalable applications. Other's decry community frameworks, arguing that they are crutches and add unnecessary complexity to solving their business problems.  I won't rehash these arguments (Google to your heart's content), but I will point out that there seems to be enough interest in CF frameworks, that Michael Smith and TeraTech are sponsoring a Frameworks Conference in February 2007. 

One framework that sadly won't be represented at the Frameworks Conference is Plum, and that's a shame, because I believe Plum rocks the house.

ColdFusion was originally a Windows-based, procedural, web development environment. When Macromedia acquired ColdFusion from Allaire and subsequently rebuilt it on the J2EE platform, we began to see the beginnings of the "javafication" of ColdFusion.  ColdFusion Components (CFCs; essentially object classes) were added and object-speak started to become pervasive in the community. This has really picked up steam recently, as many developers with Java backgrounds are moving to ColdFusion, bringing with them their best practices for Enterprise team development, Macbook Pros and Linux backends. We now have a slew of Object Oriented (OO) frameworks like Mach ii, Model-Glue Unity, ColdBox, etc. which have adopted the Model-View-Controller design pattern and use XML configuration files.  Even Fusebox, arguably the most widely adopted CF framework, has been morphing from a procedural framework into a OO hybrid.  Do I see this as a bad thing?  Not at all.  I see it as a necessity for ColdFusion to continue to evolve and be taken seriously in the Enterprise development space, which I believe must succeed for ColdFusion to have any long-term viability.

However, let's remember that CFML is *still* a procedural language, and not everyone wants to be an Enterprise psuedo-java developer. There are scores of smaller shops out there that run exclusively on Windows, and use ColdFusion to provide products and services for their customers, who have absolutely no interest in OOP, MVC, etc.  They are focused on getting their projects out the door as quickly and cost effectively as possible, and the luxury of studying and rewriting  their apps using MVC may not exist for them.

It is possible to write large, performant, scalable web applications with ColdFusion without going the down the java-lite route. If you're not into rolling your own, Plum will help do this for you, and very quickly.  Plum was originally an internal framework built by Adam and David Churvis at Productivity Enhancement.  For those of you who don't know, Adam and David co-authored the ColdFusion MX Bible, along with Charlie Arehart and Hal Helms. Their chapters on advanced database concepts and custom tag development are worth the price of the book alone. They are widely recognized for their database and scalability prowess, in addition to being excellent trainers. They espouse development best practices. Plum represents the best of their development and training philosophies.  And it is available to you as a development and/or learning tool, free of charge.

I have been using Plum as my main CF framework for about 3 years now (since the first alpha), and have built about 14 applications with it in that time. Most are not public facing (I do a lot of intranets), but some of these are outsourced projects for some well known companies including British Gas, Sun  Oil (Sunoco) and Merrill Lynch. I have not once encountered an error from one of these applications that wasn't directly induced by me or was a recognized bug in CF. 

You don't see a lot of traffic on the Plum mailing list nowadays, and I believe one of the reasons is because the framework core files are so bug free that you don't have a lot of problems. Another reason is because not many folks are adopting Plum anymore for various reasons.  Perhaps the foremost is that Adam and David are now focusing on Microsoft .NET technologies, and there are no immediate plans to further the development of the CFML version of Plum. It works fine now and there are only a couple of minor issues.  Other reasons include the fact that there is no support for MySQL or Postgres, and that the code generator is a Windows .NET application.

I plan on frequently blogging on Plum, if for no other reason that those of us in that community can share tips and tricks, show off some sites we have done, etc. There is a lot to talk about. We might even be able to address some of the shortcomings with the Churvises blessing down the road if there is interest.  So if you are a Plum afficionado, please stay tuned.

P.S. On a side note, I have just been hired as a full time developer working with a small team on a long term Fusebox project.  You can expect that I will be devoting some time on this blog comparing and contrasting the two frameworks and methodologies.