Group 1 - Promoting CFML to New Developers

Team Consisted of:

• John Farrar
• Rick Mason
• Dana Tierney
• Ezra Parker
• Jeff Coughlin
• Brian Swartzfager

You can ready the summary at: http://docs.google.com/Doc?id=dcdhg6m6_27dnzqx5gs

Group 2 - Promoting CFML to PHP Developers

Team Consisted of:

• Adam Tuttle
• Timothy Farrar
• Seth Bienek
• Dee Sadler
• Ken Auenson
• Chris Hough
• Adam Howitt - adamhowitt@gmail.com

You can read the summary at: http://docs.google.com/Doc?id=dcdhg6m6_26frtdsbv8

There is also all of the raw ideas that came out of the session. Expect blog entries from the teams which show those in more detail.

I am revisiting that topic today, in a similar format to last time, except we should have more of the materials we need to make this brainstorming session a success, AND, even more importantly, MORE TIME.

This is an all audience participation brainstorming session. We're going to figure out what our community needs to do to grow CFML outside our own community. If you are at CFUNITED, I hope you'll plan to attend.

We need "recorders" for each of the breakout groups in the session. If you are planning to attend, and are willing to just "record" the details of your breakout group, please see me before the session (as early as you can today), or email me at brianmeloche at gmail dot com. I'll send you a link to a Google Doc to record your group's findings.

People seemed to enjoy the presentation. The presentation deals with using ColdFusion to interact with an ERP system. The problem is complex enough that sample code is not really possible without a lot of work, so this presentation did not have code. I didn't get a lot of complaints about that because the session was detailed enough that it didn't NEED code.

One thing that I hit on during the presentation that has more of a wider scope and was also covered by Nic Tunney in his preso. My idea is simple: You really shouldn't try to make ColdFusion do things it's not good at doing.

ColdFusion bogs down when it needs to create a lot of objects. If you are retrieving a LARGE result set for read only purposes, you shouldn't create an array of objects. If a query will do, use a query. If not, use an array of structures. I will often add a getWidgetsAsQuery or getWidgetsAsArrayOfStructures service method and create a transfer object (not to be confused with Mark Mandel's Transfer) to give me the ability to create an array of structs instead of creating an array of objects. When you are dealing with a class with many methods and/or a large array, you will get a big performance hit, even in ColdFusion 8. On add/edit/deletes, a bean is well suited, but not on gateway style selects. You can try to get around that by going around it and doing your model in Java, but, to me, that doesn't seem like a great idea unless you NEED to (sorry, Andy P!), as doing that gives up what ColdFusion is best at in the first place - Rapid Application Development. Instead, work to ColdFusion's strengths would be my suggestion. I may be criticized by some as not being a "thought leader", but that's a crock of sh*t as far as I am concerned. As my slide says, I am not a Spock when it comes to being a CF developer. I'm a Scotty, and I'll take real world experience (using sound fundamentals) over theoretical constructs every day. During pure OO in ColdFusion is just that - a theoretical construct - and it fails in real world situations. Use OO where it makes sense, which is in MOST places, but don't try to use it everywhere in every situation.

First of all, my topic has nothing to do with Model-Glue, nor does it talk about any framework.

Adobe markets ColdFusion a lot, maybe even mostly, towards intranet development, but that sells ColdFusion short. My talk is about using ColdFusion in large scale, public facing enterprise applications, and specifically, interacting with enterprise systems, such as an ERP, and what you'll need to know should you need to develop an application that needs to interact with such a system. If you don't know what an ERP is, check out this Wikipedia entry. My topic will be at a 10,000 foot level, but I'll talk about some specific lessons learned from the trenches.

I hope if you're at cf.Objective, you'll decide to come to my presentation. I am up against sessions by Mark Mandel, Maxim Porges and Charlie Arehart, so I have tough competition!

The conference schedule for cf.Objective() 2008 has been posted.

The cf.Objective early bird ends on January 15th so register by the 15th to get the special price of $499! Register after the early bird, you'll pay full price ($629, which is still quite reasonable).