Where to start? It all happened so fast. In the blink of an eye the MAX BOF Session on CFML Language Development came and went. But what a great session.
First off, thanks Jason Delmore (Adobe), Adam Haskell and Matt Woodward (OpenBD), Gert Franz (Railo -who wins the pulling an all nighter to participate award) and Sean Corfield (CFML Advisory committee -graciously agreeing to speak).

And of course, everyone who participated in person and online.

About half the meeting got recorded on Acrobat Connect right here

Here is a recap of the first half from my recollection. I would invite speakers and participants to add in their recollections as well.

For some context, the session followed a Meet the CF Team BOF session. Most people were finishing grabbing pizza and beer from the hallway (the day at MAX starts around 7.30 am). We had a few technical challenges. Then got started.

The agenda was

• let each of the three CFML platforms state their vision of CFML.
• Get feedback from the audience on where they wanted CFML to go.
• Discussion to ensue.

Gert opened up with the concepts that

1. the core of CFML should be free,
2. the core should run equally well on all three platforms,
3. CFML platform vendors can add particularities to their flavor of CFML,
4. There should be a way for other vendors to add on extensions to the application.

Railo has begun putting this approach in action by

• being on the CFML Advisory Committee
• adding in Railo specific tags and features (Transfer, cfvideo, enclosures)
• building an extensions manager where 3rd party vendors can publish language extensions and even applications that can be added to the Railo CFML. 

Adam seemed to be generally on the same page with 

• the importance of having CFML be open source and
• allowing the community to add and extend language -both to the core and OpenBD specific implementation.

Jason talked about CFML’s long and rich history. That the mantle of rapid application development was taken from CF and that he is dead-set on getting that back. He said that Adobe is not so much focused on adding new tags or features but making it easier and faster to utilize the features that already exist in CF.

Jason did a quick poll in the room of how many people use more than 50% of tags in CF. No -one raised their hand. General consensus seemed to be that people use maybe 10 - 20% of the tags in CF.

This was a good seque to feature requests. One request that was discussed was improving CF Chart and Flash Forms.  Participants in the audience chimed in saying that CF Chart wasn’t doing the job so they looked at an extension that Ray Camden had on his blog. But then this still wasn’t enough so with about a week’s worth of work he was able to solve it using Flex. 

Jason made the point that Flex 1.5 was thrown into CF previously since the Flex server was running $40,000 or so. But now with Flex SDK being free and Flex builder being relatively cheap there was no need for CF to have this. Just use Flex. Others online added that CF Chart can be expanded by accessing the underlying Java. So one idea was to do a better job of documenting/exposing the underlying Java so that developers can extend Adobe CF in that way. (Though why couldn’t CF simply update the Flex within CF? )

Jason reinterated the value of customer feedback and was listening here to what would help the community be more productive.
This brought up the topic of what is core and who gets to decide what features are in -which brought Sean up from the audience. Thankfully.

Sean explained that

• the committee met and voted on what was core and what was not core.
• All members of the committee have agreed that whatever is core will run across all member platforms.
• They had hoped to have more visibility on what they were doing but simply have not had the time yet.

The audience started bringing up directions that CFML should go, then, from a core perspective. 

One train of discussion was on making CFML more like Java - a tightly typed language and have more parallels and access to Java.

Both Sean and Jason responded that

• there are a tremendous number of advantages in CFML not being tightly typed and
• that Java is not the ultimate model for OO development -specifically pointing to Small Talk’s place in the evolution of OO and not having interface like Java (see Patterns in Design).

There was the sense from both Jason and Sean that CFML has its own approach and philosophy, that maybe that got lost along the way/muddled a bit, but that it seems to be well in the sites now!
CFML Ain’t Nobody’s “Lite”!

Sean shared his experiences on development by commmittee pointing to his experience on the C++ committee then what Java committee with some of those features. There ultimately has to be a decision making body that says what is and what is not the core language.

Some features will be implemented differently e.g. Adobe has PDF IP that the others don’t.  But there are also areas where OpenBD community and Railo with JBoss have (and are interested in pursuing deeper)  e.g. expanding with Java.

Open BD brought up how cool would it be for OpenBD to come up with extensions that reduce the value of Blue Dragon’s CF on .NET engine.

There was then a question on how does each platform vendor get feedback from the community. Jason mentioned the bug tracker and beta testing Adobe does before a release. OpenBD and Railo both have their communities.
However, as expressed by Sean, the hope is that the CFML Advisory committee have a website where CFML users across platforms can give feedback and lobby for features in the core.  However, Sean made the important distinction between development of the core vs point specific solutions that people want solved (this seems where extensions to the language could be useful).

From here, I think the Connect Recording picks it up.

Final impressions:

• CFML has a passionate user base
• The vendors and Advisory Committee are driven by feedback
• CFML has reached a level of maturity where point solutions can be developed by third parties and
• The larger resource pools (of either Adobe or an open source community platform’s) can be focused on advancing the core in key underlining areas 
• That productivity is still the key driver behind CFML’s value - and that there are now three enterprises pushing to regain the mantle of the utlimate in rapid application development.

Stay tuned -they say big news at the keynote tomorrow -it’ll be all Ben Forta!