Thursday 26 February 2015

CFML: Dan Kraus - "more CF devs need to leave The Shire for a bit"

I'm going through my in box today, catching up with all the comments ppl have made that I've not replied to.

Dan Kraus made a comment the other week on that article "Lucee: does its future include CFML?". It and its follow-up warrant repeating and more eyeballs on it, so I said I'd promote it to an article so people would def notice it. I should have done this back when the article was fresh, but got sidetracked.

Anyway, here it is:

YES. 100% Agreed. In my anecdotal experience, a lot of CF devs haven't really seen much else out there. Maybe some PHP because of its ubiquity and I think because they can be both approached in a similar manner. Write some junk in a .cfm/.php file and view the results.

My post-college degree professional career experience only spans about 6 years but it started with CF (I'm under 30 yrs old and work with CF for a living, I often feel like an anomaly, it's true, sorry guys!). Mix in ActionScript, some PHP, whatever, but my understanding was a bit narrower focused and centered around CF. I think that's generally the case for anyone's first language. Once I started to see more stuff like .NET and Ruby, my frame of reference really changed. That stuff felt modern with incredible tooling. I argue CF tooling is pretty poor and I think one of many reasons it's lagged behind (kudos to the CommandBox guys for stepping in and doing something about it.)

Maybe I'm being condescending, but I think more CF devs need to leave The Shire for a bit and see what else is out there and what more we could be doing.

Frankly, I don't understand the outrage from the idea that Lucee begins to deviate from Adobe flavored CFML. If people are that committed and really need want the CFML they've worked with for the last 15 years, then pay Adobe their fees and get what you need from them. No dilemma about "will it or won't it" be supported. It seems like people think the goal is just be a free clone of a pay-for product. That's narrow and shortsighted.

And the follow-up:
I've been lucky to work to with some smart people in the last year who have worked with CF for much longer than me and have also worked across other technologies too. Perspective. We try to bring modern technologies into what we're doing. We're starting on a big application sticking with CF because that's where the team's skills are, and we still believe we can achieve what we need with it. However, everything else seems to stray a bit from the "typical" CFML stack. Linux, Nginx, Railo, MariaDB, virtualization on our prod servers, virtualization and with vagrant on our local boxes to replicate prod. We're working integrating Ansible for managing our server environments too. Our legacy codebase is the only thing that resembles "old school". It's the personification of the stigma and baggage we've been talking about with CFML. To be fair, a business was proven on it, and stuff written 10 years ago is still in production on it. We've outgrown it, it's time to move on. Our team went to ColdFusion Summit in 2014 and we all walked away so underwhelmed. We referred to it as PDF Summit. Everything was about PDF generation and rendering. Seemed to be what Adobe was pushing. There were some good talks. But it was also another glimpse in to where most of the community seems to be. I had started to write a blog post on just some of the out there things we heard and saw but eventually decided I didn't want to post something so negative in light of some personal things that came up immediately after the conference. I have to imagine others besides my team felt the same.

Good observations, Dan: thanks for making them. It should be food for thought for the CFMLers out there.

And with this quality of community observation, you should be running your own blog too! (as noted below, Dan does indeed have a blog:

Righto... inbox...