Thought Delimited

For those readers who didn't see my previous post regarding my conversation with Glen, Glen is a technical trainer who treated me to the usual "ColdFusion is dying" mantra we normally hear from other programmers.

The big question I had after that first encounter was why, as a trainer rather than a programmer, he had developed that opinion about ColdFusion. So when I ran into him again at the gym the other day, I asked him about that.

Turns out his belief that ColdFusion is dying is based on the job market for ColdFusion jobs. He told me he works/interacts with a dozen recruiting firms in the Washington D.C. area, firms looking to fill positions for government contractors like Lockheed, and that the number of ColdFusion positions compared to the number of positions programming in Java or Ruby is just so small. He added that a lot of the ColdFusion positions that did exist were senior positions where candidates were expected to know how create web services, work with Java, write object-oriented code, etc., making it hard for up-and-coming ColdFusion developers to find work.

I also found out that most of his training work involves training/teaching programmers OO-based languages over several weeks, so his perspective on programming trends isn't all that different from an actual programmer. Glen (who doesn't mind talking) went on to give me the standard advice given to modern-day programmers (the importance of having multiple programming language skill sets, the need to have a new job lined up before leaving your current one, etc.) before we parted company once again.

My thoughts? I don't doubt there are more jobs out there for languages like PHP, Ruby, and Java just as Glen said, but whenever I go out on websites to look for ColdFusion jobs, they're out there, and while many of them are senior-level positions, there are a few junior-level jobs to be had. Actually getting hired, however, made take some effort, as Michael Dinowitz noted in his recent blog post, "Are There Really ColdFusion Jobs?"

In the end, Glen's position on ColdFusion is nothing original: lack of marketshare is at the heart of every "ColdFusion is dying" argument we hear. But marketshare is only one metric, and it shouldn't be the main consideration when choosing a programming language.

The main consideration should be "does this technology allow me to build the web application I want?" And when that question is directed at ColdFusion, the answer is almost always "Yes."

Yesterday I ran into an old colleague at the gym (we'll call him Glen). Glen conducts training sessions on technical topics.

Glen asked me if I knew if anyone at the university was using Ruby on Rails. I told him that I didn't think so. He then told me that he was giving Ruby on Rails training at John Hopkins University because they were looking to reduce their use of Java. I responded by saying that I thought Hopkins also used ColdFusion for certain things.

I expected him to either simply agree with my observation, or perhaps offer an explanation about why Hopkins was looking at RoR verses expanding their use of ColdFusion. I was not expecting him to respond with the standard FUD about ColdFusion: it's dying, people are moving away from it, etc. He even went so far as to refer to it as the "COBOL of web programming languages."

I countered with the numbers announced at cfObjective(), that the number of ColdFusion developers had increased dramatically over the past year. That elicited a "well..." and then another assertion that ColdFusion use was decreasing in the federal government as well (there are a lot of government-oriented CF jobs here in the D.C. area).

There wasn't much else to say or do at that point. I couldn't hold him up from where he was going and I had somewhere else to be as well. So I shook my head, laughed at his prediction of ColdFusion's demise, and basically said "Yeah, we'll see about that."

My immediate feeling after the conversation was more amusement that annoyance. Year after year, we hear the annual proclamation that "ColdFusion is dead," yet the proclaimers never seem to realize that if they keep saying it year after year, ColdFusion obviously must be still around. And more often than not, they have no hard numbers to support the idea that CF use is waning, just statistic-free assertions.

Now that I'm several hours removed from the conversation, I find myself more curious about why Glen even had an opinion on ColdFusion in the first place. I don't know him all that well, but he's always struck me as a fairly rational person and not someone who's looking to make and win an argument, yet he immediately began disparaging ColdFusion the moment I mentioned it. What or who lead him to having this opinion about CF? I may have to ask him that the next time I run into him.

I just finished reading Issac Dealey's recent blog post (worth reading, by the way) where he shares some of the results from a survey he conducted regarding framework preferences back in September.

In the post, he mentioned how difficult it seems to be to get folks to participate in these kinds of surveys, and it reminded me that someone else in the community was recently pleading with folks to take their survey (unfortunately, I forget who, but I did take it). And I wonder what kind of a response Hal Helms is getting with his ColdFusion Web Developer Survey.

So it got me thinking: is surveying the community/getting feedback from the community a problem in need of a solution? Are these surveys being neglected because people don't see the point or don't have the time? Or is the low response rate more the result of a lack of publicity or poor technical implementation of the survey itself?

Any thoughts?

In my last blog post, I suggested that Adobe include a list of CFML community resources in the upcoming Bolt IDE in order to promote the community to isolated developers who might otherwise be unaware of all the resources out there.

After thinking about it a bit more, it occurred to me that maybe Bolt could take it one step further. Instead of simply using Bolt to point developers to the community, have Bolt bring the community to the developer. Build in an RSS viewer that displays the latest ColdFusion posts from Adobe Feeds. Put in a communicator tool so the developer can converse with other CFML programmers via IM or Twitter. Let the developer screen-share their code with other developers both inside and outside of their organization. Integrate geolocation into Bolt and show the developer a list of other Bolt users (and maybe Adobe user groups) that are nearby. Instead of using e-mail and message boards to communicate with CFML developers, Adobe could broadcast any news announcements to all of the Bolt installs, and Bolt users could submit questions to Adobe and other users via discussion forums displayed in a window of the IDE that gets refreshed automatically.

I'll admit, it's a pretty pie-in-the-sky idea. Given that Adobe only has a finite amount of time and resources, I would certainly not want Adobe to leave out any traditional IDE features, the things that allow developers to code quickly and efficiently, in order to take the time to add all of the things I just suggested.

But if they did have a little extra time, I think adding even one or two simple collaboration/informational features would certainly enhance the product, and perhaps set a trend for other IDEs to follow.

I was a little surprised this morning to find little or no mention of the announcements made at the Sneak Peek session at MAX last evening on any of the ColdFusion blogs aggregated by Adobe Feeds. Either I'm missing something or everyone had too much fun at the after-session party last night. :)

I don't really have any of the details about the announcements, since I was only half-paying attention to the live blogging from the event and the Twitter stream, but two items stood out for me.

One was the announcement that server-side ActionScript is in the works. For those who don't know, ActionScript is the language of Flex, which is a client-side technology. Someone on Twitter said that the announcement meant that you could run ActionScript on the ColdFusion server, so that you could code certain things in ActionScript rather than CFML, but I don't know if that's really the case or not (I'm sure that will be clarified within the next few days).

The second announcement that caught my attention was about Durango. To quote the Durango web page on Adobe Labs (it's already available for download): "Durango is a framework that allows developers to build Adobe AIR applications that can be customized by end-users." Basically, it sounds like a means of allowing user-created mashups in an AIR application. Giving end-users the ability to make their own mashups seems to be a trend in the industry lately. It remains to be seen whether users will make use of that kind of power and flexibility.

Anyway, I expect folks who are actually at MAX will blog about these items and provide some more details, but I figured I put these items out there so people know what to look out for in upcoming posts from the community.