ColdFusion isn’t going anywhere
Ok, so I’ve purposely chosen an ambiguous title – in this post I’m going to discuss why I think ColdFusion isn’t going anywhere. It’s not going to be a ‘ColdFusion is dead’ post because that’s simply not true – but I think I can say that it’s not going anywhere and will provide evidence to back that claim up.
Remember, I am a ColdFusion programmer – I spend pretty much 100% of my working day coding ColdFusion – we have invested time/effort and money into ColdFusion so I, as much as the next person want ColdFusion to be going somewhere.
I was going to start by discussing the results of my recent survey asking when did you start coding in ColdFusion but I’m not. Instead I’m going to turn my attention to the ColdFusion Evangelism Kit.
On page 2 it states 500,000 Developers and cites EDC 2007 Global Developer Population and Demographics Report as the source. Now costing around $16,000 this is not something I’m ever likely to get my hands on but that’s not the point – it’s the figure that is important. A few Google searches reveals that the figure of 500,000 has been banded around for quite sometime now, in fact the earliest reference I can find is back to CFUN (aka CFUnited) from 2000 in a talk named CF Programming Philosophy and quite aptly the content is still very relevant today (although I’m not sure who wrote this). The point though is a reference to the number of ColdFusion developers that Allaire claims there were 500,000. The year 2000 was in-between Allaire’s release of CF 4.5 in November 1999 and Macromedia’s release of CF 5 in June 2001. So if Allaire reckoned there were 500,000 developers in 2000 yet the EDC data from 2007 claims there are 500,000 developers 7 years later then something is clearly wrong. Since Macromedia bought Allaire in 2001 the releases of ColdFusion 5 through to the current version ColdFusion 8 have (according to the the data available) added no more developers – hence my post title “ColdFusion isn’t going anywhere”.
The more worrying fact is that another report (also by EDC) projected that the global developer population would see an increase of 46% between 2005 and 2009 so why has none of this growth not been seen in ColdFusion developer numbers?
I’m not disputing that ColdFusion has attracted new developers over the years but equally developers have stopped using ColdFusion and shifted to other technologies so the net effect is that ColdFusion hasn’t seen any growth in developer numbers and this is my cause for concern.
Returning to my survey of ColdFusion developers – incidentally, my sample set was obtained from posting on my blog, via aggregators, Twitter, ColdFusionCommunity.org – and whilst a sample size of 432 is relatively small, 59% responded that they’ve been using ColdFusion for 8+ years (this holds true even after a second push of the survey where I doubled the sample size but the percentage hasn’t changed). The people that have responded are ‘real’ people – I’m not working to extrapolated numbers as the results of the EDC survey will be – can anyone reading this confirm they actually took part in the EDC survey? Am I really supposed to believe that the number of people that really do ‘care’ about ColdFusion, that read blogs, twitter etc is really this small? Now that does really concern me! Perhaps they think “oh another bloody survey” and ignore it, maybe – but if they do that to my survey then they’ll do it to other surveys so the data from EDC is equally open to misinterpretation.
Returning to the link previously referenced (CF Programming Philosophy) which asked ‘where are all the people?’ and ‘why are there people out there who are not helping grow the community’ it appears that ColdFusion is finding itself now (or still) in exactly the same position as it was almost 9 years ago – people are certainly using the product, if it was dead then it would have been discontinued long ago but it’s user base appears not to be growing and ‘growth’ is a key characteristic of something to be considered alive.
I’m not sure of the solution though, I’m excited where Railo is going with the open source engine but still providing ‘commercial’ add-ons/features. I know it’s been spoken about before but perhaps now is the time where Adobe needs to go too, release a ‘core’ version of their product to compete with the likes of PHP, Python, Rails (i.e. free) and then sell features/add-ons as commercial products, in which case I’d imagine a whole ecosystem would evolve of companies offering third party add-ons – much like the case of ASP.NET, PHP etc etc. Arguably who is responsible for growing the CF community? As long as Adobe is making sales then they’ll be happy – do they care about growing the community if sales are growing? Are they waiting to see how much the free offering from Railo impacts their sales before making a decision – as surely if there’s a free alternative to ColdFusion out there then that should grow the community too? Have the likes of OpenBD and Railo done anything to further expand CF to new shores?
Thoughts anyone? But if you say “I don’t care if you are right, you are just such a self-righteous dickhead” then I will delete your comment! 
Related posts:
I really shouldn’t comment on this but I think a few things need to be said.
Surveys and reports from companies like EDC, Gartner and Forrester provide a significant source of data to the industry. Feel free to dismiss their merit and data, but do so across the board, not just for ColdFusion. The same 2007 report that noted ~516k used ColdFusion also denotes the ~3.7m use PHP, ~6.9m us Java and so on. These are the reports the industry uses to measure all communities, not just ColdFusion. Rather than focus on the specific numbers, I look at the relations to other technologies and the % of growth over time. The growth trend we see with EDC is similar to the growth trend we see with our sales data. Yes, growth. Not decline.
Everyone is free to speculate what they will about CF and it’s current market. But it seems a little silly to cite a CFun preso from 2000 as a data source.
As far as the OpenBD and Railo go, I’m still not convinced they will make a significant impact in the growth of the CFML community. Having an OS/free alternative to Adobe CF means that customers are not locked into a proprietary language (CFML). Their mere presence will prevent a number of companies from migrating to .NET/PHP/JSP. I doubt FOSS CFML will attract a horde of new users. Of course, I would _really_ like to be proven wrong. It’s just that all our research has said that price/os isn’t the primary obstacle. Time will tell.
Incidentally the 500,000 figure from the 2000 CFUN preso came from a presentation by Michael Dinowitz – I’d love to get the figure substantiated some how?
The issue is Education, until I came out of Uni I had not even heard of Coldfusion. The fact is, CF is a fantastic product, I think every company should be using it, but that one view. As a student my views were tinted by what skills do I need to get a Job and in the UK CF would have left me limited. Its was not until the company I work for starting letting me use CF that development and support time when down. Now we only employ CF developers. I mean CF is out there its everywhere I turn online now, but we still only have 4 BOOKS on CF8!
Its not Adobes fault is ours. Adobe have already started to understand Education is Key. Educational versions are Free, Adobe even run the CF exams at our local college now. Its down to us we have the tools.
as for cost, this has never been an issue with our cleints. corporate clients tend to want corporate solutions.
I think Adobe will continue to develop CF as a powerful and rapid development middleware solution. I’m looking forward to stronger Flex (and media format) integration. I speculate that Adobe is interested in deployment options in the cloud (stax.net) more than open source.
Hi Adam,
So if i’m not to dismiss EDC data, am i to take it then that having a commecial product that has been on the market for 14 years and has a developer community of ~516k compared to PHP which also debuted in 1995 and has a developer base of ~3.7m is something to shout about? Oh and not to mention Sun releasing Java also in 1995…?
@John: Don’t forget the number is not changing through the time.
@Adam: I think Adobe Evangelists are best people to comment on the issue. It simply looks like some one picked that random number one day and just sticked to it.
It’s interesting to me that we as a community so often look for answers in statistics from surveys and reports more than the product/software itself. While not intending to question results from the reports mentioned above, I must say that after spending a few years in graduate school working on an advanced degree in anthropology, I learned to have an inherent distrust of statistics. Several of our seminars actually taught us how to skew data for our own ends (not promoting this; rather, as a critical thinking exercise so you can see how to better question statistical “data”. Ultimately, any one of us could, fairly easily, pick apart the resulting numbers.
For me, the “going somewhere” really comes down to growth of a product and an investment in said product by the owner/maker/etc. In my opinion, the growth of ColdFusion since Adobe took over has been astounding. And now CF9 is looking even more impressive.
I don’t have a direct opinion, however I can share the following, also debatable, trends:
http://www.indeed.com/jobtrends?q=coldfusion+or+%22cold+fusion%22&l=
http://www.indeed.com/jobtrends?q=php&l=
http://www.indeed.com/jobtrends?q=%22c%23%22&l=
http://www.indeed.com/jobtrends?q=j2ee&l=
I think what Craig Kaminsky is stated hits the nail on the head.
I was going to write a blog post a few days ago, basically using the same sort of graphs as Calvin was using. Even though Coldfusion on its own does display a slight increase, (especially when looking at the Relative scale:
http://www.indeed.com/jobtrends?q=coldfusion+or+%22cold+fusion%22&l=&relative=1
)
but when you compare it against almost any other language, it looks static
http://www.indeed.com/jobtrends?q=coldfusion+or+%22cold+fusion%22%2C+php&l=
I’ve done some more digging into the 500,000 figure quoted in 2000 and found this ALLAIRE press release…..
and an interview with Jeremy Allaire in JAVA magazine August 2000 here again, 500,000
I really agree, CF should have at least a core version for free to make easy start up for companies and sell features separately. I know lot of companies moving out from CF just because they need to pay for that. Even they switch to other “free” solution they anyway will pay but later that is the main trick I think. Some companies never use report engine for example or Exchange integration etc, maybe later they will use that, so the flexibility with end user, Adobe can get more customers…….and probably CF become more popular tech…..
I was a long time CF developer, and slowly but surely moved into Java development, where I have been for the past 3-4 years.
Although there were CF jobs around, they were really not that challenging, and the market seemed to be headed more towards Flex whereas I’m more interested in scaling and complex business problems. Having worked in those types of environments as well as having a Computer Science background, I find that I have more in common with the average Java developer as well because we share an understanding of general concepts like threading, memory/disk scheduling, algorithms, state machines, etc.
I like ColdFusion and am stoked that there are some open source versions of it now (I mean I still read CF blogs despite being away for so long), but I don’t think it’s gaining many new developers. Mainly the ones who have already been doing it for quite awhile are the ones still doing ColdFusion.
One way to look at what future platforms are going to be like is to see what startups are doing. I think I have a pretty good overview on what people are using, and it tends to be RoR, PHP, or agile Java (ie.. Spring, Hibernate, maybe Grails).
I have seen only a very small amount “hot” technology startups using ColdFusion since the early 2000’s. CF 5 was really CF’s heyday as far as startups are concerned. Right around CF 6 is when PHP really started to get some momentum.
I look forward to whatever platform will supplant Java (no, Ruby on Rails is not it.. if you have done any development in Java for large platforms you know why) and look forward to learning it, but in the meantime Java development is what is best for me personally. I think ColdFusion has a good niche and is a good mature product which is going to be around for a long time, but I think people who want to pick up a new web language/platform are going to opt for PHP, RoR, or maybe even Django instead because they are “sexier” and don’t have any legacy pre-conceived notions about them.
Anyhow, that’s just my opinion. I’m a pragmatist and not really a fan boy of any particular technology (I can tell you a billion things I hate about Java if you ask me) because they all suck in one way or another, so don’t take what I’m saying as an attack or anything.
One other thing.. to see what I mean about few new developers are using CF, take a look at the survey John ran awhile back. While I didn’t break it down mathematically or anything, look at the split of people who have been doing CF for 1-5 years, to 6-8+. It is obviously heavily biased towards experienced CF developers.
There are many reasons that the poll can be thrown out as inaccurate and I wouldn’t base a major decision on it, but it is an interesting data point.
Regarding, “…growth trend we see with our sales data.” From publicly available documents:
1. During its last full year before the acquisition by Macromedia, Allaire revenue was about $100M.
2. When Macromedia reported server division sales separately, it appears that ColdFusion revenue had declined to about $50M/year prior to the Adobe acquisition, or about half of what it was under Allaire.
3. Adode does not report server sales separately, so it’s impossible to know what ColdFusion sales have been under Adobe.
Some questions:
a) Are the above number accurate? If not, what are the real numbers?
b) When Adam says “ColdFusion sales are growing” does he mean from the lows experienced during Macromedia’s reign, or from the high’s experienced by Allaire?
c) Are ColdFusion sales revenues today more or less than they were under Allaire?
It seems that Adobe could answer the “ColdFusion is dying” critics simply and directly by providing sales figures. Why don’t they?
@noname: it is worth pointing out that under Adobe only one CF version was released – ColdFusion 8 (with 8.1 update). So infact following statement: “ColdFusion sales are best since Adobe acquired Macromedia” is true
Section Releases:
http://en.wikipedia.org/wiki/ColdFusion
To quote Kristen Schofield (Marketing Manager for ColdFusion):
http://www.webbschofield.com/index.cfm/2008/9/15/ColdFusion-Evangelism-Kit
“Matt -
the 500k number is refernced in the kit as sourced from Evans. The precise wording should be “over 10k companies” – this is data we have internally at Adobe via customers we work with directly and through the channel. CF has a sweet spot in large enterprise companies, at the departmental level – as well as in gov’t agencies. There are some co’s / agencies like this that have 15 to 200 CF developers, *worldwide*. Many of the CF developers are scattered amongst different departments within the same company. This is the most typical scenario for CF. Not too long ago, we did an analysis of the number of units of CF shipped and compared that with what we knew to be the average number of developers associated with the units and found the data similar to what Evans claims.
Kristen”
So, in Adobe’s eyes, there are 50 developers per shipped unit of ColdFusion.
Rubbish
To address the question of “where are all of the CF Developers” in specific regard to answering the “how long have you worked with CF” survey, the majority of CF Developers that I work with don’t follow the CF blog circuit or related mailing lists. It’s unfortunate but most of them just ‘maintain’ their current level of CF skill set without looking to further improve it.
I fear that a large part of our stagnant growth is the fact that CF is mostly used in larger corporate settings where the developers tend to be older, have less time on their hands (b/c of family) and generally have less ambition to excel further.
The last part of that statement is a bit presumptuous and obviously not true for all situations, but realistically I’m amazed at how many people simple treat their career as a day job and nothing more. I am one of the few at my office that actually try to keep up with whats going on outside of the vacuum and am constantly looking to improve. Many of the folks I work with rely upon my company to push them to increase their skill set through paid training, etc… very few actually do any reading or learning outside of the office.
As an aside, I myself have migrated towards working more with Java on the regular day to day. The majority of the design and building that I perform during the day is with Java, hibernate, spring, some JSP/JSF, etc.. Its really only on my side projects and outside ventures that I use CF entirely (with some Java sprinkle of course).
I still consider CF to be the best web application language out there hands down. I don’t think that it’s going away anytime soon and I think projects like Railo and BD are helping further it’s acceptance into the younger, open source generation of developers. I think we as a community however need to figure out a better way to remove the negative stigma that seems to permeate from larger development communities.
I’m closing comments on this post now. The neccessary facts have been presented, make of them what you will…back to writing code!