I’m not a big fan of IT analysts, and have noted so previously. I find that far too many of them are all opinion and no experience, tossing supposed expertise around with little to back it up and even less transparency, and in doing so often actually impacting buying and deployment decisions.
Now to be very clear, there are exceptions. Last week in Milan I got to spend some time with RedMonk‘s James Governor who seems to relish breaking the technical analyst stereotype – he’s direct, honest, knowledgeable, and brutally blunt, and he cares less about being popular or quoted and more about finding sanity and clarity amidst cesspools of hype and hyperbole. But, of the analysts I have met (and I have met many over the past two decades), he’s the exception, not the rule.
I bring this up because I was recently pulled into a conversation with a U.S. Government agency that has been using ColdFusion successfully for a long time, and has recently started working with Flex as well. Apparently, as is the case in most large organizations, multiple products and technologies are in use, and there was some discussion internally about the platform on which to build a new application. And so, the organization (I promised not to name names) sought outside expert advice by contacting Gartner who then came back with a recommendation.
Before I go any further, there is something I should state quite emphatically, and that is that I fully appreciate that there are lots of options out there, and sometimes ColdFusion may not be the right tool for the job. If careful research is performed, and a decision is made to use PHP or ASP.NET or Java or anything else, well, I can respect that decision (even though I’d try to encourage the use of ColdFusion when and if I think it is the right choice – hey, I am biased, and I freely admit it). Ok, now back to our regularly scheduled programming …
Gartner came back with a recommendation, a copy of which I have read and reread. And they did not recommend ColdFusion. Nor did they recommend Ruby on Rails or Java or ASP.NET or … So, what did Gartner recommend for brand new development instead of ColdFusion? I hope you are sitting down … Gartner recommended the use of Perl. Yep, Perl, the same Perl that was created in the late 80s, the same Perl that has been a staple of IT departments and system admins for decades, the same Perl that primarily relies on CGI for use with web applications, the same Perl of which the current version 5 was released in 1994 (and yes, I know that version 5.10 came out a year ago and v6 is in the works).
And no, I am not badmouthing or belittling Perl – it is indeed proven and battle tested and powers some impressive applications including Bugzilla and Moveable Type and parts of some impressive sites including Amazon.com and IMDb. Still, it’s a pretty astounding recommendation, and one that is somewhat harder to fathom than, say, recommending PHP or ASP.NET.
Ok, so why was Perl recommended? The first criticism of ColdFusion is one of leveragability, that logic created in ColdFusion can’t be easily shared with other applications. Obviously, considering Web Services interfaces, and gateways to just about anything, this is flat out incorrect. I am not sure what type of integration Gartner is referring to, but I fail to see how logic in Perl is inherently more sharable than logic written in ColdFusion. The next criticism is cost, and Gartner notes that as commercial software, ColdFusion has a higher acquisition cost. Interestingly, Gartner explicitly points out that this is an acquisition cost issue, but then goes on to simply ignore any ongoing cost issues (like training time, development time, etc., even though it previously points out that a benefit of ColdFusion is that it is easy to learn and fast to develop in). In fact, the report goes on to point out that code written in Perl is more complex, and that maintenance by future developers is a potential challenge. There’s more, but you get the gist of it.
Gartner concludes by stating that Perl can run on multiple platforms, and then noting that ColdFusion can also run on multiple platforms, but that Perl can also run across multiple solutions. (I have no idea what that even means!). And finally, the report makes the argument that Perl is open source and thus provides a higher level of investment protection than single vendor solutions, apparently unaware that there are in fact other products (including open source offerings) that support CFML (to varying degrees).
The report does not state whether or not the researchers have ever actually used ColdFusion or Perl (or any other player in the space), apparently that is not a prerequisite to actually making recommendations.
I know I am generalizing somewhat, and am sure that there are analyst reports and expert opinions that are balanced and sound and well reasoned. Unfortunately, I don’t seem to run into them much, which makes the reliance on supposed experts a very worrying notion.
And that end of the day, analyst opinions are just that, opinions. They are not facts, they are not automatically valid, and they are not even necessarily driven by subject matter expertise. They are just opinions, albeit expensive ones. And as for opinions, well, you’ve probably heard the expression “Those who can’t, teach”. The phrase is actually rather absurd, I know many who can and teach (and many who can’t and thankfully don’t!). So, I propose that we update the phrase to “Those who can’t, opine”.