I’m a complete metrics nerd. I’m always fascinated by numeric data, and what we can glean from it.
Recently, I’ve been having some conversations about choosing a web framework. Should we use Struts? Tapestry? JSF?
If you were going to consult a large consulting firm, like the Gartner Group, they’d probably create one of their famous “magic quadrant” pictures: a simple two-by-two grid with dots placed on the grid to represent different choices. One axis of the grid represented how visionary the solution was, and the other represented ability to execute. In theory, the dots in the upper-right quadrant were usually favourites.
I’ve always felt that these grids were somewhat subjective, but the fact that they’re produced by reputable consulting firms usually made the subjectivity irrelevant.
But I don’t think I’ve seen the big consulting firms tackle things like open source web frameworks.
So how do you make a decision? Perhaps the obvious answer is to use a lot of the frameworks in question and figure out which ones make you more effective and more productive. That’s a great approach until you have to convince a relative stranger that your choice is the right one. Suddenly, your past experience is hard to distinguish from bias.
I was pleasantly surprised by this presentation that tried to quantify this question a bit. In particular, I thought they used some creative approaches to finding metrics.
Here are two measures that interested me: a comparison of the number of books in Amazon.com or the number of job postings on Monster.com.
Even though I’m a total metrics nerd, I recognize that not everything reduces to metrics. What stands out for me is that, looking solely at the numbers in the presentation, Struts looks like a good choice. And, um. Well.
There doesn’t seem to be a good quantification of the question: “If you use this particular web framework, will you hate every day of your professional life?”