I recommend reading this rant by a developer named Andrew Ray about Vim and SublimeText.

It is a good reminder about commitment bias and how escalation of commitment makes objective measurements of developer efficiency so difficult. Developers are especially prone to being subjective when we speak about efficiency. And at times we aren't honest (even with ourselves!) about efficiency because that kind of honesty can be painful to our pride.

But in this particular case, I'm also personally gratified that someone such as this author, who has invested four years in drinking the Vim Kool Aid has the refreshing honesty to admit regret. Where most Vim users proselytize because deep technical commitment bias seeks friends, the author says he will do that no more. That takes strength.

I have powerful commitment biases of my own, but try to remain open. Knowing the amount of investment that goes into becoming competent with something new routinely leads me to ask for credible evidence of increased efficiency. In my years, I have seen dozens of orchestrated demonstrations of products that claim to increase efficiency. Remember how amazing Macromedia's products seemed when you saw them the first time?

I've been down variations of the following road enough to be suspicious.

What? Hold on, there, chief! This tool makes building web pages that quick? Would you look at how beautiful that is? Holy cow. That is awesome. Let me try.

Hmm, that's weird. Uh, I don't get it. I can't get it to work with my design.

Wait, this isn't working at all. Oh my, this is horrible. What have they done? Oh the humanity! I want my week back.

And yet, people who had invested the time to become experts with the tool would tell me that I just didn't give an adequate amount of time.

Still, I'm happy I didn't bother learning Dreamweaver.

When Vimmers tell me they use Vim because they can avoid moving their hand to their mouse in order to select text, I don't even know how to respond. They have such a different notion of productivity and efficiency that we cannot possibly agree on subjective measures, let alone objective measures. Their goals are oddly focused on micro optima such as "avoid mouse use" rather than macro optima like "get the work done." When observing Vimmers over their shoulder, some things seem modestly efficient and others seem tragically inefficient.

I feel the same way with Unix in general, which is why I routinely express regret that computer science has lost interest in operating system innovation. We are stuck with a few "good enough" micro optima.

I also feel the resurgence of Vim is among the preeminent indicators of a harmful (not disastrous, but harmful) modern groupthink in software development. I admit to having a greater than rational distaste for groupthink. If I sense large-scale groupthink is at work, I apply an exaggerated negative score to my internal Bayesian inference calculation. But that's a topic for an entirely different rant.
About this blog