"Ability to type quickly and accurately maximizes the effectiveness of time spent in the flow of programming. Hunting and pecking means you only capture a fraction of what could have been done."
I submit that typing speed matters, but only as a component of what I'll be unoriginal and call actions per minute (significant APM, to be precise). More important components of SAPM would be pushing blobs of text around (editor knowledge) and manipulating/filtering blobs of text efficiently (Unix toolchain).
A 60wpm developer who uses rectangle commands and pipes a portion of buffer to shell to transform it may have higher SAPM than a 150wpm developer ignorant of the various shortcuts available.
To further flog the RTS analogy: typing speed and keyboard shortcuts that operate on one line are micro. Regex replace, sed/awk-fu, multiple buffer usage, etc are macro. To be an efficient dev both are necessary but not sufficient -- all the micro and macro in the world can't save you from brain-dead decisions.
Agreed, and let me add typing speed is important to 90% of a programmer's activity outside a source code editor. The changeset commit message, the bug tracker description, the status report email, commenting on Hacker News... ;)
Anyone who believes typing source code faster increases productivity by any worthwhile amount is solving an inane problem. If the bottleneck is on your fingers, then 10 outsourced developers or a Visual Studio drag-and-dropper will replace you.
I not sure the author disagrees with you here, however he doesn't specifically come out and say it and I can't really read his mind. So let me attempt to reword it anyways.
Being proficient with the editing tool(s) of choice is of significant importance to productivity and keeps the natural flow and rhythm of the thought process while working. Hunt-and-peck is absolute crap, however, likewise anything over say 50wpm in concert with solid knowledge of the shortcuts and meta commands available is going to be sufficient to achieve a good working rhythm.
The premise the author made that I completely agree with is that touch-typing (i.e., not searching for the keys) is not an optional requirement for a productive programmer. I am mildly shocked that there are people on HN (not yourself obviously, but I've read through some of the comments here, and in other posts) who get worked up about this and actually get up-voted for defending it.
I was astounded when I noticed a coworker using hunt-and-peck typing. I simply don't understand how a professional who is passionate about their work can not spend the time to master their tools - especially one as fundamental as a keyboard for someone who uses the keyboard 8+ hours a day.
I agree: many developers not only refuse to leverage their tools, but even aggressively defend this lack of automation. (Ironic, since the game of the game is automation.) Take for example copy/pasting from an Excel document (or worse yet, eyeballing and 'copying' by hand) — when it's frequent enough to justify automating.
Often these inefficiencies come from a lack of critical thinking: What are your tools' deficiences? What's better than "How everyone's always done it"?
I submit that typing speed matters, but only as a component of what I'll be unoriginal and call actions per minute (significant APM, to be precise). More important components of SAPM would be pushing blobs of text around (editor knowledge) and manipulating/filtering blobs of text efficiently (Unix toolchain).
A 60wpm developer who uses rectangle commands and pipes a portion of buffer to shell to transform it may have higher SAPM than a 150wpm developer ignorant of the various shortcuts available.
To further flog the RTS analogy: typing speed and keyboard shortcuts that operate on one line are micro. Regex replace, sed/awk-fu, multiple buffer usage, etc are macro. To be an efficient dev both are necessary but not sufficient -- all the micro and macro in the world can't save you from brain-dead decisions.