Dennis Ritchie — An Appreciation

I’m a programmer, and just about always have been. My favorite programming language is still C, and my favorite operating system is still Unix. Dennis Ritchie was, of course, jointly responsible for both. So I have definitely lost a personal hero and, to the extent that I can claim I’ve learned from his work, a mentor as well.

It’s been said that Unix killed research in operating systems. I find I don’t mind, because Unix is just about perfect. It’s said that you have to keep updating your skills in the tech world, but I’ve been programming professionally in C and Unix for more than 30 years now, and I don’t expect to have to switch anytime soon. In a field that does tend to burn down and reincarnate itself at least once every five years or so, those two wonderful little programming systems have proved remarkably durable. (And they are little, which is one of their underappreciated charms.)

Just about everybody of a certain era in programming probably considers Dennis a hero. The tech world being a bit more gregarious and less stratified than (say) Hollywood, Dennis was delightfully approachable. It was always a thrill to see a post from dmr in a Usenet newsgroup, the more so if it was in response to one of your own posts, the more so if he agreed with you. And if you got an email out of the blue — well, that was really one to be treasured. But you didn’t have to wait; any random hacker out there on the net could send an email to dmr, and he’d often reply. (I know this because he once thanked me — another email to treasure! — for being able to save time by simply pointing supplicants to the comp.lang.c FAQ list I’d compiled.)

Random reminiscence: it’s a USENIX conference, sometime in the mid-90’s. There’s a session on copyright and other intellectual property issues, and as always happens when computer types discuss this topic, there are a bunch of flamboyant statements being made about how copyrights and patents on software are Evil, information wants to be free, etc., etc. One commentator, objecting to the possibility that too-strict copyrights might stifle progress, solemnly opines that he doesn’t want to be stuck using 20 year old software. But sitting right in front of me happens to be Dennis Ritchie, who calls out in a rather commanding voice, “But you all do!”

I’d like to say I’ll miss him not only as a mentor but as a personal friend, but I only met him once or twice, so I can’t honestly say that. But I can say this: every time I simply type

r = read(fd, buf, 13);
to read 13 bytes from a file without worrying about its record structure, Dennis Ritchie lives. Every time I pipe something to grep rather than having to eyeball it for a pattern I’m looking for, Dennis Ritchie lives. Most importantly, every time I have the pleasure of using (or writing!) a software tool that’s wondrously small and simple, that does one job and does it well, Dennis Ritchie lives.

In fact, that’s not a bad epitaph. Dennis Ritchie: he did one job, and he did it well.

Steve Summit