I was a young annoying engineer at Sun, I worked upstairs in building 5, Roger worked downstairs. I met him because I had a /proc question. Being the sort of jerk that I was at the time, I just walked in and sat on the edge of his desk. I had never met him before.
He ignored me. For about 5 minutes, he was coding and he didn't even look up.
Finally, again without looking up, he said "what do you want?"
"I've got a question about /proc".
"And why should I answer that?"
"Because I'm going sit on your desk and belch and fart until you do".
He burst into laughter and said something like "OK, you win".
We're friends from that time forward.
I've shared some fun times with Roger, we had a nerds thanksgiving where a handful of Sun people who didn't have girlfriends all got together and had a great time.
I'll miss him, he was a great guy. RIP Roger.
P.S. All the comments about his protection of all things Unix are spot on. He was not a Linux fan, that's putting it mildly.
Just because you have me reminiscing about Roger, I'll add (another) one of my own to the pile: before joining Sun in 1996, I had gone to East Africa, where I had purchased a carving of a crocodile -- and have had it sitting on my desk ever since. Roger loved to pick up the crocodile by the tale (it's about 18" long) and thrust the snout into my chest when making an emphatic point -- or demanding that I approve his RTI. I still think of Roger whenever I pick it up, though I have never been able to brandish it as eloquently.
raf@sunraf if I recall correctly. He is also a big part of the reason that distributed source management exists, he was a big nselite user (turned into Teamware which inspired BitKeeper, which inspired Git, Darcs, Arch, Bazaar and who knows what else).
At the time of the great Network Division disaster (also known as a Garlick's Folly) I was talking to Roger at a beer bust and he assured me that even though I was going to be "network guy" now, he was ok with me changing the kernel when I needed to. It really made me feel great.
Don't get me started about Garlick (or Kannegard), oy. Having Roger say you have his blessing the wack the kernel is a pretty big deal, he was very protective of it.
Roger's greatest gift was "telling it like it is", but always in a way that you knew was truthful, well-meaning, and not in a way purely for the sake of offence. He always cut right to the heart of the matter, and of course his distinguished Carolinian accent helped soften the blow to your ego.
The last time I had the opportunity to collaborate with him was on a fairly minor addition to Solaris libc I was making -- pthread_getattr_np, in January of this year (2016).
Let me tell you from experience that the greatest praise any engineer could receive and the greatest sign of success for their projects was Roger's endorsement.
I miss him greatly, and UNIX and the world feels lessened by his passing.