Yeah no doubt. I really like Bjarne, he's a big inspiration to me over the years. He's just so damned persistent in what he does, never gives up trying to improve on his work.
To quote Patton - Bjarne, you magnificent bastard! I read your book!
I have met Bjarne a couple of times with his Surface and it is running standard Windows. I mentioned it once and his response was basically "Windows works best on it so I stick with it". Like he said he likes a simple and stable environment and Windows does provide that.
"Able to perform perfectly when disconnected from the web" is really vague since it already exists. You'd still have access to the broader Internet and your local device, obviously.
"Able to act as an interface to other systems" is a matter of network transparency and authenticating to a remote node manager. Example: drawterm(8).
"Simple and predictable to use" is subjective, but generally lends itself well to systems built around exploiting a few Grand Abstractions to compose higher functionality. Inferno, Spring, etc.
"Stable, never needs rebooting" is already the case for big iron and mainframe systems, among plenty of standard microcomputer server racks. Microkernel architectures with autorestarting of failed drivers and well-defined communication boundaries greatly aid availability and reliability. The more important thing besides never rebooting is maintaining orthogonal persistence with frequent checkpoints so that rebooting isn't a big deal when it finally happens.
23 comments
[ 3.3 ms ] story [ 76.5 ms ] threadTo quote Patton - Bjarne, you magnificent bastard! I read your book!
https://www.youtube.com/watch?feature=player_detailpage&v=Ug...
"Able to act as an interface to other systems" is a matter of network transparency and authenticating to a remote node manager. Example: drawterm(8).
"Simple and predictable to use" is subjective, but generally lends itself well to systems built around exploiting a few Grand Abstractions to compose higher functionality. Inferno, Spring, etc.
"Stable, never needs rebooting" is already the case for big iron and mainframe systems, among plenty of standard microcomputer server racks. Microkernel architectures with autorestarting of failed drivers and well-defined communication boundaries greatly aid availability and reliability. The more important thing besides never rebooting is maintaining orthogonal persistence with frequent checkpoints so that rebooting isn't a big deal when it finally happens.
Makes me really want to know what Ken Thompson uses.