begin quoting Tracy R Reed as of Thu, Jan 27, 2005 at 10:46:15PM -0800: > On Thu, Jan 27, 2005 at 10:32:09PM -0800, Stewart Stremler spake thusly: > > Ooooh, I really don't like that. I really don't *want* my system to > > "come back up just the way I left it", I want it to "come up in a known > > good state", which is something quite different. > > Of course you can completely reinitialize the system state if you like. > But it is not the sort of system that you are probably accustomed to where > coming up in a known good state is such a big deal.
It's from *dealing* with systems that come up "in the last state" that encourage me to want an easy option to "boot into a known good state". A laptop with "sleep" and "reboot" accomplishes this -- and a power-off will force a clean boot. > That's like coming > from a Windows background and griping about how long Linux takes to boot > up. Waiting on Linux to reboot is not likely a problem you will have > often. Um, no. > > But if the system saves state after it's buggered itself but before it > > locks up, bringing back up the box just leads to rebooting into a wedged > > state. > > They have solved these problems. It really doesn't happen like that. The I didn't see that described in the links provided. But I only spent a quarter-hour browsing 'em, so it's likely that I missed something. > idea is that this system is extremely reliable. It doesn't get wedged. Sorry, "that won't happen" is a *lousy* justification. It just means that when the machine *does* get wedged, an attitude of "that doesn't happen here" will mean that it can't be un-wedged, unless you have a Murphy-burned engineer involved throw take some precautions (which might break the proof of correctness). "Beware of bugs in the above code; I have only proved it correct, not tried it." --Knuth > > Finally... what use is fine-grain access control when the program can simply > > demand total access and refuse to do anything until it gets it? > > Why would you want to run such a malicious program? Software written for > this sort of system would not do that. This is why I've not updated my copy of OpenOffice.org -- software that demands access or refused to run counts as malicious. And yet a sizable number of people either don't have a problem, or judge the risk to be low, and comply. A good user community will say "hell, no, get away", but we don't have that sort of community, it's not likely that we'll get that sort of community, and it may be that _having_ that sort of community dooms a platform to obscurity, as the developers search for a less critical market/audience. I'm not saying the operating systems are /bad/ -- I would like to see more variety in the OS market myself -- I just don't agree that the persistence feature is necessarily a Good Thing. -Stewart "How do you solve the Two (or Red/Blue) Army problem?" Stremler -- KPLUG-List mailing list [email protected] http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-list
