dflemstr
It's a ball.
Interesting fact: I once took my main computer, started everything I could possibly think of that doesn't directly rely on a disk (2 KDE instances, a memcached maven build, a servlet server, and some random internet video streams etc) and pulled out all of my (S-ATA) disks. It survived for quite some time without any problems at all, and when I inserted the disks again, operation continued as if nothing had happened. That thus shouldn't be a problem.Stephane Hockenhull said:Squidge said:The problem with that is people would soon complain that so much of every sd card is used up by os installs, or that they can't find there SD card with the OS on.
very true.
also, it would be quite annoying to have the kernel,X,GUI,etc crash/panic when you accidentally eject the SD card.
this has happened a couple of times I tried Linux on my iPAQ from the SD card.
(its still unusable as a Linux PDA, hopefully my Pandora will soon replace it)
Also, PNDs will suffer from the same problem: if you pull out a SD card with apps you're running on it, things will crash, horribly.
I guess that my main problem now is that PNDs can't modify the main file system in any way. So let's say that I put Apache Web server (bad example, I know) in a PND, and want it to automatically add itself to /etc/init.d or /etc/rc.d or whatever is used in Ångström; that just isn't possible. Or if I want to put a "game engine" (let's say, something like the Quake 3 application) and levels for that engine in separate PNDs, there's no way that the "level" PNDs can create symlinks or something similar to make the game engine able to find the level files. Thee's just so many hoops you have to jump through with the PND system, and some things just aren't possible any more when using it.
...
What was this conversation originally about, again?
Last edited by a moderator: