Comment Fundamentally wrong security model (Score 1) 571
I can't help thinking that there's something fundamentally wrong with this whole approach to PC security.
Now, as far as I'm concerned, all my PCs are extensions of my own mind. No one else is going to be using them, and it's MY responsibility to ensure that code I don't permit never runs on them.
This implies several things, all of which are contradictory to 'how it's done' at the moment:
* There should never, ever, be any 'active executables' that must run on the PC as part of net browsing or any other activity. Flash, Java, active agents, dynamic plugins, etc - all are a bad idea. Nothing should come in but passive data, that applications already on the PC (by my permission) parse to display.
* Anything that IS installed on the PC should have full access to all PC resources. I don't set varying 'permission levels' to different parts of my own mind, and shouldn't have to put up with this shit on my PC either. At the moment the brain-computer link is so primitive (keyboard-screen) that the incompatible approaches are still workable. As technology advances, this will cease to be so.
* The whole 'permissions' ideology inevitably leads to the kind of DRM insanity that has started with Vista and 'secure computing/trusted computing', and will only get worse (if Microsoft has anything to do with it.) When one considers the computer as an extension of one's own mind, then such DRM bullshit equates to mind control. Which is probably where Microsoft would like to go next.
* The right security model for personal computing, is something more like a perimeter fence. Anything outside the fence is considered hostile. Anything let through the gate is going to have to undergo a very thorough checking out (such as being required to have all executable code in some plain text interpretable form, that can be scanned for nasties). Once inside the fence and OK'd, it is 'part of you' and has the same access to everything as you do.
* Just as your mind has introspection, a conscience, that monitors what you do, PCs need a hardware means to continuously and invulnerably monitor the computer's activities, and throw an emergency halt if something stupid is happening. Some kind of secondary CPU and firmware that acts a bit like a continuous tracer and profiler, and which can't be corrupted by the main processor's actions.
In other words, dispense with ALL the annoyances of internal security, and rely on perimeter executable exclusion, backups and self-activity monitoring to catch and recover from any hostile or faulty internal code operation.
Note that any kind of DRM management would be impossible in such a model. GOOD!
But that is why TPTB will not develop such an OS.
Now, as far as I'm concerned, all my PCs are extensions of my own mind. No one else is going to be using them, and it's MY responsibility to ensure that code I don't permit never runs on them.
This implies several things, all of which are contradictory to 'how it's done' at the moment:
* There should never, ever, be any 'active executables' that must run on the PC as part of net browsing or any other activity. Flash, Java, active agents, dynamic plugins, etc - all are a bad idea. Nothing should come in but passive data, that applications already on the PC (by my permission) parse to display.
* Anything that IS installed on the PC should have full access to all PC resources. I don't set varying 'permission levels' to different parts of my own mind, and shouldn't have to put up with this shit on my PC either. At the moment the brain-computer link is so primitive (keyboard-screen) that the incompatible approaches are still workable. As technology advances, this will cease to be so.
* The whole 'permissions' ideology inevitably leads to the kind of DRM insanity that has started with Vista and 'secure computing/trusted computing', and will only get worse (if Microsoft has anything to do with it.) When one considers the computer as an extension of one's own mind, then such DRM bullshit equates to mind control. Which is probably where Microsoft would like to go next.
* The right security model for personal computing, is something more like a perimeter fence. Anything outside the fence is considered hostile. Anything let through the gate is going to have to undergo a very thorough checking out (such as being required to have all executable code in some plain text interpretable form, that can be scanned for nasties). Once inside the fence and OK'd, it is 'part of you' and has the same access to everything as you do.
* Just as your mind has introspection, a conscience, that monitors what you do, PCs need a hardware means to continuously and invulnerably monitor the computer's activities, and throw an emergency halt if something stupid is happening. Some kind of secondary CPU and firmware that acts a bit like a continuous tracer and profiler, and which can't be corrupted by the main processor's actions.
In other words, dispense with ALL the annoyances of internal security, and rely on perimeter executable exclusion, backups and self-activity monitoring to catch and recover from any hostile or faulty internal code operation.
Note that any kind of DRM management would be impossible in such a model. GOOD!
But that is why TPTB will not develop such an OS.
