Security wasn't originally a low priority for Microsoft. I just wasn't a consideration at all. There was simply no method to differentiate programs rights at all, at least till Windows XP. The fool who decided that allowing ActiveX controls to run from web pages should be barred from the industry.

Linux/Unix is infinitely ahead of Windows are far as security goes. But even so, it is (as is Vista) flawed in design when it comes to the mainstream casual user.  The requirement for root or sudo elevation by users who have no hope of knowing the true nature of new software isn't good enough.

What needs to happen is for a totally new operating system to be created,  which for transitional purpose requires legacy operating systems to run in a sandbox under a hypervisor.  The only local crossover should be with non-executable data documents via dual filtered virtual devices.

THE NEW OPERATING SYSTEM

Software should only be installable via data packages by the operating system. The will be required to classify themselves singlely (game, document application, system utility, etc). Each shall have it's own unmodifiable application store, per user config and per user document access. No application would be allowed to create, modifiy or run executable files.

No application would be allowed to access or modify system configurations directly, and could only achieve this by requests to the operating system which would moderate user/program rights, and explicitly deny any request outside of the application's declared category.

All software packages could be subject to blacklist/whitelist checks by a trust provider as chosen by the user/administrator. System utilities, or other categories that need to modify the underlying OS, and thus create risk, would be required to be whitelisted by a trusted party. Not just a casual password.

All granted configuration requests shall exist as overlays to the system, on a per user basis, and thus removable, and movable (backup) by the operating system.

Yada Yada Yada. I could go on forever. The key is transition.

Food for thought...

Hello,

my two cents about security is, that some parts goes in the wrong direction. If a guy buy a bigger gun I must buy a bigger bulletproof vest? Why not take away the gun?

For example, someone send me an Email attachment, I open it, and get the trouble. I visit a website, click somewhere, and now....

Prevent Outlook or whatever, from open the attachment is only one side of the coin. I ask myself why eMails still using a 20 year old Pop3/SMTP protocol for the mass market. Same as HTTP stuff, why not make HTTP, javascript, active X,... on his own more secure?

Peter

Hey Roger, thanks for posting a comment on my blog. I have <a href="http://www.tssci-security.com/archives/2007/11/01/operating-systems-arent-any-more-secure-than-the-idiot-using-it/#comment-2470">posted a comment</a> in response and dre has also put a lot of thought into his reply as well. I hope you check them both out.

Did you actually read the blog postings you cited? You seem to have missed their message entirely. The common theme of all three wasn't "blaming and bashing" or "a blog war on which OS is more secure," and in fact they were quite the opposite. Those postings talked about why comparing OS security is mostly an empty and pointless exercise.

Not compare and “best of class” statements don’t mix well, in my opinion, but then I’m just nitpicking I suppose. I’m no security expert, so maybe I’m missing the point entirely, but as far as I can see, a lot (most) of the security threats have to do with the design choices of the OS in question. The amount of threats directly relates to the exposure of the OS, but the way the threats work is that they specifically target flaws in the OS in question. “Work together” could in my opinion only work on a governing level (help identify malicious emails, attachments, websites together, to minimize the time they go undetected, maybe even on a level of “trying to find the bastards who write the darn things”), but that’s only fighting the effects. The cause of the problem, insecure code, is a problem of the OS writers (unless ofcourse you’ll choose to open up the source ;-)). As long as the development of your OS stays behind closed doors, don’t be surprised you’ll be on you own fixing the errors as well.

Hey Roger, I just wanted to point you back to my blog for a couple follow-up comments, one from dre and one from David Rice, the author of Geekonomics (http://www.geekonomicsbook.com/).

Comments after you last reply begin here: http://www.tssci-security.com/?p=326#comment-2538

Cheers :)

-Marcin

It seems that I am not yet gone J . Eric Bidstrup, a colleague of mine, wrote a great blog post about