The release in the last few days of malware for the Mac and Linux underscore some old issues about how it is possible to have malware on those platforms. I have some new thoughts though. Ive begun to wonder what Apple would do if a real problem developed.
To be very clear, a real problem has not yet developed, and Inqtana.A and Leap.A are not a real problem, except to the extent that they may be bellwethers. They are more interesting for what they suggest than what they actually do.
Its true that the Mac OS has had, for many years, an important level of protection that will only emerge in Windows with Vista: By default, Mac users run at a restricted privilege level, and any malware they run will be similarly restricted.
The user may be presented, as with any legitimate software install, with a request for the Admin password, and at that point they may exercise discretion.
This is almost entirely a consumer issue I believe; in managed business networks it has been well understood for a very long time how to create Windows clients with restricted privileges, and any administrator who doesnt do it is to blame for problems that result.
This process protects against silent installations of malware, but it doesnt protect against all fraudulent installations.
Much of the garden-variety malware on Windows pretends to be a data file of some sort (as if data files were inherently safe, but thats another story). Double-click the icon and it turns out that youve actually run a program.
But much malware, especially of the adware variety, doesnt hide the fact that its a program. The user is told that they are installing some slick browser toolbar or perhaps a special viewing program for some porn, but in fact they could be installing anything at all.
And even if they were on Windows Vista and not initially allowed to install the program they would provide the Administrator password to install it because they know they intend to install a program.
When good social engineering attacks are developed for the Mac, the same thing will happen. Its not hard to imagine Web sites and e-mails offering programs for the Mac that do more than they claim to do.
Just in terms of adware, there may be some benefit to being able to deliver known Mac users to advertisers, but for the most part the "value" of infecting the user is the same: to spread itself, and perhaps to create a Mac botnet.
The small size of the Mac community has protected it so far even more effectively than the security features of the software. Few have even tried to write malware for OS X, although in earlier days of Mac OS, back when we still passed floppy disks around, it was a frequent virus target.
I still think this has to be an important limiting factor on malware for the Mac. Since Im talking about attack methods that are universal, most attackers will want to go after the larger audience.