undefined | Better HN

0 pointsRetroTechie2y ago0 comments

Perhaps a better question to ask: why, in 2024, is it still possible that running a random app infects an operating system?

Don't we have cpu's with a whole set of hardware features to limit what <insert executable here> can or can't do? Don't we have OSes with fine grained permissions, VM's, capabilities, etc, etc? Weren't these things figured out like, in the late 70's?

0 comments

4 comments · 2 top-level

chuckadams2y ago· 2 in thread

> Don't we have OSes with fine grained permissions

No, we really don’t. We have OS’s with the tools to give us fine grained permissions, but at the end of the day on a desktop OS, any program you run can do any damn thing to any other damn thing owned by the same user. Look to phones for security boundaries that make sense in a post-1970s world, but not any shipping desktop OS.

vlod2y ago

Doesn't Qubes OS solve this problem? It is linux only though.

paulmd2y ago

Average users tend not to like operating inside heavily sandboxed environments that make it difficult to move data between apps. See: iPhone.

paulryanrogers2y ago

Because security isn't perfect and attackers need only one mistake. There were some early computers built with security in mind, yet their cost and performance were impractical for most uses.

There is also the problem of everyone having different needs, so permission profiles must vary widely. OS builders cannot assume anything about the limits of their use.

Finally users can be the weak link. They may not understand the risks of what the screen is asking them when prompted.

j / k navigate · click thread line to collapse