Put all the gigabytes and gigahertz to good use to simulate a very simple computer which be run backward an arbitrary amount. The entire computer, not just a given program.
Any bug in an open source system can be quickly understood and debugged, even rare sporadic failures. The full trace is available.
The simulated system might not even be x86 but rather an architecture designed to be easily debugged. Possibly need to design an operating system and user libraries from scratch. "Machine code" might be very high level and close to the source.
Weirdly, is the biggest problem privacy?
The easiest way to sidestep that problem is have the user debug. We need a way to make code more accessible, probably better programming languages and tools.
Theoretically, homomorphic encryption could help: someone else debugs why your browser crashed without learning anything about what you browsed... i am skeptical.
I feel we are getting no closer to this goal: things seem to just be getting more complicated to take advantage of more powerful computers.
Is the second biggest problem DRM?
No comments :
Post a Comment