h2g2 The Hitchhiker's Guide to the Galaxy: Earth Edition

While I enjoyed the piece (thank you for that) and agree with the math, it occurs to me that a simple solution is now possible.

Many computers make use of dual (or more) processors.

Perhaps the best bet is to write a program on a computer with three processors where #1 processor monitors numbers #2 and #3, while the #2 is running the given program with the given input.

By having #3 'ghosting' the #2's processing (but several steps behind), #1 can be alerted to the halted program on #2 and instruct #3 to avoid the problem or stop the program.

#1's program could then continue to restore processor #2, and leave the current state in #3 for debugging.

A sort of real-time error-trapping.

Just a thought.

Unsurprisigly, such heartbeat monitoring methods are being used in the high-availability linux project.

although in that case, you use two machines, and a stonith (shoot the other node in the head) power supply. this type of power supply is a network controllable on off switch which you plug the machines into, and you have the surviving machine power down the failed one, and power it up again soon afterwards, at which time it spots that it has not shut down cleanly and runs recovery software.

this is partly what I meant in the other thread, when I said that the halting problem was getting less relevent.