JVM bytecode is verifiably safe (AFAIK all exploits have involved convincing a faulty JVM to run your malicious bytecode without verifying it), which allows for incredibly fine-grained security. Does Taos do this, or have a capability model, or does it have the same old memory-protection overhead and "everything you run gets all your privileges" problems?
Slashdot Mirror
JVM bytecode is verifiably safe (AFAIK all exploits have involved convincing a faulty JVM to run your malicious bytecode without verifying it), which allows for incredibly fine-grained security. Does Taos do this, or have a capability model, or does it have the same old memory-protection overhead and "everything you run gets all your privileges" problems?