> Perl makes a secure implementation language due to its tainting
> feature. It also makes a secure remote execution language due to
> its "Safe" and "opcode" modules.
Is it good enough that you are prepared to bet *everybody's* machine on the correctness of this safety box?
Personally, I wouldn't do that without a thoroughly examined semantics for the allegedly safe box.
Please understand: I'm aware that the Perl safety box has been in use for a long time and that it is quite good in practice. What I need to be satisfied that it is "safe enough" is a reference to a paper in which its safety has been rigorously and formally verified.
Jonathan