Earlier quoted context omitted.
So then you're up against the halting problem at the "digital border" and you've only reduced the problem to say one in 300 million. There are many differences. I already mentioned locality and scale. Another is that it's possible to make secure software (aka math) that precludes undesirable behavior a priori, whereas such thing is impossible in the real world.
A) It's not a halting problem. B) Digital borders exist all over the net. We use them every day to secure all sorts of things.
> Digital borders exist all over the net. We use them every day to secure all sorts of things
Erm, how do you square these two sentences?
I took your first comment to be arguing against software security in general, presumably in favor of more post-facto enforcement when people violated authorization boundaries.
Your response then seemed to focus on mitigating the cross-jurisdictional issues that make post-facto enforcement hard, by having some sort of software-based security enforcement at a "border", and then relying on post-facto enforcement inside of that.
Now you seem to be supporting software-based security in the form of firewalls everywhere?
If we continue along this trend to even more local, we'll get to fewer firewalls (because they aren't that good of a technology), with security pushed out to the edges. Which is where best practices seem to be headed (BeyondCorp, etc), but is directly antithetical to your initial comment.