I was wondering how someone would go about adding Capability-based security to Erlang without making it into a new language and preserving maximum backwards compatibility on the application side?
If you really are interested in this you should read the Master's thesis by Gustaf Naeser on Safe Erlang