Re: Digital self-destructing possibilities?

I find linear/substructural types very useful for a large variety of
applications.  These correspond more or less to 'use once' messages.

In particular, if I provide an affine capability, I'm saying you should use
it at most once. I can enforce this on the receiver side by disabling the
capability after first use. Relevant (at least once) capabilities aren't
really enforceable outside of the type system, but are still useful in
expressing expected protocol behavior. Linear capabilities are both affine
and relevant, use exactly once.

Note that I apply this concept to the capabilities, not to the messages. In
general, messages cannot self-destruct because they don't do any
computation. Of course, we could provide messages containing linear or
affine values, including more affine capabilities.

_______________________________________________
cap-talk mailing list
[email protected]
http://www.eros-os.org/mailman/listinfo/cap-talk