On Fri, May 24, 2024 at 1:21 PM Dale Schumacher
<
dale.sc...@gmail.com> wrote:
>
> Part of the Friam discussion today examined this issue. Some took the position that functions could not be safely communicated via serialization/deserialization. Others wondered where the risk was, and if it could be mitigated by making closed-over data immutable.
What do you mean by "safely"?
- That the data will reach the recipient without being snooped on or
modified? Seems like cryptography handles that.
- That the recipient will not be able to alter it? If the recipient
can only execute in the context of a runtime, then the runtime can
handle that; otherwise it seems impossible.
- That if the function closes over state, the state will not be shared
properly among all instances? Whether this is possible depends on the
guarantees you want about access to the state.
--
Mike Stay -
meta...@gmail.com
https://math.ucr.edu/~mike
https://reperiendi.wordpress.com