> In BoC, the equivalent of a message is received by multiple actors and operates with exclusive access to the message and all of the receivers.
(emphasis mine)
IIUC with actors, messages are processed by exactly one actor so it can be difficult to express transactions (e.g. transferring funds from A to B cannot be done atomically). Erlang somewhat fixes this with "selective receive" which re-introduces the possibility of deadlocks. BoC fixes both issues.
Was wondering how this is using a separate GIL per action when the actions are pure Python code. "PEP 684 introduces a per-interpreter GIL, so that sub-interpreters may now be created with a unique GIL per interpreter."
I was a bit curious to learn what the differences are between this and the actor model, and I found this lobste.rs discussion to be helpful: https://lobste.rs/s/gsjskz/behavior_oriented_concurrency_for
> In BoC, the equivalent of a message is received by multiple actors and operates with exclusive access to the message and all of the receivers.
(emphasis mine)
IIUC with actors, messages are processed by exactly one actor so it can be difficult to express transactions (e.g. transferring funds from A to B cannot be done atomically). Erlang somewhat fixes this with "selective receive" which re-introduces the possibility of deadlocks. BoC fixes both issues.
Was wondering how this is using a separate GIL per action when the actions are pure Python code. "PEP 684 introduces a per-interpreter GIL, so that sub-interpreters may now be created with a unique GIL per interpreter."
- stupid question: what is different about this concurrency compared to say asyncio?
Did you read it the paper at all? For one, asyncio is primarily still single-thread, while this can scale with number of processes.
the example is atrocious, seems like erlang actor model but for python.
It is Erlang inspired. They write
> One of the enabling technologies for BOC is Erlang-style send and selective receive.
And the example with match, receive, case is a few lines below. It's more or less what one would write in Erlang.
Is it just me or is the example they use for the locking extremely badly implemented?
Why would one ever result at that methodology?
It's researchers trying to contrive an example for why their research effort deserves a budget every year.
Of course their example will be detached from reality. But it doesn't matter.
The nested `with`s are weird, but otherwise what would you do differently? Seems like a standard pattern.
I still don't really see behaviors helping with this. Like yeah you avoid deadlocking, but now you're abstracting away mutexes :S