Actor Model: Difference between revisions
Adamcolton (talk | contribs) Created page with "I'm looking at the actor model and it seems interesting, but a little problematic because it's too recursive. I like this; an actor has 3 properties * Storage * Processing * ..." |
Adamcolton (talk | contribs) No edit summary |
||
| Line 9: | Line 9: | ||
One model has it that after sending a message, you can send any number of responses. I'm not sure I like this. Even the concept of a response gets complicated. | One model has it that after sending a message, you can send any number of responses. I'm not sure I like this. Even the concept of a response gets complicated. | ||
Another issue is that communication is complicated and this covers it up. There needs to be a way for any message to send itself over any channel. | |||
* passing reference in same program | |||
* socket | |||
* file | |||
* internet | |||
Questions | Questions | ||
* How does an actor stop? | * How does an actor stop? | ||
Latest revision as of 17:53, 21 September 2018
I'm looking at the actor model and it seems interesting, but a little problematic because it's too recursive.
I like this; an actor has 3 properties
- Storage
- Processing
- Communication
but I don't like the Communication model. Instead, a sort of Object with handlers seems good. But then there's the problem of what is communicated. I like the idea of messages being separate entities. So a message is storage and processing, but no communication. However, following the Rust model, when you send a message, you lose access to it, it is not shared.
One model has it that after sending a message, you can send any number of responses. I'm not sure I like this. Even the concept of a response gets complicated.
Another issue is that communication is complicated and this covers it up. There needs to be a way for any message to send itself over any channel.
- passing reference in same program
- socket
- file
- internet
Questions
- How does an actor stop?