I understand that SAFE will have a secure message service, and that because messages will not be persisted by the network they will be free to send. Is that correct?
If messages aren’t persisted then does that mean they will only be sendable to nodes that are connected and listening for messages, and otherwise dropped/forgotten if not received by the end recipient?
To whom does a message get sent? To a vault that is connected? How do you identify the “address” of a connected vault? Say I’m using some chat app, somehow my app has shared its address with the chat network and my peers know how to identify and contact me… so I’m talking with person A and person B, then B starts sending a constant stream of spam to me flooding my chat app/network… How do we prevent this sort of message spam being targeted at a specific user?
With centralized systems, the intermediary can throttle that sort of thing. But on SAFE, wouldn’t spam like this clog the network traffic?
The idea previously discussed (not finalised) is that to send a message you send it to a SAFE ID but the message itself is stored (you pay) and sits in your own limited outbox until the recipient accepts it.
So this limits spam - you pay to create the message and you are limited as to the amount of messages sent but not yet accepted (though you can delete them in order to make room to send others).
I don’t know how this would be implemented with the new data types, or if the scheme may be different now we have them, but I think that is what @Maidsafe were thinking of a while back.