When I last looked (probably v0.4.0), it appeared that the code sent to a set of nodes sequentially, which could presumably be very fast or very slow (depending on the connect/read timeouts configured, and the state of the other nodes at the time). Is there any reason why this can't be done as parallel tasks?
Yes, parallelization is more sophisticated. Current implementation is called Nxt
Reference Software. We could optimize networking by using binary protocol instead of JSON, but this would increase entry barrier for newbie programmers. The same is applied to other parts of NRS.
Maybe it would be good to have 2 types of protocols JSON for local communication
and binary for remote communication or maybe use
data compression while sending data.
For example you or the developer that would continue work on the project
could create a procedure or library to convert JSON data to binary data.
What are your thoughts about this?!