How to discover SNMP devices

Question

I'm looking to set up a monitoring system and was reading quite a bit about SNMP. I don't yet have practical experience with it.

I didn't quite understand how an NMS discovers SNMP capable devices. To put it simply, what I gather is that it's a two step process:

1. Get some IP addresses/subnets
2. Send SNMP requests to these addresses (assuming knowledge of the correct community string) and wait for replys

My questions in this context:

1. Is this how it works? If not, then how it does?
2. Is possible to set an SNMP agent to send a trap to a multicast address?

Thanks