The short version:
The original versions of the SmartNet controller envisioned two formats for SmartNet trunking.
In "PTT Trunking," the voice channel would be taken down as soon as the originating subscriber unkeyed. Any responses would require a new group call channel request and grant over the ISW/OSW path, with the ID of the responding unit contained in the data. The advantage of "PTT Trunking" is that all units participating in the exchange had to send their IDs to the controller. The disadvantage (at least from a public safety perspective) is that barge-ins are not permitted (since the responder's radio, when keyed, goes immediately to the data channel).
In "Message Trunking," the voice channel assignment dwells after the originating subscriber unkeys for a set duration, and during that time any other unit can talk through the voice channel repeater simply by sending RF on the voice channel input with the correct connect tone. Each such re-key resets the voice channel dwell timer. The advantage of "Message Trunking" is that barge-ins are permitted.
Public safety wanted both the capability of barging in and the each-unit-ID, so a revision to the controller brought out "PTT-ID Trunking." In this format, a responding (or barging in) subscriber goes to the control channel just long enough to ID, and then reverts to the voice channel, so long as the voice channel is still up.
If you are listening to a SmartNet voice channel and not on scan, whether you see the IDs of all participants depends on how the controller is programmed (and whether your radio is similarly programmed). If you are scanning a SmartNet II system programmed for "Priority Monitor," you may not see the current ID of a unit talking on a non-priority talkgroup if you join that conversation in mid-stream. You should see all succeeding IDs so long as your scan delay timer has not expired.
There is no explicit parameter in the PSR-500/-600 for selecting or detecting SmartNet trunking format, so I have no idea how the radio handles these issues.