Chuck,
Yes, there has been a significant change since a decade ago. In those days, I was using NEC4WIN95, which is based on MiniNEC-3. Although NEC2 gave better results with wires close to ground, its card-format input scheme and lack of a convenient data editing feature made it unwieldy for quick turnaround modeling, and its antenna gain output was not directly usable in ICEPAC. I switched to 4NEC2, published about 6 years ago, because it added a 3-D graphical user interface and a spreadsheet geometry editor to NEC2, and generates 3-D antenna gain tables for direct use in ICEPAC. I mentioned the transition to 4NEC2 in my 2006 paper "Harmony and Power".
The wire frame model comprises 1069 wires which, at 3105 kHz, happens to be the number of segments. Of course, as you noted, the number of segments increases with frequency.
As for insulators, my intention was merely to convey that there were insulators at those three locations, the point being that the wire section aft of the wire junction on the starboard side was not inactive.
Bob