HeelX: Blame this on my over-simplification. Thread TWO will be working on an array
of data that is SOMETIMES replaced by thread ONE.
Thread TWO will be re-reading (never writing) the array several times per update by thread ONE.

And both thread ONE and thread TWO both have many arrays to manage. So they will never (and I WANT it like this),
never be working on the same one at the same time.
And I only want one thread each for these two tasks, because they are either time-hungry(ONE) or CPU-hungry(two).
So by buffering data 'to be processed' by these threads, it lets me create a 'known' bottleneck that will allow me to
limit CPU/HDD utilisation by these tasks.


JustSid: Only just starting to read yours now...

[PS] They are actually models PRETENDING to be terrains.
And they will actually be hidden (INVISIBLE) during the vertex-adjustment phase, then swapped into replace the old one.