(PPE: I see I didn't check for new page of thread when I started to reply!)
That is correct! Thanks Meph and Japa
I don't like that as much. What you have with that is the inner point of the ramp being the only 'steppable' access up the ramp from the bottom. If you tried to come straight south in the midde of the tile (also straight east) you'd come up a cambered slope and have half a Zed's step to climb up unless you moved somewhat to the corner then back to the midtile position upon crowning the slope.
The topology is
always going to be weird, because the first version of this that (to my eye) meets flat at top of the slope on both edges asks for the centre 'fold' in the ramp (see
Lego brick example, with the diagonal slope of the first like the diagonal slope of
this brick, but with the orthagonal slopes being built up to straight-sided abuttments) and presents a full triangular cliff to the transition
onto the ramp from the lower level.
What you maybe ideally (or better than either of these?) want is to have the half-way points flush with the surfaces they should be flush with by making it a compound slope that doesn't go out of joint with either end.
I was thinking the following. The strings of numbers between the *s are the orthogonally adjacent tiles (which match at all the midpoints), the *s themselves you can assume as the diagonal corners meeting the tile whose heightmap we are actually framing. The numbers are of course 0 at the lower level and 8 at the higher.
edited to change the edges. The Ls are the (flat) Lower level tile neighbour edges, assume @0, the Us are thr similarly flat and level but elevated Upper neighbouring tile edges, which you can assume are @8. Hopefully it looks slightly better.*LLLLLLLLL*
L000001234U
L000012345U
L000123456U
L001234567U
L012345678U
L123456788U
L234567888U
L345678888U
L456788888U
*UUUUUUUUU*
That means theoretically that if arriving at the slope from around the corner-cliff it provides access to (or descending to go round there) it might be thought more necessary to dodge sufficiently to the west/north side centre to get
onto the slope, but you basically do anyway (insofar as the granularity of moving entity positions is, anyway, on DF as it always has been).
Does this make sense, without actually digging up an art program of my own (which would take time to make available for viewingmon forum, sorry!) that might still not remain clear?
But it's just my thoughts, I'm sure the artists of this thread don't need my guidance in this. (And, while I'm having On-Topic thoughts, I can now properly apologise to the self-same artists for my somewhat off-topic jest the other day. I've dialled it back to Strictly Relevent, since then, among other things.)