Lisp to draw line between two lines - dividing space equally
@barry2104 wrote:
is it possible to extend this Lisp to be able to pick two curved polylines (e.g. two S-curves which run alongside eachother but not necessarily parallel to one another, but start and finish around the same chainage) and get the curved "midline" as a (polyline or likely rather spline) Output?....
You can use PathAverage.lsp with its PAv command, >here<. Read some of the subsequent posts for more considerations.
Kent Cooper, AIA 
You can also try CurveAxis - see:
https://www.cadforum.cz/cadforum_en/how-to-create-an-axis-of-two-curves-average-curve-tip11025
Vladimir Michl, www.cadstudio.cz
Using that lisp, Output looked good at first glance, but upon closer inspection it's obvious that this doesn't work accurately (at least not for my curved Situation)
Here the red line (Output) doesn't cross through the midpoint of the selected line, and for the curved part of the Output, lies obviously Closer to the top black curve. Why is this?
@barry2104 wrote:
.... the red line (Output) doesn't cross through the midpoint of the selected line, and for the curved part of the Output, lies obviously Closer to the top black curve. Why is this?
Assuming that routine works similarly to my PathAverage.lsp routine [link in Message 41 here], here's why:
It divides the two paths equally, and draws Lines between the corresponding division points [here, the green, though PathAverage doesn't actually draw the Points]. Here, the highlighted/gripped red Line is between the Points at the mid-point of the overall length of each path. The "contortion" in the relative shapes of the paths is what "throws off" the relative positions of equally-spaced locations along them.
So tell me, how did you come up with the locations at which to put the endpoints of the blue Line in your image, that the averaged path doesn't go through the midpoint of? It's not the shortest distance from either of its endpoints to the other path. Where is the distance between paths to be measured for a midpoint along it, and how? For more discussion on the difficulties of deciding the distance between two such paths, see >this<.
Kent Cooper, AIA
since the lines aren't parallel, one can not simply draw a line from one arc perpendicular to the other arc and take this midpoint, because this midpoint would vary depending on which of the arcs you started from.
Hence, to get that line, I drew a (yellow) line from the below Point 1 to (perpendicular to) 2, then continued in green to (perpendicular to) 3. Then i used a "middle line" Lisp to create the Magenta line between yellow and green, trimming it at the blue arc. The midpoint of this Magenta line (labelled M in the screenshot) is what the Lisp should calculate to depict a more accurate (nevertheless nicht perfect) representation of the midpoint line/arc.
The Lisp Looks to be currently flawed if it breaks each of the polylines up into the same number of segments - this gets less accurate with a larger difference in length between the two polylines.
Not sure how much effort it would be to use my calculation principle described above to rewrite the Lisp though...
You don't need to construct perpendiculars - for such cases the "rolling ball" algorithms are being used. Try to search for a Rolling Ball routine by John Leavines or by Lee Mac. The result of such routine:
Vladimir Michl, www.cadstudio.cz www.cadforum.cz
