Using the new curve option for Point Deviation, I noticed that the evaluation does not update when manipulating the input curve. Is that intended?
Furthermore I would like to propose adding visible Min And Max indicators to the tool. That way the tool could also be used to take actual measurements!
The points test statistics even show a Maximum Deviation. But only after closing the tool. Having that in real time would be great help for us CV pushers!
Best regards!
Hello - here the panel updates when I move control points and indicates max and min deviation (in the panel).
I get about the max and min dots at the appropriate locations - we can look at adding that but the max and min should be updating as you go, in the panelā¦
-Pascal
Hm, it might be my command window. The whole bottom part is not visible for me. I used the last WIP from June 30th. I will look into that again.
Thank you, the Min and Max indicators would be a great addition.
Hello Pascal,
I tried again today and found that the analysis does indeed not update when manipulating the curve. Strangely it does update when moving the surface! I tried some variations, like a planar surface and a circle, to see if that makes a difference, but was still the same.
Can you reproduce that weird behavior?
Hello - the curve that gets sampled for points does not update - what happens is, you pick say one curve, that you want to edit, and one that you want to check against. The second curve is sampled for points - these are then static, and are measured against the first curve, the one that you edit. Monkeying with that curve should update as you go - do you see that?
-Pascal
Iām sorry I do not understand what you mean. Where does the second curve come from?
I am probably using the tool in a wrong way, but what Iām trying do is manipulating a (blend) curve to match as closely as possible to the big vertical fillet (in the pic above), then using the new TrimRefit command, I build a linear matched corner fillet. That is the standard Class-A way of filleting. It is the most reliable method of approximating that big fillet surface, while still using a single span curve for a clean linear match.
Hello - whatever you are measuring against just provides points - - in the example, youād need to get the points from the surface (render mesh) or get them from an extracted isocurve of that vertical fillet. The curve you are manipulating then is measured against those points, if you see what I mean.
-Pascal
Yes I understand that. I picked the curve first and then the surface to measure against.
I understand the tool this way: The preset number of points are generated on the first curve. Those points are then projected in normal direction onto the object I want to measure against, and those two sets of points are what actually gets measured. Each point against its projected counterpart. Is that right?
I think the missing bit is that the points are static - they do not travel with the geometry that provides them, so you cannot get them from the curve you want to edit. Does that make sense? I see what you are getting at and I see how that could be a good thing but it does not work that way now.
-Pascal
Now I understand. However the logic behind that tool having two sets of points, one static and one dynamic, is already there, it obviously works. Couldnāt this just be swapped? Not being able to use the tool that way keeps Rhino from having high quality fillets unfortunately.
Hello - for now, use an extracted isocurve or a projected curve to provide the points:
Red is my curve that I want to adjust, the green is sitting on the surface - it has provided the points.
-Pascal
The problem with ācorrectā filleting is, that this curve never is an isocurve. Thats is why you use a blend curve and TrimRefit. That way it ābecomesā an Isocurve. The parameterization of the surface is changed, but the āshapeā stays the same. This way the TrimRefit Surface becomes a substitute for the original Surface. That is the whole āTrickā behind that technique.
Right. Iām trying to help you get a curve that is closer to the surface.
-Pascal
Thank you very much Pascal!
This approach is also the reason why I was so excited about all your Fillet Enhancements and the new TrimRefit. command. It opens up a whole new chapter of surfacing possibilities.
FWIW, what I do is get my blend pretty close, Pull it to the surface and then use the resulting curve as the point-provider, if it is not an isoā¦ when the fillet has the same structure - degree 5/6 say, then itās point distribution is of course a good indication of where to put the blend curve points.
-Pascal
Yes I see. But that way you get a curve that sits perfecty on the surface underneath, but will likely have tangent deviation. Using the Point distribution of the surface as an indicator is the the right way btw!
The problem is that you need the"green" curve to get the āgreenā curve.
And without measurements to keep everything within tolerance, we are running in circles.
The trick to Class-A surfacing is that everything always gets measured. All those little tools are used in conjunction with each other.
Unfortunatly Rhino is missing the last 5% of those little tools that keeps it from getting 80% of the quality.
I know Rhino always gets close. The surfacing tools are great, but it is the match quality and the analysis tools that keep Rhino back in topology quality.
@Cadworx - just to see if something useful might come of it - here is a first stab at a curve >surface deviation checker. This currently does, assuming it works, the following:
Run the script:
Prompts for a curve and a surface - it then shows max and min clearance at these locationsā 32 samples for now. (Negative numbers indicate it is on the āfarā side of the surface normal. I imagine it might be good to change dot or text colors for negatives.)
Run the script again:
Turns it all off. Youāll need to re-select to get it running again.
Eventually, assuming this is somewhat correct in real life, we can make it remember the selection, draw hairs etc.
CurveDeviationDrawingTest.py (6.5 KB)
To use the Python script use RunPythonScript, or a macro:
_-RunPythonScript "Full path to py file inside double-quotes"
-Pascal
Thank you so much Pascal, Iām so happy that you take care of this! Unfortunately I cannot test this before Monday at work ā¦ 
My initial hope was to have a dedicated āDistance Analysisā in Rhino, using Curves, Edges and Surfaces interchangeably, as one tool that just returns Min and Max Distances with the only settings being the point count and the hair scale.
However thanks to you the Edge to Edge case is already covered with the Min value addition to the Edge Continuity Analysis. Now this Script covers the Curve to Surface case.
The only two cases left would be Edge to Surface Distance and Edge to Curve Distance. Could your Script be tweaked to also measure these combinations? The logic is always the same in all those cases.
Hello - yes, all that should be possible - Iāll take a look. Meantime, here is an updated script to try.
It now finds and shows the max and min locations more exactly and uses the samples to draw the lines only.
ā
Curve-Curve:
Updated these and made a plug-in to make testing a bit easier -
DrawCurveDeviation.rhp (48 KB)
Youāll need to unblock the rhp in its Properties, then drag and drop onto Rhinoā¦ seems to work ok here.
DrawCurveSurfaceDeviation
DrawCurveCurveDeviation
DeviationDisplayOptions
Run either command to toggle it on and off.
-Pascal
I was in the same seat, but different use case:
You script seems to work great! I just have a question about the drawing of the green line that indicates max deviation (which, btw, should perhaps be blue to stay consistent with the EdgeContinuity tool):
It seems to be drawn sideways in the above image? What does that indicate?
EDIT: Btw, I use the .py files, becase nobody has told me a good way to do network deployment for .rhp files yetā¦
