Arc tool preview magenta inference from a node point

Any idea what the logic is behind the arc tool previewing with an arc or a straight line? It appears to default to straight line if the arc is started at a node. This appears to make it impossible to get a magenta (or light blue) tangent inference if your intended arc begins at a node. Is there a trick to this I have yet to discover? It’s a silly little thing I know, easily overcome, but it was bugging me tonight, happy to be enlightened.

1 Like

I find it mildly annoying too, but I think the logic is it is a point in 3d space and it doesn’t have a fixed direction to be tangent to, which is what the teal and magenta inferences are.


Yeah, that’a a good point (ha ha)… can’t be tangent to a point. But it does seem that the inference engine could sort that out from the plane you are glued to and the adjacent edges. As you said, mildly annoying, thanks for commiserating.

1 Like

That is not quite right. Each node is the endpoint of the line. At the points where the lines meet or intersect, the node is formed from the endpoints of all the lines joining it. The program itself cannot decide which one to use as an inference and from which direction to build the arch. This problem does not exist for line lonely endpoints or straight line division points.




a point at which lines or pathways intersect or branch; a central or connecting point.

@endlessfix More appropriate for this case:


[ nohd ]
Geometry. a point on a curve or surface at which there can be more than one tangent line or tangent plane.

Fair enough, I didn’t mean to argue the point. I get that the inference engine distinguished between the two definitions. You are right that the arc inference recognizes endpoints as possible to tangent from, but not intersections. I just think that if the curser is glued to an adjacent plane after setting a start point that the inference engine should have enough information to successfully tangent from an intersection as well. Its not a big deal, just a quirk.


This is the case, but the absence of a cyan or magenta tangent line does not indicate non-contact with the plane. This indicates that the program does not understand which of the several lines joining the starting node should be tangent.

I get it, I understand how the program currently works. My point (ha) is that it could be improved. Once starting an arc at a node or intersection, if the curser is placed on an adjacent connected plane, any one of them, the inference engine should have enough information (starting point and edge direction) to look for tangents just the same as if an arc were started mid line. This video shows first the way it currently works, and second how I believe it should work.

As I said it’s not a big deal, it’s way, way down on my FR list, just a thought.

1 Like

@endlessfix I can imagine enough real situations where I don’t want the Arc tool to behave that way, BUT… developing your idea, maybe it is possible to combine these two options? For example, with a modifier similar to Alt for the Line tool in the latest SU version. What do you say about that?

Sure, I guess it could be a modifier key option. But why? It’s not stopping one from clicking anywhere else, and it still shows global axis and other inferences as well. Can you show me an example where it would be a problem to help me understand?

Yes. For example, if I want to draw an arch on a face from a corner point with a specified distance, direction and radius (or curvature) that is not tangent to any of the edges.

This can be done much faster and more accurately with the 2 point Arc tool (see drawing above) than by constructing the center of the arch from the endpoints and then using the Arc tool (see drawing below).

What is preventing you to do this even when ‘Tangent to edge’ is showing up: drawing an arc with chord of specific length in desired direction and with specified radius? (also from midpoints or any location). To overrule the arc being tangent to…
So I see nothing against @endlessfix’s suggestion so far.

Maybe. However, I prefer the additional option of not having to deal with any kind of unnecessary snapping in any situation. Much like it was with the Line Tool I mentioned.

So how would you specify the desired direction of the chord? If there’s nothing there?

It depends on the specific design. In most cases with guide.

I think there is nothing to discuss here. Working methods for achieving the same for each user will vary depending on the specifics of the work process. I am very happy with the option @endlessfix is proposing, as long as it is technically possible to provide it, but I do not want the new option to completely replace the old one.

Couldn’t agree more.

In my lifetime, I have experienced enough changes made by software developers, where when creating something new, easier to use for inexperienced users, the old, well-functioning one is lost. The leader in this respect is Microsoft. So don’t be angry that I am so conservative and suspicious of every innovation.