user1234 wrote: ↑Sat Oct 01, 2022 3:58 pm
I would more generalize the request, constraint line parallel (offset) curves. But this was already requested.
Greetings
user1234
wrong thread? I don't see the correlation with concentric constraint.
IMO all we need is that coincidence command recognizes that when selecting circles we mean their center points. Side note: if this is done then that's a good reason to keep coincidence and point on object separate.
adrianinsaval wrote: ↑Sat Oct 01, 2022 6:30 pm
wrong thread? I don't see the correlation with concentric constraint.
No, is the right thread. A concentric constraint on a arc is the same as a parallel offset constraint of a arc. My suggestion is just a generalization of all types of geometry, as long is possible.
user1234 wrote: ↑Sat Oct 01, 2022 8:00 pm
No, is the right thread. A concentric constraint on a arc is the same as a parallel offset constraint of a arc. My suggestion is just a generalization of all types of geometry, as long is possible.
It's not really the same, concentric only implies that they share the same center, for an offset constraint you have to specify a distance between edges.
adrianinsaval wrote: ↑Sat Oct 01, 2022 8:34 pm
It's not really the same, concentric only implies that they share the same center, for an offset constraint you have to specify a distance between edges.
A offset of arcs always matches their center (they are always coincident), it is intrinsic of an offset on an arc.
user1234 wrote: ↑Sat Oct 01, 2022 8:39 pm
A offset of arcs always matches their center (they are always coincident), it is intrinsic of an offset on an arc.
Yes I understand that, but the request is for a concentric circle, which doesn't imply any fixed offset at all, so having an offset constraint doesn't really solve this request. Consider the case where you want the diameters to be driven by other constraints.
adrianinsaval wrote: ↑Sat Oct 01, 2022 9:06 pm
Yes I understand that, but the request is for a concentric circle, which doesn't imply any fixed offset at all, so having an offset constraint doesn't really solve this request.
When i read the request, it is absolutely the same and it implies it per the types of geometry of an arc/circle.
adrianinsaval wrote: ↑Sat Oct 01, 2022 9:06 pm
Consider the case where you want the diameters to be driven by other constraints.
That have nothing to do with that at all. You can still constrain it with an arc or diameter. The question is there is a value of distance. Never said there should be one, but yes should be also somehow handled, abut not per default. Imaging it like the parallel constraint, just more in generalized, not just lines.
2022-10-01 23-28-47.png (22.29 KiB) Viewed 692 times
There is more than lines and arcs. The offset of an ellipse (more general: a conic), is no longer an ellipse - although it looks like one. With conics I immediately agree that concentric means common focal points.
chrisb wrote: ↑Sat Oct 01, 2022 9:33 pm
There is more than lines and arcs. The offset of an ellipse (more general: a conic), is no longer an ellipse - although it looks like one. With conics I immediately agree that concentric means common focal points.
Yes the result can be an other type of geometry, depends of its source. An offset of an ellipse is no ellipse. This is why i wrote, as long is possible, i had should write, as long it is technical possible.
user1234 wrote: ↑Sat Oct 01, 2022 9:31 pm
That have nothing to do with that at all. You can still constrain it with an arc or diameter. The question is there is a value of distance. Never said there should be one, but yes should be also somehow handled, abut not per default. Imaging it like the parallel constraint, just more in generalized, not just lines.
I think you are deviating into unrelated things, concentric is a very simple concept and easily achievable (it's just coincidence constraint on the centers). Deviating from this objective is not a very good idea IMO. If you want a generalized "parallel curve" constraint that's a request on it's own, and a big one at that IMO, offset constraint requires some complex stuff on the solver so it will take much more work. Not too long ago someone already took a look at it and in the end decided not to go with that route because it involves a lot of work for little reward.
adrianinsaval wrote: ↑Sat Oct 01, 2022 10:10 pm
I think you are deviating into unrelated things, concentric is a very simple concept and easily achievable (it's just coincidence constraint on the centers). Deviating from this objective is not a very good idea IMO. If you want a generalized "parallel curve" constraint that's a request on it's own, and a big one at that IMO, offset constraint requires some complex stuff on the solver so it will take much more work.
I think you already know me enough to know that is not an easy task. But what speaks generally against it? It is not needed or suggested to support all geometry all at once, but adding support every few one type of geometry should be realistic. Yet there are only lines supported. Now if someone can do that, circles/arcs would be nice. And some years alter, conic geometries would be nice. Correct me if i am wrong, but i am sure the ellipse in the sketcher are internal pure conic with auxiliary geometry (yes in the element list there is written ellipse). If someone add an offset on it, it could be handled like when you add on a horizontal line a symmetry on a vertical line, which is also overconstraint.
