USER FORUM
(you are viewing a thread; or go back to list of threads)
Identifying objects acted on my constraints (by Alex Yeilding)
When fixing an over-constrained model (in a plane if that makes a difference to the answer), I go through the list of solutions suggested in the text window, and see the constraints highlighted in the drawing window as I move through the text. However, sometimes I can't tell exactly what constraint I want to undo.
Eg, I have a vertical line A
Line B has an endpoint on line A
Lines C and D also have endpoints that are constrained to the same point on A
Going through the text list of constraints that might solve the over constrained condition, I see one point on line constraint (which I know from history is endpoint of B to A, but couldn't tell that without recording that history) and two point-on-point constraints (but I can't tell which constrains line C and which constrains line D, or if I don't remember how i created them, which constrains line B.
Is there a way to tell which constraint constrains which objects?
Eg, I have a vertical line A
Line B has an endpoint on line A
Lines C and D also have endpoints that are constrained to the same point on A
Going through the text list of constraints that might solve the over constrained condition, I see one point on line constraint (which I know from history is endpoint of B to A, but couldn't tell that without recording that history) and two point-on-point constraints (but I can't tell which constrains line C and which constrains line D, or if I don't remember how i created them, which constrains line B.
Is there a way to tell which constraint constrains which objects?
(no subject) (by Jonathan Westhues)
Interesting question. As you say, the constraints themselves are highlighted as they're drawn on the sketch; but in the cases that you describe, the constraints are drawn identically. So there's no way to distinguish among them just from that.
I suppose that I could also highlight the entities referenced by that constraint, perhaps in a different color. I'm not sure whether this case is common enough to justify that, though.
I suppose that I could also highlight the entities referenced by that constraint, perhaps in a different color. I'm not sure whether this case is common enough to justify that, though.
(no subject) (by pedrodude)
I came to ask about something else then saw this, I'd definitely have a use for some contextual highlighting exactly as you describe. Perhaps as an option that becomes available in the browser upon selecting a constraint.
Regards,
Regards,
Graphic solution would be nice, but text meets the need (by Alex Yeilding)
I think what Jonathan suggested would be nice, but I was thinking the same thing as pedrodude. In fact, that could be more powerful. If I have bound endpoints of lines B, C, and D to Line A, just showing the bound objects doesn't tell me enough -- all three will just be points to the line. I need to know what else those points are bound to (i.e., the line segments of which they are the endpoints.
I'd love it if you could step the constraints one step at a time, the way you can step the dependencies in an Excel sheet. you could eventually see the full chain of constraints that lead to the redundant or inconsistent one you just added, and this could work very well graphically, I would think.
I'd love it if you could step the constraints one step at a time, the way you can step the dependencies in an Excel sheet. you could eventually see the full chain of constraints that lead to the redundant or inconsistent one you just added, and this could work very well graphically, I would think.
(no subject) (by Jonathan Westhues)
Such a feature could be useful. Note that dependencies in SolveSpace (or other similar constraint solvers) are fundamentally different from dependencies in Excel, though. A circular reference in a spreadsheet is generally an error, but most sketches will contain cyclic dependencies. The point of the constraint solver is to solve in spite of that, like a geometrically-conscious multivariate "goal seek".
Post a reply to this comment: