-
Notifications
You must be signed in to change notification settings - Fork 133
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Overlap between Line2D and LineSegment2D #242
Comments
Possibly related to #35 |
I agree with your comment, that this is a bit messed up. For my understanding there are cases when one wants to consider a line segment (having a specific start and end), where I would say a line is infinite and independent of direction (eg. if you switch the direction vector of the line, it's still the same line). I think the core functionality should be part of Line2D and LineSegment2D should be more specific, something along these...lines (pun intended :)) Line2D
LineSegment (inherit from Line2D, but have a given Start and End)
|
A small bit of history, originally there was only Line2D/3D and they implemented the concept of a line from 1 point to another point, which was inaccurate, so LineSegment2D/3D was born to handle the point to point case and Line2D/3D was intended to be deprecated and eventually replaced with a model for an infinite line to allow the library to work with more pure geometry cases. However for various reasons this never happened. |
Thanks for the info. So if I read this correctly, you would also agree if LineXD covered the infinite case and LineSegmentXD the “start-end” scenario, correct? |
I already thought it was halfway a replacement. I was mostly wondering why it wasn't deprecated cause it took me (a new user) some time to figure out the current state of things and the difference between Line2D and LineSegment2D I'd definitely not make these classes inherit from each other as I don't think a segment should ever be treated as an infinite line or the other way around. Maybe a LineSegment.AsRay() => Ray(StartPoint, Direction) would be enough for most use cases. Apparently there is already an infinite 3d line called Ray3D, that you get from intersecting planes, which is a ThroughPoint and a Direction. Apparently there is also no UnitVector2D to do the same for a Ray2D 🤔 The equal checks if both the ThroughPoint and Direction are the same, which might also not be the expected behavior as #35 even implies that the direction orientation shouldn't even matter, let alone the ThroughPoint. Although calling it Ray might imply that the direction matters. But then what is the orientation of the Ray when you do an intersection of Planes? Is it then that doing A.IntersectionWith(B) = C that C.Direction = A.Normal x B.Normal? Intersection methods are also a bit inconsistent. Some use the bool TryXx(, out result) (LineSegment2D), some a nullable return values (Line2D) and others like Plane.IntersectionWith throw an exception (which I found out the hard way 😆). But that's another issue altogether. As I'm right now working on cropping and slicing functionality for triangulation, that heavily use the intersections, infinite lines, line segments and projections between 2D and 3D, I'm thinking about forking this library and doing a lot of breaking changes. |
That might indeed be the best option you have at the moment. Once I was accepted as a member of the team, I tried to clean up things a bit and keep evolving the library, but it faced considerable backlash from the community (#229) We could create a separate branch from main, where all these changes are implemented, so anybody has access to them in the future, without having to redo the work, but I need to find the time and primarily the motivation to do so. |
As a new user, the current state is a bit confusing.
LineSegment2D/3D is clear:
It is the segment between two points
Line2D/3D is unclear:
the class documentation is exactly the same as LineString, with the added "projections to" operations
Looking at Line2D, it has:
And equality is checked by start an.d end
So it's also a line segment
But IntersectWith, LineTo and ClosestPointTo (optionally) work as infinite lines.
As LineSegment2D is already split off, would it be better to also split off an infinite line and deprecate the Line2D/3D?
The text was updated successfully, but these errors were encountered: