Michaël Michaud

OK. In CoordinateArraySequence and CoordinateArraySequenceFactory, there are some documentation and tests verifying that dimension is not > 3. Ex. https://github.com/locationtech/jts/blob/4804f760a1e58c1991dddb2967bd7dd0bed423e1/modules/core/src/main/java/org/locationtech/jts/geom/impl/CoordinateArraySequenceFactory.java#L65-L73 Should I remove this limitation and try to make it...

Hi Martin, Agree that the definition of dimension and measure are confusing. After all, it seems to me that throwing exception in builders everytime d > 3 (builder with d...

Maybe this paper is worth some consideration : https://agile-online.org/conference_paper/cds/agile_2012/proceedings/papers/paper_ledoux_automatically_repairing_invalid_polygons_with_a_constrained_triangulation_2012.pdf

Didn't JTS already include a CDT algorithm ? There is a repo on github with a few examples of invalid polygons: https://github.com/tudelft3d/prepair The approach you propose seems not too far...

I know JTS is a 2D library by design and z (or m) values will always have a place appart, but Coordinate has a z attribute and it is the...

I thing the first change (Adding the point to the graph) is worthwhile even if touches still does not return an intuitive result (may also be the case for other...

I will add a comment. I've tried to change the returned dimension of LineString and Polygon when they collapse to a single point, but I've also checked the whole test...

I had to go a little further to keep all tests positives. Beside the change in GeometryGraph, I changed - the returned dimension of collapsed geometries (0 in case of...

I wanted to explore your proposition to change the dimension during operations instead of changing the semantic of getDimension, but in IntersectionMatrix, all public methods where dimension has an impact...

About polygon collapsing : maybe we can handle the simple cases by traversing all segments of the outer ring from the starting point in cw and in ccw and check...