Open breannansmith opened 10 years ago
Cork truncates precision internally before performing arithmetic so that it can use big-integer arithmetic when needed.
Can you clarify why a double-precision interface would be useful? If it's a convenience thing, that's not necessarily hard to implement. If there are real concerns about the output precision, then it would be helpful to have more context.
Is there any reason not to include a double precision interface to Cork (e.g. a double precision variant of CorkTriMesh)? A quick glance under the hood suggests that each input CorkTriMesh is converted to a CorkMesh, which stores vertices in double precision, internally.