As most of us know, the boolean functions (difference, intersection, etc.) should be used sparingly due to the extra time they use during processing. It is also known that it is good practice to make the elements involved should be made slightly larger to avoid edges touching edges (for example, increasing the height of a cylinder when making a hole with difference). I have a polygon from which I want to remove several areas that touch the edges. The reason I am using paths instead of modifying the perimeter is because the area being removed actually touches opposite edges, basically splitting the polygon into multiple pieces. The two solutions I have come up with are:

Use difference(), but as mentioned earlier, this should be used sparingly
2.
Create multiple polygons that do not require me to create extra paths to subtract

Is there a way to avoid this problem? I am wondering if this is possible because paths are specific to polygon, along with the fact that they are 2d instead of 3d.

Nathan Sokalski
njsokalski@hotmail.commailto:njsokalski@hotmail.com

It's important to know that the 2D boolean functions are a great deal
faster than their 3D equivalents.  I wouldn't worry about performance
with them at all.

And even the 3D functions aren't that bad.  They weren't that bad
with CGAL, and they are much faster with Manifold.  Yes, they take time,
but I suspect no more than a union does - and if they're what your
project needs, it's better to have your project be clearer and easier to
understand than to minimize its render time.

Mostly, don't worry about performance until it's a problem.  It isn't
quite true that "premature optimization is the root of all evil", but
especially in the 3D printing world the tradeoffs can be inverted from
traditional software development.  How many times are you going to
render your project?  It depends on the project, of course, but mostly
it's going to be a lot of previews during development, and then a
handful of actual renders to make and tune the production STL... and
then zero.  It's not something that a user is going to do every day. 
Again, there are exceptions, but I think this is true for most projects.

By all means adopt patterns that improve performance, where they don't
conflict with understandability.  If you're going to make a block with a
bunch of vertical holes in it, do everything in 2D and then extrude into
3D, because that's a lot faster than doing it in 3D in the first place -
and, critical, it's not any less understandable.  But if you are
hand-building a polyhedron with a hole in it, to avoid a difference, I
think you're making the wrong tradeoff.