Empathy List Archives

RP

Ronaldo Persiano

Tue, Aug 7, 2018 2:40 PM

Em ter, 7 de ago de 2018 12:27, Carsten wrote:

I used a different approach in my own code (not openSCAD), realising
that any convex polyhedron can be defined from just an array of vertex
coordinates.

A similar approach in OpenSCAD language might be:

module octahedron(r=1) {
verts = [ [r,0,0], [0,r,0], [-r,0,0],
[0, -r,0], [0,0,r], [0,0,-r] ] ;
hull( polyhedron(verts, [0,1,2,3,4,5]) );
}

Em ter, 7 de ago de 2018 12:27, Carsten wrote: > > I used a different approach in my own code (not openSCAD), realising > that any convex polyhedron can be defined from just an array of vertex > coordinates. > A similar approach in OpenSCAD language might be: module octahedron(r=1) { verts = [ [r,0,0], [0,r,0], [-r,0,0], [0, -r,0], [0,0,r], [0,0,-r] ] ; hull( polyhedron(verts, [0,1,2,3,4,5]) ); }

T

Troberg

Thu, Aug 9, 2018 5:50 AM

Yes, you can do it this way, although it is "cheating" as it relies on
the OpenSCAD discretization of a cylinder instead of defining the
polyhedron explicitly. It is more compact, but less readable.

I agree on the readability, but I wouldn't consider it cheating. It's a well
defined behaviour, it's not just some random effect of how it currently
works.

--
Sent from: http://forum.openscad.org/

> Yes, you can do it this way, although it is "cheating" as it relies on > the OpenSCAD discretization of a cylinder instead of defining the > polyhedron explicitly. It is more compact, but less readable. I agree on the readability, but I wouldn't consider it cheating. It's a well defined behaviour, it's not just some random effect of how it currently works. -- Sent from: http://forum.openscad.org/

A

arnholm@arnholm.org

Thu, Aug 9, 2018 8:56 AM

On 2018-08-09 07:50, Troberg wrote:

Yes, you can do it this way, although it is "cheating" as it relies on
the OpenSCAD discretization of a cylinder instead of defining the
polyhedron explicitly. It is more compact, but less readable.

I agree on the readability, but I wouldn't consider it cheating. It's a
well
defined behaviour, it's not just some random effect of how it currently
works.

That is why is said "cheating" (with quotes). It relies on
implementation details such as the discretization of a cylinder (or a
cone rather) plus it also relies on implicit union of 2 such objects.
Granted, this is well known current OpenSCAD behavior, but still.

It is also a special case that works only for creating an octahedron. If
you wanted a different kind of convex polyhedron you would be out of
luck with this approach.

But for sure it solved the original problem as presented.

Carsten Arnholm

On 2018-08-09 07:50, Troberg wrote: >> Yes, you can do it this way, although it is "cheating" as it relies on >> the OpenSCAD discretization of a cylinder instead of defining the >> polyhedron explicitly. It is more compact, but less readable. > > I agree on the readability, but I wouldn't consider it cheating. It's a > well > defined behaviour, it's not just some random effect of how it currently > works. That is why is said "cheating" (with quotes). It relies on implementation details such as the discretization of a cylinder (or a cone rather) plus it also relies on implicit union of 2 such objects. Granted, this is well known current OpenSCAD behavior, but still. It is also a special case that works only for creating an octahedron. If you wanted a different kind of convex polyhedron you would be out of luck with this approach. But for sure it solved the original problem as presented. Carsten Arnholm

discuss@lists.openscad.org

understanding the