Added geom classes based on doubles

[TPMoyer]

I need to hack about with geometry using double precision.
Forked your lib to add double versions of your geom classes.
Would like to share.

Your lib is much more replete than my legacy home-brew C library (1990's). I have converted over to Java/Scala, and Processing, so it seems like adopting use of toxicLib is WAY desireable.
BUT, am developing a CNC technology which is intended to have 30 meter span with 0.01 mm movement control. So I need doubles.

Know the continuing 32Bit speed advantage for graphics, even outside the GPU. Had written an OpenGL 3D app used by IBM failure analysis folks to fly thru the real 3D paths of the 100,000 wires in our mainframes. That app worked on a 486, so understand speed vs size. Believe there are also folks with application spaces where the trade-off of precision vs speed favors doubles.

Implementation details are the fork under Doubles_README.md

PS: Figure you would be one of the few who will appreciate an efficiency neither of us implemented.
We implemented a plane as a point and a normal (6 parameters), which is better than the plebeian 9 parameter: 3 points define a plane. I recently came upon the suggestion (no idea of the source) that a plane can be described in only 4 parameters: a normal and scalar. The scalar is the distance from the origin, along the normal, to the plane.

[postspectacular]

Hey @TPMoyer - thanks so much for this massive PR / addition. I had started on something like this separately (though never committed) just before I abandoned this project 8 years ago and am completely sold on the benefits of using doubles for digi fab use cases. All my later libs (in this field) are predominantly using doubles too...

The thing is I haven't touched this code base (or Java) for at least 4-5 years and it will take me some time to review all these additions (even though at first glance this all looks splendid!). So just a heads up & please bear with me...

As for your plane as vec4 comment: yep, that form is how I dealt with planes in these newer projects, e.g.

• Clojure: https://github.com/thi-ng/geom/blob/feature/no-org/src/thi/ng/geom/plane.cljc
• TypeScript: https://github.com/thi-ng/umbrella/blob/master/packages/geom/src/ctors/plane.ts

[TPMoyer]

No problem with timing, Was just hoping it would be seen as an attempt at art, not just graffiti.. Awoke with a concern about the Vec4 plane thing. (dreaming about shapes is a good night for me). A usual interest is a face on some closed polygon. For these cases, the Vec6 uses a local basepoint, so any imprecision in the normal is cantilevered over short distances only. Faces far from the origin would get off-plane by epsilon times the in-plane radius from the normal intersection.

…

On Sun, Oct 20, 2019 at 8:18 PM Karsten Schmidt ***@***.***> wrote: Hey @TPMoyer <https://github.com/TPMoyer> - thanks so much for this *massive* PR / addition. I had started on something like this separately (though never committed) just before I abandoned this project 8 years ago and am completely sold on the benefits of using doubles for digi fab use cases. All my later libs (in this field) are predominantly using doubles too... The thing is I haven't touched this code base (or Java) for at least 4-5 years and it will take me some time to review all these additions (even though at first glance this all looks splendid!). So just a heads up & please bear with me... As for your plane as vec4 comment: yep, that form is how I dealt with planes in these newer projects, e.g. - Clojure: https://github.com/thi-ng/geom/blob/feature/no-org/src/thi/ng/geom/plane.cljc - TypeScript: https://github.com/thi-ng/umbrella/blob/master/packages/geom/src/ctors/plane.ts — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#9?email_source=notifications&email_token=ABIOVQCEH22EOE72YVNGYEDQPTYMPA5CNFSM4JCWSGA2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOEBYXWEI#issuecomment-544307985>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ABIOVQCPRER7SIWR2XJF3VDQPTYMPANCNFSM4JCWSGAQ> .