There is just over a year to go on the Python 2.7 death clock
Have Mcneel plans moved on at all in terms of Python 3 support? Two major features of Python are the large standard library and the large number of available technical and scientific libraries.
This has really accelerated over the last 2 years with some great tools in recent versions of Python 3 so supporting Python 3 would give Rhino developers access to some amazing machine learning, data visualisation and other tools. There are also lots of tools and support for switching code from 2 to 3.
I see the 3dm project is py3 compatible but is that a moonshot? is there an official strategy for Python 3 support in Rhino and Grasshopper?
By the way : thank you very much for integrating Python in Rhino! This makes my working life much less repetitive and tedious and I feel like I am just starting to explore the possibilities
I know that Rhino currently uses IronPython, and I see that that project picked up life again earlier in 2018 after a long hiatus. I am also aware of the Python .Net project which supports Python 3 but donāt know if that is suitable for Rhino integrationā¦?
I wanted that py3 compatibility for Blender 2.8, as I wanted to start import_3dm as a new Blender add-on. the rhino3dm more-or-less lives outside Rhino itself, so no direct influence from the py3 compatibility from that to the main Rhino project.
I will have to try rhino3dm outside Rhino - I sometimes need Pandas dataframes and I am targetting a Rhino ā Matlab ā Paraview toolchain. Iām a bit nervous about the newness of it all - I donāt want to have to rewrite everything if the API changes.
I donāt forsee large, architectural, changes in the rhino3dm API, especially considering that the Rhino file format is quite āflatā. I think @stevebaer probably will agree with that.
Now the compute/inside APIs I donāt know how stable they are, but I think you should be able to get quite far with already ājustā the file format module.
Nathanās right; I doubt weāll make a whole lot of changes as this is simply writing OpenNURBS bindings in python and I am following the patterns that we have used in RhinoCommon. Iām not one to go dramatically changing things just for changeās sake.
OK thanks both - I shall get stuck in next time Iām linking data to geometry. Iām still not sure how best to fit all the pieces together but this could be the missing elementā¦
Latest update on Ironpython 3 development from those who are working (unpaid, open source) on it: āI wouldnāt say itās ready for any kind of release, but it works for a lot of scenariosā
So if people are feeling brave then I guess they could download, (compile?) and install it outside of Rhino.
Looks like theyāre implementing Python.NET, which Iām sure comes with its own can of worms (and is another potentially brittle dependency ala the dead Enthought stuff). But itās cool to see them attempt to provide official CPython 3 support out of the box.
given 
