In the what aspects is Julia weak in thread, the topic of ongoing support for DelaunayTriangulation.jl was raised and prompted a good amount of subsequent discussion:
I’ll move some of the posts from there into this topic and let’s centralize the discussion here.
We are already pooling it into the Meshes.jl efforts. The package was moved to JuliaGeometry and we are discussing the integration in the following months.
EDIT: so to my point, the same faces (me and you) maintaining more and more packages…
Ehh but my crew is also growing as I get older so decent. Again could always be better, but we are on the verge of decent.
Good to hear you’re taking this one. I want to have a call that includes some more of the PDE folks (Trixi.jl, Gridap.jl, Ferrite.jl) and really map out the ModelingToolkit general PDESystem for arbitrary meshes in the near future. This space has a lot of people and I want to find out how to best connect it all.
That call is really necessary. We’ve been working on Meshes.jl for years already, polishing it on various aspects that are not just PDEs. Ideally we will be able to use the same mesh types under the hood to connect the PDE world with the geospatial data science world in GeoStats.jl, which is working great in industry.
Do we want to do a Birds of a Feather meeting for Meshes.jl and PDEs at JuliaCon? Did anyone submit that?
@mkitti I never organized a BoF, don’t know the responsibilities of organizing one. How does it usually work? We enter into a video call to discuss software design or is it more high-level discussion about the requirements of each of these worlds?
Good point, I forgot to submit it.
Feel free to send an invite to a call anytime. We missed JuliaCon, but can shape these requirements when everyone is free.
As far as I understand, a BoF is just of meeting of like minded people on a topic of current interest. It’s meant to be a free form discussion.
Anything dealing with Meshes.jl or GeoStats.jl is certainly welcome within the Scientific Data Minisymposium at JuliaCon. As organizer of the Scientific Data Minisymposium, I have discretion to invite talks outside of the CFP process. The HPC Minisymposium is actually composed entirely of invited talks.
I understand Chris wants to have a more focused conversation about PDEs that may not quite fit within a data minisymposium. As far as I’m concerned though, a BoF can go in any direction that its participants want.
I would be interested in that BoF for sure.
Regarding maintenance of DelaunayTriangulation.jl: If anyone else does want to assist with its maintenance, let me know. Probably doing so over on Zulip in the meshes.jl stream will be good so that @juliohm also knows. I will be writing up a document kind of going over the package and its design/internals + some unreleased work to assist with the integration into JuliaGeometry that can also help any future dedicated maintainer.
I could be another maintainer if that’s still needed. 