search

anowacki / Geodesics.jl

Calculate great circles on a flattened sphere in Julia
Other
10 stars 5 forks source link

Register package? #3

Open anowacki opened 5 years ago

anowacki commented 5 years ago

In #2, @tclements would like to depend on Geodesics.jl. This would required registering this package, and probably changing the name.

anowacki commented 5 years ago

Possible alternative names for the package if we register it:

GreatCircle[s].jl (also pretty general?) Vincenty.jl (scope too narrow?—I'll happily improve the algorithm) SphericalGeodesics.jl

tclements commented 5 years ago

Apologies on the delay, just got back from fieldwork. SphericalGeodesics.jl could work or maybe GeoDistance.jl?

On Thu, Jul 11, 2019 at 07:40 Andy Nowacki notifications@github.com wrote:

Possible alternative names for the package if we register it:

GreatCircle[s].jl (also pretty general?) Vincenty.jl (scope too narrow?—I'll happily improve the algorithm) SphericalGeodesics.jl

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/anowacki/Geodesics.jl/issues/3?email_source=notifications&email_token=AE36WXF7UWDKDYRBKG6XNVTP64LY7A5CNFSM4IBBKRHKYY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGODZWNV7A#issuecomment-510450428, or mute the thread https://github.com/notifications/unsubscribe-auth/AE36WXDCWS5334GH6EVB57DP64LY7ANCNFSM4IBBKRHA .

-- Tim Clements PhD Candidate Earth and Planetary Sciences Harvard University

lazarusA commented 2 years ago

Hello, I had been using this package for some great circle calculations. Any chance of registering the package and bumping compat to the latest Julia. As it is it works just fine in Julia 1.7.2. Thanks.

anowacki commented 2 years ago

Hi @lazarusA. Great to hear this is useful! I think if this is proving helpful I'm prepared to register it, or really a package which combines it with GeographicLib.jl and gives the user the choice of which algorithm to use. There are a couple of wrinkles which I want to iron out on the way:

  1. I really do think 'Geodesics' is too general a name. @tclements's suggestion above of GeoDistances is good, but opinions on that or other suggestions are very welcome.
  2. I don't think Vincenty's algorithm is really good enough as the default algorithm for most people because of the near-antipodal problems it has.

See anowacki/GeographicLib.jl#2 for a summary of why I have been reticent until now to register this. But I think that those are no longer stoppers.

However, this leaves me with a couple of options on how to proceed:

a. Copy the code from GeographicLib.jl into here, and improve that in time. Has the advantage that it can be totally rewritten without affecting the public (more Julian) API available here. b. Register GeographicLib.jl, and depend on that here. However the former cannot change its interface without breaking compatibility and hence needing a version bump.

I am reluctant to register GeographicLib.jl because I think the name implies that I am committing to keeping up with improvements to Charles Karney's code, and I don't want to do that (or imply it's related to his work in any way). I think I prefer there to be a Julia spherical geodesic distance package.

I've collected what I think is needed to be done before I can register this in #8. Any contributions are welcome.

tclements commented 2 years ago

This is great! I like option a. A Julia geodesic distance package, imo, will be easier to maintain and interop with other Julia packages. Could check with the JuliaGeo folks what they think about the name GeoDistances? @c42f @visr The other option would be to include this as a submodule within Geodesy.jl (https://github.com/JuliaGeo/Geodesy.jl/issues/40).