Feature Requests

[mileslucas]

Here are some features I'd love to see here:

• get the position angle between two coordinates
• Unitful.jl integration (this may require some restructuring of the core SkyCoords type
• Given a separation and position angle, calculate a skycoord from another
  • EG c1 = ICRSCoord(0, 0); c2 = offset(c1, sep, theta)
• FK4Coord (Add Ecliptic and FK4 coordinates #8)
• EclipticCoord (Add Ecliptic and FK4 coordinates #8, add EclipticCoords #46)
• AltAzCoord (or maybe HorizontalCoord)
• CartesianCoord (x, y, z, requires a known distance or can normalize to unit distance) (add CartesianCoords #45)
• String representation of coordinates
  • astropy does this really nicely, you specify a format string like "hmsdms" to get "hr min s, deg arcm arcs" just like simbad would show
• some more utilities for converting between hour angles, degrees, and radians

I'm happy to crack open some of these but some of these implementations would make good first issues for possible contributors.

[mileslucas]

So a big usability thing I've noticed when actually using SkyCoords recently is that humans don't work well with radians. Most of the errors as well as the time spent scratching my head recently has been simply formatting the SkyCoords into proper hour angle or degree formats for calculating offsets and for display.

This was really frustrating, because I know all of the hard spherical geometry is coded up really nicely, but the coords don't really feel usable. This has shown me that working on the string representations (and corresponding parsing methods) will benefit this package the most, currently.

For some ideas, I'm curious how Dates uses their DateFormat type. I think if we can nail down the design for a struct or two that provides a nice, clean interface for

• specifying text formatting
• distinguishing hour-angle and degrees
• inputting coordinates -> getting out radians

This way, we don't run into questions with "how do we store the coordinate"- we can always store the numerical value in radians, and use the formatter for the display. This also gives a struct that could be used for bulk parsing of coordinates.

[activexray]

Long time Julia user, just getting into astro for my PhD work. I'd love to help on this as it would be super useful for my research as well as for a few of my colleagues. Specifically, I would love to see the Unitful integration as well as the AltAzCoord type (as we use them frequently in radio astronomy). I think I can get a PR done soon for the latter - looking forward to helping!

[giordano]

That would be great! 🙂

[aplavin]

From the list in the first post, Unitful, EclipticCoords, and CartesianCoords are already present – for literal years :) @mileslucas can be worth checking these in the list.