julia> Pkg.add("SkyCoords")

julia> using SkyCoords

# create a coordinates object
julia> c1 = ICRSCoords(0., 0.)  # inputs are ra, dec in radians
ICRSCoords(0.0,0.0)

# access ra, dec individually
julia> c1.ra
0.0

# convert to a different system
julia> c2 = convert(GalCoords, c1)
GalCoords(1.681404315278054,-1.0504869904089078)

# Note that galactic coordinate fields are l, b
julia> c2.l
1.681404315278054

# FK5Coords is parameterized on equinox.
# Equinox refers to Julian year and can be floating-point or integer
# (though using integers seems to be slightly faster).
julia> convert(FK5Coords{2000}, c1)
FK5Coords{2000}(1.1102233710147402e-7,4.411803426976326e-8)

# Arrays of coordinates
# =====================

# create an array of coordinates 
julia> c1 = [ICRSCoords(0., 0.) for i=1:3]
3-element Array{ICRSCoords,1}:
 ICRSCoords(0.0,0.0)
 ICRSCoords(0.0,0.0)
 ICRSCoords(0.0,0.0)

# convert entire array to a different system
julia> convert(Vector{GalCoords}, c1)
3-element Array{GalCoords,1}:
 GalCoords(1.681404315278054,-1.0504869904089078)
 GalCoords(1.681404315278054,-1.0504869904089078)
 GalCoords(1.681404315278054,-1.0504869904089078)

# There's no performance gain from using this "vectorized" convert,
# except conversions to/from FK5Coords, where the equinox precession
# can be done just once for the entire vector, leading to a modest ~2x
# speed up.
julia> convert(Vector{FK5Coords{1975}}, c1)
3-element Array{FK5Coords{1975},1}:
 FK5Coords{1975}(6.277595732508468,-0.0024292220493946897)
 FK5Coords{1975}(6.277595732508468,-0.0024292220493946897)
 FK5Coords{1975}(6.277595732508468,-0.0024292220493946897)