CGS units type / other type systems

[MilesCranmer]

Some domains use non-SI unit systems. For example, astronomers like to use CGS units (also known as Gaussian units). Although CGS units are falling out of use (and have been dead in most fields for a long time) it could be nice if these other domains could use DynamicQuantities.jl.

One should generally not mix cgs units with SI because:

1. There are only 3 base dimensions: length, mass, and time. All others (even charge) are derived based on these.
2. Some equations need to be rewritten. i.e., you could not write one version of Maxwell's equations and have it work for both SI units and cgs units – there is an extra $4\pi$ factor in the cgs version.

I think all that would be required is a CGSQuantity <: AbstractQuantity and CGSDimensions <: AbstractQuantity (requiring #24), with one fewer dimension.

However, maybe it's just easier to define erg as one of the available units, and be clear that the SI convention is used. We could even throw a warning when someone uses erg – letting them know that it is being converted to the SI equivalent.

[icweaver]

Would a PR to add a few more common astro/cgs units that tend to pop up in our field (e.g., erg, dyn, Jy, etc.) still be helpful? Basically just handpicking a few from UnitfulAstro.jl. AFAICT, we can get away with just using their SI representations for the bulk of our applications.

This is still a very long ways off, but my evil master plan is to see if we can slowly start to wean ourselves off of Unitful over in JuliaAstro. Also cc'ing one of our other steering council members who has been thinking about unit handling lately with spectral flux densities and such @cgarling

[MilesCranmer]

Sure!

[icweaver]

Sweet, thanks! Just opened #196