Dark matter-only density?

[ntessore]

Currently, the API describes $\Omega_m$ and $\Omega_b$. Do we require dark-matter only quantities for $\Omega_{\rm dm}$? And further subdivision in cold dark matter $\Omega_{\rm c}$ and warm dark matter?

[nstarman]

Astropy doesn't have one. Do any other libraries?

[ntessore]

Most cosmology codes will treat all of those separately. I'm not sure if it makes sense to introduce a "dm" component, or keep it at "matter" and "cdm"/"wdm". Pinging @gcanasherrera who has more experience in the low-level details of other codes.

[ntessore]

The entire thing is problematic because there are no neutrinos in $\Omega_m$, according to the docs. This is probably the wrong thing? I guess matter = all non-relativistic matter and separate dm components would make much more sense here.

[gcanasherrera]

From my experience working with cosmo codes, $\Omega_m$ is often defined to include neutrinos, making its interpretation a delicate matter. Therefore, if you assume (as another user might, by analogy with other codes) that you can infer $\Omega_{dm}$ from $\Omega_m - \Omega_b$ while accounting for neutrinos, this could lead to complications.

I would reconsider: define $\Omega_m$ taking into account the poor neutrinos.

(And thanks for asking for my input!)

[ntessore]

Right, I guess it makes sense that $\Omega_m$ is the catch-all for everything going $a^3$, $\Omega_r$ is the catch-all for everything going $a^4$, etc.

Where that isn't enough, I guess we do need individual components for everything. So perhaps the best way forward for now is to change the docs to say $\Omega_m$ includes all matter components (incl. neutrinos) and add a new $\Omega_{\rm cdm}$ component (or $\Omega_{\rm c}$ but there is a chance of confusion with "critical"). What do you think @nstarman?

[nstarman]

I largely agree. It should be made abundantly clear that Ωm is a^3 and Ωr is a^4 and that it is up to implementing libraries to properly deal with components that transition between them.
We have to add a RadiationComponent etc, since we currently only have PhotonComponent.
Adding other matter subdivisions is AOK.