Dear Yousuf Khan:

On Saturday, October 15, 2016 at 10:45:21 PM UTC-7, Yousuf Khan wrote:
....
Maybe the issue here is not to find a new
modification of Newtonian gravity, but perhaps
our reliance on still using Newtonian gravity
even 100 years after we found a better theory
of gravity might be the problem here?

I find it more likely that a nearly 100 year old assumption that luminosity is directly proportional to the amount of mass present, when it has long been known that luminosity drops off rapidly with surface temperature. If you have cooler objects, they simply don't put out as much light... especially in the visible light bands.

We're still using Newtonian gravity after all
of these years, because it's frankly much easier
to calculate with than General Relativity.

Paper on this subject for a "simple" galaxy, and evaluating the possible error between Newtonian gravity-as-a-force and GR, and in that galaxy, it is a 1% (or so) error, not the necessary 300% error.

But in a many-body system such as stars in a
galaxy or galaxies in a universe, those simple
inverse-distance squared relationships simply
don't work out anymore?

They do work out "well enough", for simple gravitation.

But we are "blind as bats" at these scales, and have a full complement of "flatlander fallacies" that we have to divest ourselves of.

We're still using Newtonian gravity in this day
and age because we still don't have computers
strong enough to do a GR calculation for an
entire galaxy.

False. The amount of computer time might still be abysmally long for an interesting galaxy, but it would still be doable. After all, Nature does this math in real time...

Using even our strongest supercomputers we can
do perhaps a simulation of only a few million
stars in a galaxy using GR, but our galaxy contains
perhaps as much as 400 billion stars, so we keep
approximating with Newton.

Yet, even small spirals show a need for Dark Matter. Globular clusters, essentially don't.

If one day we can do a full simulation of the
Milky Way with all of its entire 400 billion stars,
then likely we'll see surprising results coming out
of GR that are inconsistent with Newton, and then
we'll be finally shaken of our illusion that Newton
is "still good enough".

Maybe. But the speeds and curvature on something the size of a galaxy, even the Milky Way, should present minimal error in using Newton.

Now what I wonder is, if the "perfectly mirrored, massless box, containing photons", which has rest mass, exists between a star and the gases/dust/planets that give that star a background temperature higher than the CMBR. So some Dark Matter (probably less than 1%) might still be photons in transit between intersystem objects...?

David A. Smith