I've been considering my hypothesis that Planck's constant is really twice the accepted value, based on that the photon energy is related to its frequency by half the accepted amount, for the composite photon, as I proposed two posts ago. The hypothesis fails in an obvious way. Since it implies the electron g-factor is actually unity, it is inconsistent with the Zeeman effect, where it is essential that the electron g-factor be (close to) two, in order to agree with observation.
I now have an alternative hypothesis about the photon energy being half the accepted amount in relation to its frequency. That hypothesis is that it explains the spin-orbit coupling anomaly as well or better than the Thomas precession does, while also not otherwise differing noticeably from observation.
If the photon energy for a given atomic transition is half that currently assigned (that is, if the current energy assigned is twice the correct value), then the measure of the spin-orbit coupling energy from emission spectra will appear to be doubled as well. So, correcting the transition energy to be half the accepted value will recover the spin-orbit coupling energy consistent with the original prediction. There will be no need to apply the Thomas factor of one-half in order to find agreement with observation.
As I think I have stated previously, and various people have been pointing out for many decades, it makes no sense that the Thomas precession, a purely kinematical effect, can alter the interaction energy.
Halving the energy per photon has another consequence of halving the expected magnitude of the Zeeman effect generally. I'm not sure if this would be noticeable or not.