Late Roman Army during the 5th century

[Robert Vermaat]

I’m sure he does, but then I don’t recall that he has sufficient evidence to prove his case for that doubt.

He made an interesting point: barbarians were able to take over areas, apparently with minimal if any resistance, which in theory were guarded by Roman units (which suggests the units didn't actually exist). North Africa was supposed to be protected by "31 regiments" totalling about 15,000 men, plus 10,000 or more limitanei. These should've been able to deal with a Vandal/Alan force of 20,000 or fewer. But Geiseric doesn't seem to have encountered resistance until he got almost as far as Hippo. Apparently unsure of his strength, Bonafatius must have hoped a long trudge would wear out the Vandals, so he made his stand far to the east. But it still didn't work; they had to send an SOS to the east which sent Aspar. The evidence suggests much of Bonafatius's strength was just on paper.

I’m not sure which areas he meant, but Africa is not a correct one. True, on paper these numbers were available in Africa, BUT all over the diocese I think, which is such an enormous region that you can’t mass all units and leave the borders unguarded (as the regulars did in Gaul).
Plus, those numbers we know date from AD394 (being the Notitia Dignitatum) and it would be bad science to assume that they were correct a generation later.
Plus, that generation had seen quite a bit of conflict within the Empire, causing all kinds of troop movements, which would at least make it logical to assume that Africa, like Gaul and Britain), would have to send troops for the defense of Italy.
Plus, I think you’ve misunderstood Bonifatius’ plans, because the Vandals seem to have arrived as an invited force to begin with, not as an invasion force to be met by all available Roman forces of the beach. That relation may soon have turned sour, but it might well have prevented any massing the available troops for that part of Africa would perhaps have become impossible with the fox in the hen house. It’s perhaps nice to mention that the traditional date of the Anglo-Saxons in Britain (early dating) is in exactly the same year, also by invitation (Vortigern), also with the same results.
About arming the population:

During various emergencies of the republic, the Romans recruited all kinds of untrained people--even slaves after Cannae. And the slaves won a victory too. mile:

I’m sure that was a jest. Most of these slaves were captured tribesmen, most citizens had served in the military. Times changed.
Adrianople:

The post-Adrianople period has learned us it took years to rebuild the Roman army, and even that may never have been achieved.

Because by then most people seem no longer willing to serve. Rome had big armies again only a few years after Trasemine and Cannae and revived militarily after the catastrophes of 252-260; Gallienus slaughtered the barbarians at Naissus in 268.

That’s because the Roman citizens had military training. That, and the armies back then were smaller than the numbers involved with the late Roman army. Training took a lot more time and that was not available. I think they took the wrong approach by hiring barbarians under their own leaders, but perhaps it was necessary. I have no information from sources that unwillingness to serve was a large-scale problem at any time. At any rate there’s no connection between that and the use of barbarians as regulars.

Sure, Caesar hired Germans, there were balearic slingers etc. But the difference is that non-citizens then supplemented a strong central core of legions. Whereas in the late empire barbarians became a substitute for regulars.

Indeed, and that’s what had a large part in demise of the West. But it was no logical process – the East managed to overcome it. The Roman army, when circumstances allowed, did not come to depend on barbarians permanently.

Attila is said to have considered the Roman army of 451 "beneath contempt."

Pride. His biggest mistake. That same army defeated him a few years later, before he succumbed in bed with yet another bride to an overdose of his own vomit. A real warrior’s death.

As for the east, starting in the late fifth century, it was no longer the same, since reforms are said to have reduced its dependence on mercenaries and barbarians.

Political reforms, yes. But not military reforms – that army operated from the same handbooks.