Where ever did the "threaded rod" method originate?

It's still comes up, all these years later, so must have been published at some point?

Can you even imagine trying to determine how much over-movement (due to the considerable elasticity) would need to be dialed in to those hex nuts?
And can one imagine how many such time-consuming releases and re-tensionings, perhaps thousands of nut rotations, would need to be applied in order to to "home in" on the needed correction?
And that assuming that one chainstay wouldn't do all the bending, which it pretty much always would.

Can this threaded-rod method ever die?

It must be published somewhere, if perhaps only in the internet rumors archives. Perhaps it was even demonstrated in a Three Stooges skit, can you imagine?

As for me, I've corrected hundreds of fork legs and chainstays over the years, using only my foot, by myself, using balanced dynamic application of body mass and energy through my four limbs, which literally can make for some strenuous exercise.
Trial and measure, repeated iteratively and patiently, with an understanding of visual sighting, measurement, metal-bending behavior (varies wildly among various tubings) and steering behavior.

In the field, one can even use the bike's wheel axle as a width-measuring indicator as one proceeds (using no tools) to heave bent chainstays or fork blades into perfect alignment. A carefully-sized and placed block of wood placed on the ground, used as a stop, can prevent the likelihood of over-bending or buckling a tube. And a glove here and there can protect bike parts like saddle from abrasion against terra firma.

It's always good to learn the needed techniques on less-precious frames, though the late Sheldon Brown's published methods make this about as fool-proof as can be. And yet I learned using my feet.