Was just browsing (insomnia) chapter 9 and I think that credit for the "two-transistor" bandgap voltage reference topology that doubles as an error amplifier as used in the LM317 ("Anatomy of a 317", page 604 and figure 9.13 page 606) has been wrongly attributed to Widlar-Dobkin.
To date I was of the understanding that this bandgap topology is originally attributable to Paul Brokaw, as detailed in:
A Simple Three-Terminal IC Bandgap Reference, IEEE JOURNAL OF SOLID-STATE CIRCUITS,VOL.SC-9,NO. 6, DECEMBER 1974
I have attached a pdf of this document. Note that this paper references the paper by Widlar mentioned in chapter 9 (ref#14 page 604):
New Developments in IC Voltage Regulators, IEEE JOURNAL OF SOLID-STATE CIRCUITS,VOL. SC-6, NO 1, FEBRUARY 1971
For convenience, I have attached a pdf of this document also.
Brokaws paper begins with the "conventional bandgap circuit" as outlined by Widlar in the latter, and then builds upon the concept, describing the "new" combined reference and regulator topology that is topologically/functionally equivalent to figure 9.13 in AOE3.
AOE3 "Anatomy of a 317" states that the LM317 was designed by Widlar-Dobkin "around 1970" and gives ref#14. However, ref#14 doesn't actually detail the LM317 circuit shown in fig 9.13, but that of the LM109 fixed 5-V linear regulator, which uses a somewhat different variation of the "conventional bandgap circuit" (as far as I can currently see now at 1.22 am).
It would still be nice to know exactly when the LM317 was designed, but I suspect it wasn't before the Brokaw paper unless there was a highly coincidental/improbable independent development.