I don't think we fully agree, because you are making it sound like the key aerodynamic features of the aircraft are optional, whereas we should instead view the aerodynamic configuration as immutable and required to meet their aerodynamic performance requirements.
The LERX of J-20 is not curved, it is straight (the initial tech demos s/n 2001 and 2002 hard curved LERX among other differences, but those were modified and refined by the 201X series prototypes that led to production aircraft).
As for the aircraft's rear section -- the LERX are about as signature reduced as they can be, and the most major difference vis a vis F-22 that remains is the 2D nozzles, which of course are lower in signature but a circular LOAN nozzle that J-20 has with WS-10s and WS-15s is also in the same category of nozzle outcome/solution that F-35 has.
Let me be clear -- I'm not talking about how stealthy J-20 is versus F-22 or F-35 or how effective they are as airframes.
I am talking about your prior assertion that: "It appears to me that the basic layout was selected for primarily aerodynamic reasons without full understanding of stealth shaping at an early stage of development, and then subsequently revised to substantially reduce the RCS without deviating too much from the original design."
My counter argument, is that on the contrary the stealth/signature reduction efforts made to the aircraft, based on the key aerodynamic features and configuration that they needed to use for the purposes of meeting aerodynamic performance requirements, do not show a poor understanding of stealth shaping.
On the contrary, if one was given the task to shape a stealthy airframe based on the following brief: "twin engine aircraft with circular LOAN nozzles, all moving twin tails, fixed ventral rear fins, dihedral canards, LERX, delta wing" then you would probably see something like J-20.
As I mentioned in the article write up/commentary, the paper itself likely didn't have a classification level so the "numbers" (including RCS, and things like radar performance etc) at the time were likely just hand-waved standins.
That's fine, I appreciate the discussion.
But I want to re-emphasize my participation in this discussion was not about the relative performance or stealth of J-20 versus other aircraft, but rather the extent to which one can infer CAC's understanding of RF signature reduction based on the stealth shaping for the specific aerodynamic configuration they had.
