To date, there are only two formal/official documents known to the public related to China's fifth-generation stealth fighter. One is obviously Song Wencong's paper, which laid the design foundation of J-20. Another recently disclosed one is a high-level requirements doc for China's fifth-generation aircraft, authored by Gu Yongfen (with a team) in 2003, former Chief Designer of J-8 and a senior academician of Engineering Academy of China. Here is the first and last page of the document.
In the last page, the document stated that the PLAAF requires the fourth/fifth-generation stealth fighter be competitive with F-22 and clearly superior to F-35 (I understand people may have different opinions regarding whether J-20 can achieve the requirements, but that's not the point here). There are lots of other interesting information in the document, such as range, stealth, supercruising, engine, radar.
We know that CAC chose the canard design for its superiority in achieving supercruising and to compensate for the relatively weaker engine. The trade-off for that design is more complex flight control and some slight sacrifice at stealth (but still meet their design goal in stealth).
So at least from the two documents, China has clear design goals and bench-marked against F-22 and F-35, made their design choices and trade-offs. In a video posted some pages back in this thread, the chief designer of J-20, Yang Wei, was interviewed at the sideline of the Zhuhai Airshow and said that J-20 has been under-work for "twenty years." Song Wencong's paper was published in 2000/2001, which is quite consistent with Yang Wei's statement. We also know CAC had gained a lot of experiences in canard design from their experience with J-10, DSI from FC-1, and had used J-10B/C for experimenting with avionics. Yang Wei made his name with designing the flight control software for J-10, which would be extremely important given J-20 rather complex aerodynamics. The CAC team was very well prepared and confident when they presented their design in competition with SAC for the fifth-generation stealth fighter project. They won the competition in no small part due to that confidence.
In the last page, the document stated that the PLAAF requires the fourth/fifth-generation stealth fighter be competitive with F-22 and clearly superior to F-35 (I understand people may have different opinions regarding whether J-20 can achieve the requirements, but that's not the point here). There are lots of other interesting information in the document, such as range, stealth, supercruising, engine, radar.
We know that CAC chose the canard design for its superiority in achieving supercruising and to compensate for the relatively weaker engine. The trade-off for that design is more complex flight control and some slight sacrifice at stealth (but still meet their design goal in stealth).
So at least from the two documents, China has clear design goals and bench-marked against F-22 and F-35, made their design choices and trade-offs. In a video posted some pages back in this thread, the chief designer of J-20, Yang Wei, was interviewed at the sideline of the Zhuhai Airshow and said that J-20 has been under-work for "twenty years." Song Wencong's paper was published in 2000/2001, which is quite consistent with Yang Wei's statement. We also know CAC had gained a lot of experiences in canard design from their experience with J-10, DSI from FC-1, and had used J-10B/C for experimenting with avionics. Yang Wei made his name with designing the flight control software for J-10, which would be extremely important given J-20 rather complex aerodynamics. The CAC team was very well prepared and confident when they presented their design in competition with SAC for the fifth-generation stealth fighter project. They won the competition in no small part due to that confidence.