Based on the original equipment build specs and definitions circa 1980s, PAPI is very much superior to the then-days VASI as it is based on a focused / projected light beam. This gives a clean red/white changeover at all heights, eliminating the ‘pink’ VASI. Thus PAPI can be used below 200 ft and due to its higher power is better in poor visibility. More recent installation definitions (the equipment is basically unchanged) use mixed terminology and in some instances we have PAPIVASI or PVASI (Precision VASI). Our OZ friends have ‘T’VASI, which has a completely different ground installation; it is reported as being even better.
The original PAPI design concept assumed that the aircraft ILS Glideslope receiver antenna to pilot eye height was small, circa 5 ft when in the approach configuration / attitude; thus on an instrument approach the ILS glidepath and the visual glide path would be aligned. For small / medium sized aircraft with small pilot eye height to wheel heights (configuration / attitude/ and eye-main wheel longitudinal distance), the wheel clearance at the threshold is acceptable. However, for large / long body aircraft, or aircraft with large wheel / antenna to eye height, the required wheel clearance at the threshold may not be met, thus at some airports there is a second set of PAPIs located further into the runway. Only one set may operate at any time. The larger aircraft use the ‘long’ PAPI, which gives reasonable visual / ILS glidepath correlation and meets the threshold crossing height requirements.
The original concept was for the change in the PAPI lights (2/2 to 3/1) to corresponded with a 1 dot deviation on the ILS glideslope.
PAPI is an angular system (as ILS), thus if the PAPI is flown 1 light high/low the aircraft will still aim at the same visual glideslope origin, only the height over the threshold (or at any range) will change. I understand that some PAPI installations (possibly non instrument runways) have slightly tighter or slacker values between the lights from 2/2 to 3/1 etc. These changes increase / decrease the precision of the system, but due to this, the perceived sensitivity the system makes it more difficult / easier to fly. I also understand that this sensitivity is changed for steep approaches to keep an equivalent sensitivity (height change per lamp) vs range with that value chosen for 3 deg approaches.