PAPI Origins
You might be interested to know how PAPIs became a replacement for VASIs.
In the mid 1970s PAPIs were used as a visual aid for use by (principally) aeroplanes and a helicopter (HS748, BAC 1-11 and Wessex 2) operated by the Blind Landing Experimental Unit (BLEU) at RAE Bedford to assist pilots flying normal (3 degree), steep and two-segment approaches. Thanks to an initiative shown by the two airfield lighting boffins, JJ and TS, ICAO agreed that PAPIs might be employed as an alternative to VASIs on civilian aerodromes.
As one of the BLEU pilots (the clue is in my PPrune name), I used them on many occasions, with a chap on the ground adjusting them to a range of different approach slopes in the time it took us to fly a circuit. One trial we conducted on a clear day with our flight being monitored by highly accurate kine-theodolites involved each of three pilots flying three approaches to Bedford's long runway: one with PAPIs, one with VASIs and one with no visual light aids. The most accurate approach was that flown using PAPIs and the next most accurate was that using neither PAPIs nor VASIs. With VASIs it was all too possible to 'bounce' in and out of the relatively wide approach slope corridor.
We also conducted trials of the first (and only) British-designed GPWS, and when demonstrating its operation at Bedford, i.e. diving towards the runway, on a day when the visibility was not especially good, I used guidance from PAPIs to help me demonstrate to the aviation Press standing behind me how the GPWS functioned 'in a timely manner' whilst ensuring that we remained over the aerodrome. Interestingly, the Press were more fascinated by the PAPIs than by the GPWS!
Although ordinarily you would not expect to expect to use PAPIs below 200ft above touchdown elevation since by that time on a manually-flown approach you should have acquired sufficient visual cues to have confidence that you could follow-through to a safe touch-down, we noticed - as you might expect - that they still provided a measure of guidance below that height. This, however, was not and should not be relied upon as primary guidance for reasons already stated by other respondents.
After leaving Bedford I enjoyed (!) a spell in the MOD where, when a general question was asked if anyone had any ideas as to how finance that had been arranged for the introduction of Tornado but not spent due to delays might be used, I proposed and had accepted a suggestion that industry might be persuaded to invest in produced a few sets of PAPIs for use at Northolt, to facilitate accurate approaches over built-up areas, and at a Lightning station to reduce the risk of tail-strikes, again by helping to promote stable approaches.
After that, the idea caught on and now PAPIs are used across the globe. Incidentally, our proposal was that a set should comprise eight boxes with four on either side of the runway co-located with the glide slope origin. If you've flown an approach with this arrangement you will have observed that the gap between each group of four helps define that origin, and of course for the occupant of the right-hand seat the right-hand group is more in line with his/her view of the runway.
And this all began 40 years ago!