Vishay datasheet ( Motorola design) defines several max currents with assumptions for each.
Motorola defined voltage rise from
IZT1 =1 mA with a high knee resistance to the low knee resistance at max power of 250 mW at IZT2.
If you cannot apply a heatsink to the leads and keep the leads at 25'C then you use a low current like 5 mA which became an industry standard by other suppliers to dissipate about 10% of this. The reason is that the tempco. of voltage changes from NTC at low
Vz to PTC above 5V and keeps rising. The regulator current is the maximum at 900 mW but only if you regulate the temperature of the leads @ 25'C.
The
IR is the max pulse zener current for 10 ms is the must not exceed max value.
You do not want to use 250 mW unless you predict the temp rise from 110 deg C/W above ambient and know the tempco for that Vz.
Thus 5 mA became the test method used for manufacturers other than Motorola. With dynamic resistance specs you can predict the error voltage with a dynamic load and interpolate voltage between the Izt1 and Izt2. The rest of the world decided later to standardize on 5mA so that you use a constant current source when it matters or is needed.
I understood this 25 yrs ago when I started comparing the Vf of 5 mm LEDs. The wide variation of Vz is all due to knee resistance and not the threshold for each colour. Mfg quality controls the tolerance on these nominal knee resistance values which also varies with Zener Vz and LED colours.
Using the std. 20 mA LED test the best R/Y LEDs have a Zzt of <10 Ohms and the best 20 mA white LEDs are < 15 Ohms . There is also a direct quality relationship with knee resistance times Pmax @ 85'C that is predictable.
Each Iz test method has its pros and cons.
FWIW Vishay has grown by acquisitions such as General Semiconductor in 2002 and I speculate has purchased the rights to produce many OEM components that have expired on patent rights such as Sharp and Panasonic Opto's and light sensors.
I found that my select suppliers of LED's have much lower knee resistance than all the rest. This allowed me to get LED's better than Zeners in two price ranges +/-50 mV or +/- 100 mV @ 20mV. rather than 2.9 to 3.8 for white.