please see iec61966-2-1
Section 4.1 Reference image display system characteristics:
The reference image display system is a computer controlled cathode-ray tube display and shall be as follows.
- Display luminance level: 80 cd/m2
- Display white point: x = 0,312 7, y = 0,329 0 (D65)
- Display model offset (R, G and B): 0,0
- Display input/output characteristic (R, G, and B): 2,2
and then
Setion 5.1 Encoding transformations Introduction:
The encoding transformations between CIE 1931 XYZ values and 8-bit RGB values provide unambiguous methods for representing optimum image colorimetry when viewed on the reference display in the reference viewing conditions by the reference observer. The CIE 1931 XYZ values are scaled from 0,0 to 1,0, not 0,0 to 100,0. These non-linear sR′G′B′ values represent the appearance of the image as displayed on the reference display in the reference viewing condition. The sRGB tristimulus values are linear combinations of the CIE 1931 XYZ values as measured on the faceplate of the display, which assumes the absence of any significant veiling glare. One impact of this encoding specification is the creation of a mismatch between theoretical reference display tristimulus values and those generated from the encoding implementation. The advantages of optimising encoding outweigh the disadvantages of this mismatch. A linear portion of the transfer function of the dark-end signal is integrated into the encoding specification to optimise encoding implementations. Recommended treatments for both veiling glare and viewing conditions are provided in annexes D and E.
It does not matter so much to me what people say they might have meant in the sRGB standard.
It is very unambigoiusly specified how the sRGB should be implemented.
The pure 2.2 power law is numerically unstable around zero, especially in 8bit. So they looked for an encoding method which handled multiple linear to none linear transforms and back without breaking the image. The slight mismatch to the Display Chararateristics was then just accepted as a better of two compromises. Again see here:
The advantages of optimising encoding outweigh the disadvantages of this mismatch.
To me this is clear enough.
Nowadys we do not use 8 bit processing anymore so the compound function is not needed anymore.
Also we have better methods (DRTs) to introduce flare/glare compensation
I hope this helps
Daniele