This is not an area where I am an expert, but I do not believe that Gibuld is correctw where he says
[Number of Least Significant Digits] meint also, zur prozentualen Ungenauigkeit wird noch auf die letzte Stelle die Zahl der Nachkommastellen hinzuaddiert, um die absolute Ungenauigkeit anzugeben. It is
not the number of digits to the right of the decimal that are added to the specification of accuracy; it is a
multiple of the least significant digit.
Maybe an example will help.
Suppose that the specification of accuracy for a device is given as ±0.2% + 2, and suppose the device reading is 1.234.
The first part of the accuracy is ±(0.2% of 1.234), or ±0.002468.
The second part of the accuracy is ±(2 x 0.001), or ±0.002.
The total accuracy of this measurement is then ±0.004468.
The measurement could be given as 1.234 ± 0.0045.
See
http://www.physics.mun.ca/~cdeacon/labs/gen_e... for another example.
(I cannot say how Gibuld's interpretation would apply here.)