Yes you have misunderstood in this instance, probably because I didn't convey it well enough.
What I was trying to convey is that all the base measurements are done using a linear scale. For example a densitometer measures the transmission of light as a ratio using a linear scale i.e. normal numbers, typically expressed Logarithmically which we often choose to use for convenience but you can always use the antilog to get back the base measurement. Digital processing often requires getting back to the original base measurement so the calculations can be done correctly. I.e. the ratio of light in the scene, or ratio of light being reflected from the print, monitor etc.It is also further confusing because some might consider a response curve which when expressed as a logarithm that is a straight line as being linear and others are looking at the antilog of that which in that instance will be a curve (unless of cause the slope is 1), and will see it as non linear.
