Actually, measuring out FL at any focusing distance isn't that difficult. Let's take a practical and real example:

With my Tamron SP 35-80 mm Model 01A CF Macro lens, I set the FL on the lens to 80mm (the lens sets itself at that position at closest focusing distance) and photograph a ruler at the closest focusing distance, which I shall call S.

Now I don't know where the optical centre of my lens is but I can measure the distance S from the ruler to my sensor plane. I find S=27 cm.

Next, I study the image of my ruler. I take 1 cm (called b1) in the following and find that the image (hereafter referred to as b2) of that one centimeter is 660 pixels long on my K200D sensor. That sensor has pixels that are 0.00606 mm wide and thus, my image, b2 is 0.4 centimeters long. I have M = b2/b1 = 1:2.5 as Tamron has also promised me in their lens sepcification.

Now I have enough - measured! - data to compute my FL:

In this figure, I know S = 27 cm, b1=1 cm and b2=0.4 cm. From the figure you can see (from the principles of equidistant triangles) that M = b2/b1 = s1/s2 and S = s1 + s2. Thus, you have two simple equations with two unknowns and you find:

s1 = S/(1+M) and s2 = S*M/(1+M) where I can measure both S and M.

I finally need to know the general lens equation that tells me:

1/FL = 1/s1 + 1/s2.

Inserting my actually measured numbers, S=27 cm, M = 0.4 I find that the focal length of my Model 01A at closest focusing distance at the "80mm" setting is actually 55 mm !

And all I needed to find out was a ruler and the pixel size of my sensor.