Simon,
I was quoting your post to offer support to your arguement, but also to address a couple of 'minor' (percieved by me!) errors....
Half of the system may well be working, but due to the piston travel NOT operating over ALL of it's travel (due to having to close the valves to the reservoir), the 16% does not halve to 8%, as the first bit of both pistons' travel has little effect on the braking pressure, and this bit of movement is present whether there is a leak or not.... (but for the sake of your arguement against Phil, it may well be irrellevant.
The 0.5mm comes from the piston travel excess between operating at 8% and slightly different, to allow for the portion of travel where pressure doesn't increase.
The underfelt refers to the amount of room that the pedal can travel before coming to a stop (as per an earlier post of mine), and was perhaps, a trifle flippant of me to include!
I still maintain my principle of fluid compressibility, and dissagree that ALL fluids are compressible.... Some more so than others, and not *ALL*.
It is accepted wisdom that Glycol based brake fluid is about 1.9x less compressible than Silicone fluids, but in normal use, it is also accepted that brake fluid doesn't really compress to an appreciable amount (otherwise they wouldn't use it!).
In fact, you need to apply some huge amounts of Pascals of pressure to cause a very small compression (and then it is down to the temperature increase more so than the actual pressure - (Physical Laws), but, yes, I accept that glycol will diminish in volume by 12.3 ppm during a temperature increase of 100 Celcius, IF 14 x 10 (to power) -11 KPa is applied within a sealed container.
However, we will never see that sort of pressure in a brake system, but continual applications of the system WILL increase the temperature, leading to boiling of the fluid, and as soon as it goes over about 112C the water absorbed in the glycol will come out of solution and boil, adding gas (steam) to the mix, which is VERY compressible - and that is where the real problem lies