OK, several things are wrong here.
First, I never said anything about refractometers in my post, you you can't claim I said anything "false" by talking about them. Since you brought it up, I'll go there.
We use both hydrometers and refractometers measure the same thing, through different means (and some conversions have to take place). Did you bother to use your instruments to measure salinity, like I suggested? If you did, you'd see that I'm correct, and have provided a very easy to observe and understand counterexample. Seriously, try it. Your statement that these instruments only measure sugar is 100% false.
Consider this: when both instruments are calibrated, distilled water must be used, right? That's because the unknown amounts of dissolved solids in other types of water make it unusable to calibrate with. You
could calibrate an instrument with other types of water, if you knew exactly the density or refractive index of the solution, but it's rare that we have anything like that on hand.
If you can't bother with taking 5 minutes to drag out your equipment and try this, consider these points. The reason we can rely upon the measurements we do, even though we're measuring all dissolved solids, is that a properly made wort is
mostly fermentable sugar, and we're generally not adding or taking away any dissolved solids other than what gets fermented. We make wort to be that way so it makes better beer, but it means that the gravity changes are mostly due to increases or decreases in sugars.
Since I said I'd talk about refractometers as well, look over
this link. People who work with salt water aquariums and in ocean water environments use the very same refractometers we do to measure salinity. Refractometers are used in commercial industries, medicine, an other fields to determine the purity of a solution with known ingredients. Even the
Wikipedia entry for refractometers states this. The only difference in the "types" of handheld refractometers is the difference in scale, which is tweaked to suit the needs of the user (i.e., the refractive scale measured and the assumptions and calculations that are used to determine what the "scale" looks like in the viewable readout). The base product is the same, and it measures optical refraction, nothing else.
This thread has good reading on the subject. See post #38 for a nice discussion of how these things work. These instruments are useful, but only to estimate the stuff we're really after. We have to know about other variables and make certain assumptions along the way.