use
Mounting a conventional strain gage just at the ends would give you the same output assuming you had the dexterity to glue just the ends. dl/l should be const. over the entire area of the strain gage site.
Correcting for thermal expansion would require temperature data over the entire length.
Adding a +/- nanometer error to a +/- meter error just gives you a +/- meter error.
GPS wouldn't be useful for the actual data.
They claim laser interferometry won't work over long distances, i.e.,
50 km, but they are probably thinking of measuring an entire unknown distance when all that is really necessary is a _change_ in distance.Maybe use several different meters at different wavelengths so that you don't skip a wave.
Bret Cahill