> Since your simpler approach does not obtain an integrated measurement, one > complication would be significant swings in the temperature of either the > device in test or the control, with the control running hotter than the > functional unit for stretches of time. >
That would be bad. But, if Parkhomov's results are correct (I would like to belive so), I expect simply a large difference in temperature for large periods of time, with the loaded reactor hotter than the control reactor. If it is done using live science, MFMP style, nobody will argue about it anymore (OK, there are people that will argue, but replications will abound and, with time, all complains will be just history). Alberto. > Eric > >
