Sampling vs. taking some - 59349

Collecting a sample is a delicate task that is Not naively equivalent to simply 'taking some of the material'. The question examined is: 'What is it exactly?' The problem of sampling in general, and for nuclear decontamination in particular, is properly defined. A theory is presented (Gy's Theory of Sampling, a.k.a. TOS) that brings all the answers and allows us to put them to work. The author draws form his lifelong experience in research, teaching and practical applications in this domain to emphasize the critical odds (i.e. risks) of not taking sampling explicitly into account when assessing grades and concentrations. The evolution of the acceptance of this theory in the nuclear industry is finally illustrated, and a hopeful glimpse into the future concludes the presentation. Equally interesting, however, besides what has already been achieved at the CEA along these years, is the realization of what could not be done with TOS, and therefore had to be treated in some other ways - e.g. using mapping tools (geostatistical). It is one the great side-advantages of using a consistent theory that it warns you, before it is too late, that what you are trying to do will not work: TOS, indeed, much like its Geo-statistics sister, besides preventing many a disaster, can provide pragmatic lessons in scientific humility that are best not being left ignored. In conclusion, there are great tools out there, such as TOS, that are well worth investing into, and that our community should be much more attuned to. (author)