Correlations

One should implement the possibility of deriving correlations among observables. This may either be done using TUnfoldBinning (see backup of this presentation during the SMP General workshop 2021), or using bootstrap techniques (see this note by the ATLAS Collaboration). The former has the advantage of ensuring perfect consistency in the treatment of several observables; the latter offers the opportunity to still provide correlations despite different publication time scales (and at the risk of inconsistencies). Both approaches are not contradictory.