During testing of the function we found a bug in the atomic function (a wrong field being used, undiscovered so far since in R globals propagate inside functions) and adapted the CST_MultiEOF interface to expect a $data element instead of $exp in the CST object it is receiving