Data Dependency Trees

The analysis framework can be a useful tool for developing new analysis algorithms, but it serves a much more important purpose by providing a history of every piece of metadata in OME. Each attribute, in addition to having values for the fields of its semantic type, specifies which module execution created it. This module execution records its actual inputs, and those input attributes similarly record the module executions that created them. By tracing back to the set of modules with no inputs, it is possible to reconstruct a full data-dependency tree of the original attribute. This preserves a complete history of the original, non-computational attributes associated with an image, and all of the specific analytic routines used to create the rest.