mechanism for binding data provenance to annotation rows

calcId provides a way to bind the code generating an annotation row to its result, and the input parameter settings for an alignment view to the code that performs autocalculated alignment annotation. In other situations, calcId is used as a 'type' string - but there is no way to associate input parameters with different annotation results derived by the same code for the same associated sequence/group/alignment but with different input parameters.

This specifically occurs for data integration processes - where different instances of the same data that lead to different result annotation are associated with the same sequence (e.g. PDB files linked to the same region of a sequence).

A properties pattern would allow arbitrary data to be associated with annotation rows, enabling instance specific data to be stored with the results. Alternately, a full provenance model would also allow this, but potentially would be more expensive to traverse when checking for completeness of operations such as:
for all associated input for sequence q: does result annotation from each input on the dataset sequence/alignment ? if not, perform calculation to generate annotation.