The GSIM v1.0 Production group has a mismatch between the number of subtypes for Process Input and Process Output.
|Process Input||Process Output|
|Parameter Input||Process Metric|
|Transformable Input||Transformed Output|
|Process Support Input||?|
Separately, at the last GSIM IG meeting, the issue of a missing process output object was also raised. It was felt that there is a need to have an object to allow for output objects that are neither Transformed Outputs (e.g. statistical data or metadata) nor Process Metrics (e.g. measures of the performance or effectiveness of the process step), but more the types of outputs that could be future Process Support Inputs.
"56. A Process Support Input influences the work performed by the Process Step, and therefore influences its outcome, but does not correspond to a Parameter Input or a Transformable Input. Examples could include:
- A Code List which will be used to check whether the codes recorded in one dimension of a dataset are valid.
- An auxiliary Data Set which will influence imputation for, or editing of, a primary dataset which has been submitted to the process step as the Transformable Input."
This could be accommodated within the v1.0 style framework by adding a new object, Process Support Output, which would be the output analogue to Process Support Input.
In the "GSIM Level 2" proposed for discussion, IMF contributors dealt with this issue somewhat differently. The concept of Process Metric is replaced with a concept of Process Measure, which is defined as recording metrics associated with the Process Step and logging work and knowledge objects used in the Process Step and work and knowledge objects created by the Process Step. This model assumes that the "work and knowledge" outputs created need not necessarily be transformations of the transformable inputs, but may be some type of reusable or otherwise informative byproduct. These byproducts may fit an existing structured object from elsewhere within GSIM, or may be unstructured explanatory information (which would fit within the Referential Metadata object in this model).
Arofan also had a number of ideas on the call and offered to comment on this issue.