(feedback from Linking GSBPM and GSIM task team; May 2020)

Task team on Linking GSBPM and GSIM (2019-20) under the Supporting Standards Group is working on identifying information needed for GSBPM sub-processes and representing them as GSIM information objects. The team has found few issues related to using Process Inputs and Process Outputs in context of non-Process GSBPM Phase (e.g. Design Phase, Analyse Phase).

Data Set is often considered as the primary Transformable Input but there are also different types of inputs that are “transformed” in a very different way. For example, when we validate Data Set (e.g. GSBPM Sub-process 6.2), what is being transformed here is some document (metadata) containing the status of the Data Set or information about quality, but not the Data Set itself per se (if there is any problem detected during validation, the Data Set should be sent back to other sub-processes and modified there, not in Sub-process 6.2). So in this sub-process, would it be right that Transformable Input is NOT Data Set but its referential metadata? Or should Data Set be still under Transformable Input to highlight that primary element under the sub-process is Data Set.

  • No labels

1 Comment

  1. InKyung Choi

    14 Jul, 2020

    More discussion on Trasnformable Input from Linking GSBPM and GSIM task team (Meeting from July 7 2020)

    Team discussed whether business register should be Transformable Input orProcess Support Input (GSBPM sub-process 4.1 Create Frame and Sample): 

    • Business register is maintained by other business unit, so its maintenance is a separate business process. It is not "transformed" in the same sense as in, e.g. E&I where dataset is edited and imputed (we cannot say frame is "transformed" business register)
    • However, business register plays central role in this Process Step and it is weird to put it in Process Support Input along with other objects that are not so critical without differentiating their roles. Business register is the starting point of this Process Step and we cannot create sample without business register, is it right to have it in the same category as, e.g. variance of auxiliary variable, which is almost optional to this Process Step? If Transformable Input is the object of interest of that Process Step, business register should be Transformable Input. 
    • We have a similar problem in, e.g. GSBPM sub-process 5.7 Calculate Aggregates. Here, micro dataset is Transformable Input, but macro dataset (which is Transformed Output) is not exactly "transformed" micro dataset, they are different datasets with different data structure. Operation here is the aggregation function applied to micro dataset to produce macro dataset which is similar to how a selection function applied to business register to create the frame in this example. We should review how Transformable Input was used in other sub-processes
    • The issue is the definition of Transformable Input. Should it cover things that are not "transformed" but are core to the process that is being executed, a read-only input that we use for selection, extraction, aggregation, etc, but NOT Process Support Input because it is more than simply "supporting" that Process Step?