Since the Select Children Process [SR.1] only defines behavior for selection happening exactly once, all unavailable children can never have any attempt data. Yet the pseudocode for all the rollup processes (Measure Rollup Process [RB.1.1 a], Completion Measure Rollup Process [RB.1.1 b] and Rollup Rule Check Subprocess [RB.1.4]) iterate over the children rather than Available Children like the other processes.
Is this an oversight in the pseudocode? It seems the outcome would be mostly equivalent as the children are further filtered for consideration prior to being evaluated for the rollup.
