Compound steps are currently a pain to accommodate with the active-to-passive translator, because that script assumes that transcript steps are 1:1 with script steps. We should make the transcript more self-contained, e.g., by explicitly stating what each step represents. Maybe this could even be an annotated version of the script language, where the steps are atomic (expanded from compound steps) and each step is annotated with its outputs.
Compound steps are currently a pain to accommodate with the active-to-passive translator, because that script assumes that transcript steps are 1:1 with script steps. We should make the transcript more self-contained, e.g., by explicitly stating what each step represents. Maybe this could even be an annotated version of the script language, where the steps are atomic (expanded from compound steps) and each step is annotated with its outputs.