From Kesha: we need to ensure that the translator relations constrain the op queue columns to be zero beyond the minicircuit size. (Also seems possible that this is implicitly accomplished through the merge protocol itself plus the merge-translator consistency checks).