The layout is the archive's, discharged once and generically in $Encoder$: the two counts, one entry per job for each of the processing times, deadlines and weights, the $n + 1$ offsets, and the eligibility lists run together. What is about this construction is only that every machine number it lists is a machine — the validation machine is $C(k,2)$ and the edge selection machines are numbered below it by $pairIdx$, which $Construction1Index$ shows lands exactly in $[0, C(k,2))$.

The word is the instance followed by the threshold, as a decision instance is.