Clause 5 is the method. An organisation names its clusters of PII by purpose, sets retention and deletion periods, groups clusters into a small number of deletion classes, and documents the policy and the procedure. The DIS sharpens the language on deletion processes and adds current storage technology. The clause is best read as a build order. The vocabulary in clause 3 gives the parts, and clause 5 puts them together. Start with purpose, because a cluster named after a system will drift as systems change while a cluster named after a function survives. Then decide the retention periods that the law and the business actually require, and resist the urge to invent one period per record type. The draft wants a small set of standard periods that can be reused. Those periods become the rows of a deletion class matrix, and each cluster is allocated to a row. The final step is the procedure: who runs the deletion, on what schedule, with what check, and where the result is recorded. The DIS makes two shifts that matter to a team updating a policy. It talks about deletion processes as well as mechanisms, which brings the operating routine into view rather than only the tool. It also describes the mechanism family in clause 5.1, including overwriting, device commands, encryption with key destruction and physical destruction, and it names current media. The archives and backup paragraph is the one most often misread. A backup is not a retention store. If a record must be kept for a legal period, it belongs in an archive with its own rule, not in a backup that is recycled on a short cycle. The clause also says that deletion must reach every system and every storage place, including paper and the copies held by processors, which is the sentence that turns a policy into a scope exercise.
The graph holds this control, the 0 it maps to, and the evidence behind each claim, over MCP and REST.