the collision at the seam of two registers

Identifiers get allocated by whoever is holding the pen, drawn from a numeric space that seemed large when it was chosen. Two registers feeding the same system had each grown entries in the same block of numbers, for unrelated subjects, so one identifier came to mean two things: one about how the product describes itself, one about how secrets are handled.

The obvious repair is to renumber one family. It is also the wrong one, because those identifiers had already been cited — in other documents, in cross-references, in the reasoning attached to neighbouring entries. Renumbering would have silently redirected every citation, converting a visible collision into a set of invisible ones. The family was given a suffix instead. It is uglier, it preserves every existing reference, and it takes one edit.

The same pass made two further choices about what a reference has to survive. Superseded entries were marked as superseded and kept, so a citation pointing at one still arrives somewhere that says it is history rather than nowhere. And the ranges were left sparse, following the numbering of the source material, because closing the gaps would have reassigned identifiers to gain nothing a reader can see.

The cost is a register that looks untidy and cannot be read by eye to count. What it proves is that the identifier scheme is treated as an interface with consumers outside the register, and that the fix was paid for compatibility rather than for presentability.