compatibility on its own axis

Once there are many independent lineages, the question of whether two of them are compatible stops being answerable by comparison. They will never share a label, so there is nothing obvious to compare. The record answers by giving compatibility a plane of its own: a declaration made and maintained separately from the identity of either side.

The tension is that labels are cheap and usually right. Two things built from the same ancestor are compatible far more often than not, and a rule inferring compatibility from shared descent costs nothing to operate. Explicit declarations are a maintained burden, they can be wrong, and a stale declaration is worse than no declaration because it is trusted.

The decision is to state agreement as something two parties assert and something a third party can check, rather than as something inferred from a family tree. Compatibility lives on a level where it can be declared once and consulted by anything that needs it, so a boundary between lineages becomes a place where a claim is tested instead of a place where an assumption is made.

What it costs is a maintenance surface that grows with every lineage, declarations that go stale when the thing beneath them moves, and the admission that compatible here means asserted and checkable rather than proved.

What it proves is that the team identified the axis its own success would break — many lineages, none of them the centre — and gave that axis a home before the breakage arrived. Design that anticipates its own success is rarer than design that anticipates load.