Registered before computing: do AOSP's projects share conventions that Qt's do not?¶
The one mechanism for organizational coherence still standing is shared code conventions. It predicts something the convention-shape probe can measure without any adapter: within an organization, how separable its C++ projects are from one another when read only as convention shapes. Separability near 0.5 means the projects write alike; high separability means each project has its own conventions.
Prediction: AOSP's C++ projects are less separable from one another by convention shapes than Qt's C++ projects are. If Qt's projects are as alike as AOSP's or more, shared conventions do not explain the coherence gap either, and the mechanism is open with no candidate left.
Outcome of the shared-conventions prediction: refuted, in the opposite direction again. Code
convention shapes, .cpp only:
| within-organization project pairs | separability |
|---|---|
| AOSP frameworks/av vs system/core (the only pair that clears the size floor) | 0.806 [0.687, 0.872] |
| Qt, six pairs among qtbase, qt-creator, qtdeclarative, qtmultimedia | 0.534 to 0.608 |
| AOSP against Qt, pooled | 0.679 [0.614, 0.723] |
AOSP's two measurable projects differ in their conventions more than any pair of Qt's do, so shared conventions do not explain AOSP's coherence either. That is the fourth candidate refuted (pairwise coherence, centroid coherence, collaboration overlap, shared conventions), and, as committed above, the mechanism for why AOSP is a coherent organization in update space and Qt is not stays open with no candidate left. (AOSP has one scorable pair against Qt's six, which makes this a weaker test than the others; it is reported because it was registered.)