An errata section, and three DOIs in one day

Version 1.0 described a simulation framework in the present tense, said its Weibull parameters were calibrated to space-grade electronics and cross-validated against Voyager, and called the design validated in simulation. None of that framework is built. No parameter is calibrated. Layers 0 and 2 to 5 are not built, and Layer 1 is built in part. Version 1.1 says each of these in a section called Errata and Changes from Version 1.0, seven items long. “Validated in simulation” is now “to be validated”.

What went in its place is what exists. A no_std crate with nine Kani proofs. The dusk table with its two new baselines, including the one where the shrinking quorum loses. The bench, 150 pairs of runs with identical counts on three processors, and the known-answer coverage, including the half of stuck-at bits that one test input cannot see. A prior art subsection cites STAR, the Shuttle and deep-space fault protection, and says what Mru adds is narrower than version 1.0 implied. Em dashes were replaced, and the processors are called processors.

v1.1 was archived on Zenodo at 20:50. Its PDF put status labels on the layer boxes in Figure 4, and they overlapped the side arrows. Three panel headings in Figure 5 sat on different baselines, the rollback loop in Figure 7 was too loose, and a label in Figure 8 crossed the phase dividers. v1.1.1 fixed the four, two references’ capitalization, a stray comma and one heading, and was archived at 21:31. The text is unchanged from v1.1, and says so.

Three version DOIs, and a concept DOI that always resolves to the latest. The research pages cite 1.1.1. The GitHub profile card cites the concept, so it stays right through 1.1.2.