Superstring construction, byte objectives and practical packaging

Family 128. Edition: 9 October 2026. Source metadata. Revision fd4aeeb2ee4fc729c18d98444fed42fd0529eeeb.

Source and enabled use

The constructed forced-count/layer/request/cycle/Euler algorithm claims factor two against unrestricted optimum in symbol length. All seven main argument files (627 lines) and selected scope/config/Comparator/65-line solution entry point were read. Figures/bibliography, imported proof closure, source execution and independent acceptance remain open.

Practical boundary

An immutable indexed byte-view build plugin is a possible application. Containment does not provide drop-in zero-terminated strings, pointer identity, mutability or minimum compressed artifact bytes. Existing GCC/linker pooling and compression are strong incumbents. Source substring/rule enumeration is polynomial but no practical construction cost is established.

Current business decision

Prototype a bounded complete-artifact comparison; defer a source-guaranteed backend. Commercial confidence is low: a public 48-title sample saves 28 raw bytes, gives no gzip reduction and grows XZ by 32 bytes. A synthetic periodic sample saves bytes, already through ordinary greedy. Assumed specialist delivery leaves a weak margin before overhead. Updated dossier.

Concrete evidence

1,445 controls compare 469 tiny collections with all binary output words through length nine, verify paper-derived forced counts within that finite domain and check indexed artifact round trips. Seven saved public/runtime/synthetic comparisons emit 31 SSP1 artifacts. The conventional greedy/DP and source count stage are different from the unimplemented full factor-two construction.

Next action

Measure an actual incumbent length-aware build representation, workload/API fit and delivery effort before claiming savings or demand. Review/implement remaining source stages only with exact correspondence and realistic cost evidence.