Nishi Family › Compare › Code Genealogy and Provenance (Genealogist)

Nishi vs the Field

Code Genealogy and Provenance (Genealogist) — mechanically measured, liar-killed, sovereign.

Nishi vs Sourcegraph and CodeQL and jscpd and dependency-cruiser

How this is scored. Every Nishi cell is measured: the generator reads the real organ source on disk and requires the implementing symbol to exist (no self-grading). Competitor cells record documented capability presence. Best=leads this axis, Yes=present, Part=partial, No=absent. This is capability presence, not depth or scale: the majors lead on index size and neural ranking. Nishi's genuine exceeds are the sovereignty / neutrality / determinism axes.
CapabilityNishiSourcegraphCodeQLjscpddependency-cruiserNotes
Organ ancestry and relatedness graphYesYesNoNoNogen_register + gen_relatedness build the module family-tree (relatedness by shared ancestry); ancestry momentum measured 2673 on the 2026-07-10 radar = the hottest frontier signal; the field tracks references + git blame, not ancestry
Genesis lineage traced to the seedYesNoNoNoNogenesis_lineage.tsv + idea_genealogy.tsv record where a capability CAME FROM (its genesis), not its first commit -- provenance to the root; no code-intel tool ships genesis lineage
Caller and blast-radius analysisYesBestPartNoPartSovereign whole-tree reference search, no timeout on 89k files; powers the checked rename method (22 collisions to 14, zero loss, 2026-06-23); Sourcegraph precise refs lead cross-language
Duplicate and shadow-name detectionYesNoNoBestNoFinds organs that run the WRONG code by name (name-shadowing); jscpd is the token-clone bar; shared with the janitor census
Prioritized dedup / resolution worklistYesNoNoYesNoRanks every collision by external-caller count so the safest are resolved first; the field reports clones, it does not rank the fix order
Dependency graph of the module treeYesYesYesNoBestlineage_graph.tsv is the sovereign dep graph; dependency-cruiser + Sourcegraph lead on precision + visualization
Semantic code-graph navigation (go-to-def / refs)NoBestYesNoYesSourcegraph cross-repo code intelligence is the bar; ours is textual reference search, not a language-server graph
Semantic dataflow / dependency queriesNoYesBestNoYesCodeQL treats code as a queryable database (dataflow, taint); we have no query engine
Precise typed call-graph (AST + types)NoBestBestNoYesReference search over-reports (safe but imprecise); a type-resolved call-graph is the named next rung
Cross-language analysisNoBestPartPartBestSourcegraph/CodeQL span 20+ languages; our genealogy is NishiLang-shaped (extension-level for others)
Coupling and cohesion / architecture metricsNoYesYesNoPartNDepend-class coupling scoring; dependency-cruiser rules are partial; absent here
Clone detection at token / AST granularityNoYesYesPartNojscpd token-hash clones are the bar; ours is name-shadow only, not body-similarity
Visual dependency graphNoBestYesNoPartdependency-cruiser + madge render SVG graphs; ours is TSV data, no diagram
SBOM and supply-chain attestationNoYesPartNoYesin-toto / SLSA provenance attestation; ties to the janitor SBOM gap (mom 1864) and the warden artifact-policy gap
Historical churn and hotspot analysisNoBestYesNoNogit-history churn/hotspot (CodeScene-class); we have lineage but not time-series churn
IDE / CI integrationNoPartBestYesPartSourcegraph editor extensions + CodeQL Actions are the bar; our genealogy is console + registry
Genesis provenance to the seed (ancestry to the root, not the first commit)BestNoNoNoNoThe spore/genesis lineage records a capability's true origin back to the seed idea; git-based tools start at the first commit -- we start at conception
Sovereign whole-tree reference search with zero timeoutBestNoNoNoNooc_grep walks all 89k files with no external index, no timeout, no SaaS -- the caller/blast-radius engine that made the checked collision resolution possible; the field needs a hosted index
coverage 444/1000EXCEEDS 2PRESENT 6ABSENT 10
Honest verdict. The Genealogist is FamilySearch-for-code, and its signature capability is one the code-intelligence field does not have: it traces a module's ancestry all the way back to the SEED. genesis_lineage and idea_genealogy record where a capability came from -- its genesis, not just its git history -- so provenance runs to the root, not to the first commit. On top of that it keeps the family tree clean: a relatedness graph, a sovereign whole-tree reference search that does not time out on 89,000 files and drove the collision resolution that took 22 shadow collisions to 14 with zero functionality lost, and duplicate/shadow detection with a prioritized worklist. What it lacks is the field's semantic depth: no language-server code graph (Sourcegraph), no dataflow query engine (CodeQL), no typed AST call-graph, no cross-language analysis, no coupling metrics, no visual dependency graph, no SBOM attestation, no churn/hotspot history. Our analysis is textual-reference and NishiLang-shaped. The climb: an AST-precise call-graph, then coupling metrics, then SBOM attestation (which also serves the janitor and warden). Lineage momentum is the hottest signal on any radar (3121) -- this crew sits on the field's fastest-moving frontier.

Generated by nx_swcompare_matrix (sovereign NishiLang organ) from knowledge/compare/genealogy.matrix — every Nishi cell verified against real organ source on disk. Zero JS, zero trackers.