Simplification

TECHNICAL_GUIDE.md

@@ -194,17 +194,21 @@ Scale labels appear above chord symbols at collection changes and system starts.
 
 The solo path is `composer.compose → construction.Builder → phrases.Phrases → placement → validate → notation/export`. Source pitch/rhythm relationships seed one realized idea. The composer coordinates whole rhythmic groups, written-chord arpeggios, complete scale routes, deliberate rests and harmonic arrivals into a complete phrase candidate. It considers at most 24 alternatives per incoming history, retaining four histories at common source-derived horizons (up to eight measures). These are bounded search settings, not a library of 24 licks. Boundaries may remain open; source breaths inform them without forcing a cadence at every bar. Introductory identity and current development remain distinct.
 
-Rhythm choices are purpose-aware. Arrival values come from authored arrival cells. Fast subdivision groups have a difficulty budget rather than being selected uniformly from every catalog entry. Sequence indexing retains octave crossings. Catalog families contribute representatives, not probabilities proportional to XML file count. Source notes are never changed by generated rhythm transformations.
+The `directed-phrases-2` constructor chooses the destination pitch, harmonic time, sounding duration and release before the incoming route. It compares direct, on-time, delayed and explicitly contracted anticipated arrivals. Local settling can precede further development; completing arrivals and source-defined handoffs preserve the full selected hold. Adjacent common-tone motion can share a sounding event with an arrival when complete-duration validation allows it. A dominant-to-dominant movement is local settling, not automatic cadential completion.
 
-The complete-phrase evaluator reports continuity, contextual repetition, rhythmic interruption, broad activity feasibility, harmonic exposure, soft register and declared return fidelity. Histories covering the same span use duration-weighted phrase totals. The final report also contains an independent whole-stream diagnostic; it is labeled separately from the selection total. No reference-fitting schema, target-loss term, constructor bonus or optional training dependency remains. This evaluator is an inspectable engineering preference, not a model of listener enjoyment or a guarantee of globally optimal melody.
+Rhythm retrieval retains ordered attack/rest timing and the trailing breath, rather than reducing patterns to attack counts or Boolean categories. A clear return preserves the introduced rhythm. Development can rephrase that rhythm while keeping selected opening intervals; extension can reopen a provisional ending. An answer responds through its register and continuation rather than merely inverting every interval. These transformations affect generated material only. Source notes remain untouched. Fast subdivision groups retain the existing tempo/difficulty budget. Catalog families contribute alternatives without probabilities proportional to file count.
+
+Up to three complete routes remain available for each realized group. The route leading into an arrival knows its actual destination; scalar displacement uses octave-aware scale indices. Earlier groups can retain thematic character while approaching the planned register. Approaches and enclosures are constructed together with their destination before the preceding body rhythm; a small local refinement pass still supports prepared suspensions and resolving colors. Protected identifying events and all committed arrivals remain intact. Optional stable colors have their own explicitly validated pitch pool, so the ordinary-map filter cannot silently make an approved color unreachable.
+
+The complete-phrase evaluator reports continuity, recent phrase repetition, rhythmic interruption, broad activity feasibility, harmonic exposure, soft register, declared thematic relationship and preparation of planned arrivals. Returns compare ordered intervals and exact relative rhythm; development, answers and extensions compare the selected opening intervals while allowing rhythmic change. Insufficient notes cannot satisfy a multi-note identity. Required holds are distinguished from unplanned broad notes inside moving material, and a breath after an arrival is not treated as a broken run. Repetition evaluates attacks and relative figures separately from sustained duration. Arrival fulfillment is a hard check, in addition to the approach preference. Contributions sum exactly to the selection total. Constructor labels earn no bonus. Histories covering the same span use duration-weighted phrase totals. The final report also contains an independent whole-stream diagnostic; it is labeled separately from the selection total. No reference-fitting schema, target-loss term, constructor bonus or optional training dependency remains. This evaluator is an inspectable engineering preference, not a model of listener enjoyment or a guarantee of globally optimal melody.
 
 Both solo roles allow all six strings at frets 1–12 without opens. Overlapping register preferences never impose octave copies or string quotas. Ordinary notes remain in the mapped collection throughout their full sounding interval. Typed color, approach, enclosure, anticipation and suspension rules validate real targets, preparation and exposure. Tied fragments reconstruct one sounding event; reattacks remain distinct. The 21 scale patterns and authoritative written chord semantics are unchanged. Unsupported written harmony fails rather than silently substituting another chord.
 
 Accompaniment remains source-derived and canonical across iterations, with independently chosen sparse voicings and rhythms. It neither reads nor reacts to solo notes. Source-defined participation masks keep accompaniment off the physical guitar while it solos. Source vocals, electric melody, pickup timing and runtime representation are preserved.
 
-Production exports `*-solo1.xml` (acoustic), `*-solo2.xml` (electric), and `manifest.solo`. Optional `python -m tests.solo_compare` accepts references only after completed generation; it cannot affect candidate selection. It preserves annotations and raw encoding and separately applies the confirmed guitar-written octave convention. No exact-output reproduction is promised. Compare whole pitches, onsets, releases, rests and phrase relationships; attack counts alone are insufficient. The previous fitting experiments and `audition/` and `docs/` directories are removed.
+Production exports `*-solo1.xml` (acoustic), `*-solo2.xml` (electric), and `manifest.solo`. Optional `python -m tests.solo_compare` accepts references only after completed generation; it cannot affect candidate selection. It preserves annotations and raw encoding and separately applies the confirmed guitar-written octave convention. No exact-output reproduction is promised. Compare whole pitches, onsets, releases, rests and phrase relationships; attack counts alone are insufficient. The report also exports four distinct completed opening candidates where available, marking which survived into the full solo and showing its evaluation separately from alternatives. The optional comparison command renders these source-only candidates without rescoring or recomposing them. The previous fitting experiments and `audition/` and `docs/` directories are removed.
 
-Source songs, scales and useful rhythm/sequence catalogs remain in `data/`. Versioned general rules live in `data/composition`; optional references stay in `build/`, and compatibility expectations live in `tests/fixtures`. Keep only the current comparison and one compact baseline under `build/solo_comparison`. Cached catalogs are checksummed derivatives of production data; a clean build recreates them. Legal TAB is a feasibility approximation, not a finger-assignment or ergonomic guarantee.
+Source songs, scales and useful rhythm/sequence catalogs remain in `data/`. Versioned general rules live in `data/composition`; optional references stay in `build/`, and compatibility expectations live in `tests/fixtures`. Keep the current comparison and compact pre-change baseline under `build/solo_comparison`; staged score-only assets remain unpublished until validation succeeds. Cached catalogs are checksummed derivatives of production data; a clean build recreates them. Legal TAB is a feasibility approximation, not a finger-assignment or ergonomic guarantee.
 
 Current resource and code migration: