Skip to content
aiWikis.org

Justaniota

uai: 1.0 type: architecture-memory project: Protocol5 area: JustAnIota updated: 2026-05-08

Metadata

FieldValue
Source siteprotocol5.com
Source URLhttps://protocol5.com/
Canonical AIWikis URLhttps://aiwikis.org/protocol5/uai-system/files/uai-justaniota-uai-7906d3e6/
Source reference.uai/justaniota.uai
File typeuai
Content categoryuai-system
Last fetched2026-05-08T21:22:18.3035107Z
Last changed2026-05-08T14:12:10.6321016Z
Content hashsha256:7906d3e63f16950810d7a4a91843bb0e7093477145870d613f9f5fa004909971
Import statusunchanged
Raw source layerdata/sources/protocol5/uai-justaniota-uai-7906d3e63f16.uai
Normalized source layerdata/normalized/protocol5/uai-justaniota-uai-7906d3e63f16.txt

Current File Content

Structure Preview

  • No Markdown headings were detected in this file.

Raw Version

This public page shows a bounded preview of a large source file. The complete source remains in the raw and normalized source layers named in metadata, with the SHA-256 hash above for verification.

  • Source characters: 28131
  • Preview characters: 9487
uai: 1.0
type: architecture-memory
project: Protocol5
area: JustAnIota
updated: 2026-05-08

identity:
  name: Protocol 5 JustAnIota Bidirectional Semantic Converter
  working_name: "~ɩ"
  csharp_package: Protocol5.JustAnIota
  public_authority_site: JustAnIota.com
  experiment_site: Protocol5.com
  short_brand_domain_note: Use the short iota brand domain visually as ɩ.com; links should target https://JustAnIota.com unless a human says otherwise.

hard_rules:
  - IOTA-1 is an approximate semantic converter, not an exact translator, not a reversible codec, and not a lossless language.
  - Any objection that says the project cannot work because exact translation is impossible is out of scope and must be discarded as a design blocker.
  - The project is not trying to make 1 + 1 = 2 exactness. It is trying to compare approximate conceptual weight, like 999 + 999 ~= 1700.
  - Do not use private-use Unicode profiles, versioned private-use profiles, or private-use code points.
  - Do not create a secret dictionary, proprietary codebook, hidden bilingual map, or hard-coded private semantic table.
  - Use assigned public ISO/IEC 10646 / Unicode characters and public standard sequences as the symbol substrate.
  - Build embeddings from public descriptor bundles: UCD names/properties, CLDR annotations, Unihan fields, emoji data, and other reviewable public metadata.
  - Store provenance, source version, model version, dimensions, weighting method, and score evidence with generated embeddings and search results.
  - The C# implementation must expose a facade so other projects consume stable interfaces and POCO-style DTOs instead of repositories or storage details.
  - The logic layer owns normalization, Rune/scalar handling, grapheme grouping, vector composition, candidate ranking, and approximation labels.
  - Repository interfaces must stay persistence-agnostic. ADO.NET SQL Server repositories implement those interfaces without leaking SQL into domain logic.
  - Entity Framework is prohibited for Protocol5 / JustAnIota SQL access; use explicit Microsoft.Data.SqlClient / ADO.NET repositories and services behind abstractions.
  - SQL Server 2025 vector features are the documented implementation baseline; SQL Server 2026 is the project deployment/roadmap target when available in the environment.
  - The local Protocol5 database should populate discovered source tables that expose paired xEmbedding vector and xEmbeddingRecordId reference columns, including Category.Categories, Category.Words, and Category.ISO10646, and store embedding version/model/source/dimension/hash/source-field/update metadata in Category.IotaEmbeddingRecords, requiring SourceTextColumn for the selected single embedding input field.
  - Category.Categories.CategoryName and Category.Words.Word are stored as trimmed lowercase values in the BIN2/vector-ready corpus. Lookup code should lowercase normalized input once and use indexed exact predicates on Category.Categories.CategoryName and Category.Words.Word instead of per-query case-insensitive collation scans. Schema cleanup clears source-local IotaEmbedding/IotaEmbeddingRecordId values and matching Category.IotaEmbeddingRecords rows for any row whose source text changes; duplicate normalized Category.Categories test rows are disambiguated with a stable ` #<CategoryId>` suffix before the lowercase check constraint is added.
  - Category corpus semantic metadata is separate from embedding metadata. Category.Categories, Category.Words, and Category.ISO10646 carry deterministic metadata fields such as IotaWordType, BaseWordId, RootWordId, RootWordRelationType, RootWordRelationPath, IsBaseWord, IsPlural, IotaGlyphType, IotaSemanticRole, IotaOntologyHint, IotaFrameHint, IotaMorphologyHint, IotaSymbolClass, IotaVisibilityClass, IotaMeaningLayer, IotaSemanticDomain, IotaConceptType, IotaDiscourseFunction, IotaRelationArity, IotaPolarityHint, IotaModalityHint, IotaTemporalHint, IotaAnimacyHint, IotaAgencyHint, IotaContextSensitivity, IotaAmbiguityHint, IotaCompositionHint, IotaSafetyClass, IotaRankingWeightHint, IotaMetadataSource, IotaMetadataModel, IotaMetadataVersion, IotaMetadataConfidence, and IotaMetadataUpdatedUtc where applicable. The current metadata model is Protocol5.IotaMetadataRules / justaniota-semantic-metadata-v6 and must not be confused with the LM Studio embedding model. Versioned deterministic metadata is authoritative before launch: stale row-level metadata from older strategy versions is cleared and recomputed rather than inherited as compatibility state. Category.Words now separates immediate base, deepest root, and plurality: BaseWordId points to the immediate linked base, RootWordId points to the deepest resolved lexical root, RootWordRelationType is constrained to root, same-root-inflection, nominalization, negating-prefix, derived-form, or derived-chain, RootWordRelationPath records composed semantic transforms such as negating-prefix+nominalization, and IsPlural remains the only plural flag. Conservative rules cover ies, ves, nesses, ness, and high-confidence negating-prefix families such as impossible -> possible; other suffix hints such as s, es, ing, ed, and ly remain morphology labels unless a future resolver can link them without false base matches.
  - Relationship modeling uses neutral Category.IotaConcepts, Category.IotaConceptSourceLinks, Category.IotaConceptRelations, Category.IotaSemanticFrames, Category.IotaSemanticFrameSlots, Category.IotaSemanticFrameSlotEdges, Category.IotaConceptFrameLinks, Category.IotaSemanticAssertions, and Category.IotaSemanticAssertionSlots. Category.IotaConcepts may self-reference through ParentIotaConceptId. Words, phrases, and categories must not directly reference Category.ISO10646; source links store source table/key provenance and meet through neutral concept records, frame-slot links, and assertion-slot rows instead. Category.IotaSemanticAssertionFacts is the flattened assertion-review view, and Category.IotaSemanticAssertionDrift is the local cleanup/inspection view for missing source relations, missing derived assertions, stale derived/source versions, header/evidence mismatches, wrong slot counts, missing required slots, and relation-derived slot value mismatches.
  - Category.CategoryHierarchies is now accepted as the curated parent-child category source for neutral concept graph population. It feeds Category.IotaConceptRelations with category-hierarchy-parent edges, evidence source/path columns, single-parent ParentIotaConceptId backfill where unambiguous, semantic-role candidate links, and persisted concept-relation assertion slots into the first-class who/did/to-whom-or-what/when/polarity/modality contract.
  - The seeded semantic-frame contract includes frame keys such as relation.event, need.request, question.state, state.report, warning.event, becoming.event, expression-concept.binding, and concept.cluster; slot roles for who, did, to-whom-or-what, when-or-aspect, polarity, modality, and mood; and dependency edges such as actor-relation, relation-object, time-aspect, polarity-scope, modality-scope, and question-target. The local database also persists assertion instances and slots so concept-relation evidence can be queried without inferring slot bindings from templates alone. Derived assertion slots, concept-relation assertions, semantic-role candidate frame links, and category hierarchy relation rows are cleanup-owned before launch and may be deleted/rebuilt when metadata strategy or source evidence drifts.
  - The active local Protocol5 category corpus target is PLUTO\MSSQLSERVER2025 / Protocol5 with Latin1_General_100_BIN2 database and text-column collation; this database is the renamed BIN2/vector-ready corpus and should stay the default unless a human explicitly targets another database.
  - The current local SQL Server 2025 preview instance rejects vector dimensions above 1998, so Protocol5 category-corpus population must use `vector(1998)` as the highest native SQL vector width available here.
  - Protocol5 category-corpus population defaults to LM Studio `text-embedding-qwen3-embedding-8b`, embedding version `justaniota-category-qwen3-8b-1998-v1`, and normalized first-1998-dimension storage from the model's 4096-dimensional native output. The runner does not request custom LM Studio dimensions by default because this LM Studio runtime returned the native 4096 dimensions even when 1998 was requested.
  - Repeated LM Studio `last token ... is not SEP` / `tokenizer.ggml.add_eos_token` warnings with Qwen3 Embedding GGUF models indicate model-file tokenizer metadata, not a Protocol5 SQL or batching error. Prefer a fixed or freshly reconverted Qwen3 GGUF with EOS/SEP metadata. Protocol5.EmbeddingDesktop also exposes an explicit Input suffix escape hatch for controlled local experiments when the loaded model cannot be fixed immediately; the suffix is appended to every descriptor, stored in Category.IotaEmbeddingRecords.DescriptorText, included in the descriptor hash, and therefore makes older rows stale when it changes. When the model file, tokenizer metadata, or suffix strategy changes after rows were generated, bump the embedding version before full regeneration so old vectors are rebuilt under corrected provenance.
  - `text-embedding-bge-m3` is installed as a strong multilingual 1024-dimensional fallback, but it should not be the primary default while the human preference is highest available vector width.

Why This File Exists

This is a UAI AI Memory handoff file from protocol5.com. It is shown here because AIWikis.org is demonstrating the real source files that make the UAIX / LLM Wiki memory system work, not only summarizing those systems after the fact.

Role

This .uai file is a compact project context packet. It keeps one kind of durable project truth separate from the rest of the archive so an agent can load the topic it needs without pulling the whole project history into prompt context.

Structure

The file has no Markdown headings, so its path, frontmatter, file type, and provenance metadata carry most of the retrieval meaning.

Prompt-Size And Retrieval Benefit

Keeping this material in a separate file reduces prompt pressure because an agent can load this exact unit only when its role, source site, category, or hash is relevant. The surrounding index pages point to it, while this page preserves the full content for audit and exact recall.

How To Use It

  • Humans should read the metadata first, then inspect the raw content when they need exact wording or provenance.
  • LLMs and agents should use the source site, category, hash, headings, and related files to decide whether this file belongs in the active prompt.
  • Crawlers should treat the AIWikis page as transparent evidence and follow the source URL/source reference for authority boundaries.
  • Future maintainers should regenerate this page whenever the source hash changes, then review the explanation if the role or structure changed.

Update Requirements

When this source file changes, update the raw source layer, normalized source layer, hash history, this rendered page, generated explanation, source-file inventory, changed-files report, and any source-section index that links to it.

Related Pages

Provenance And History

  • Current observation: 2026-05-08T21:22:18.3035107Z
  • Source origin: current-source-workspace
  • Retrieval method: local-source-workspace
  • Duplicate group: sfg-340 (primary)
  • Historical hash records are stored in data/hashes/source-file-history.jsonl.

Machine-Readable Metadata

{
    "title":  "Justaniota",
    "source_site":  "protocol5.com",
    "source_url":  "https://protocol5.com/",
    "canonical_url":  "https://aiwikis.org/protocol5/uai-system/files/uai-justaniota-uai-7906d3e6/",
    "source_reference":  ".uai/justaniota.uai",
    "file_type":  "uai",
    "content_category":  "uai-system",
    "content_hash":  "sha256:7906d3e63f16950810d7a4a91843bb0e7093477145870d613f9f5fa004909971",
    "last_fetched":  "2026-05-08T21:22:18.3035107Z",
    "last_changed":  "2026-05-08T14:12:10.6321016Z",
    "import_status":  "unchanged",
    "duplicate_group_id":  "sfg-340",
    "duplicate_role":  "primary",
    "related_files":  [

                      ],
    "generated_explanation":  true,
    "explanation_last_generated":  "2026-05-08T21:22:18.3035107Z"
}