Roadmap to v1.0: stable Workshop tooling platform

[Teakowa]

Goal

Define Wright's pre-1.0 roadmap by user-visible product readiness and real workflow evidence rather than internal migration phases.

Current execution state

The canonical ownership cutover is now substantially complete:

• raw Workshop convergence is complete through released workshop-rs contracts;
• opy-rs#8 completed the first OPY→Workshop compiler baseline, with opy-rs / opy-compiler v0.1.5 released;
• wright#228 merged the source-language ownership cutover and consumes released OPY and DEL owner contracts without Wright-side language-semantic fallbacks;
• the hosted Wright integration gate passed OPY owner integration, Workshop real-project integration, DEL/OSTW owner integration, benchmarks, scenario validation, LPP, CLI smoke, and distribution validation on the merged cutover head.

The current OPY frontier is no longer #42 → #36 → #37 → #38. Those baseline stages are complete. Residual executable gaps are now attributed by owner:

• opy-rs#88 — residual OPY frontend/full-project corpus gaps;
• opy-rs#89 — residual native-WIR lowering divergences;
• workshop-rs#123 — missing/uncertain canonical WIR carriers exposed by OPY compatibility evidence;
• workshop-rs#124 — canonical enum-domain ambiguity blocking direct semantic-WIR oracle comparison.

Wright release PR #233 is the pending distribution boundary for the merged ownership cutover; until it lands, GitHub's latest published Wright release remains older than the merged integration reality.

Tooling-first product order

Wright remains the unified tooling/product layer over independently owned language implementations. Product priority remains:

1. check / diagnostics
2. lint / static analysis
3. inspect / semantic query
4. validated source edits / refactoring
5. agent / embedding workflows
6. CI / language services
7. Workshop stability / cost analysis
8. compilation / conversion as required by real workflows

Compiler work is product-enabling infrastructure, not the sole progress metric. Wright must not claim source-language behavior beyond what the owning implementations actually support.

Ownership and dependency rules

• Raw Workshop semantics/catalog/settings/WIR/emission gaps belong to workshop-rs.
• OPY syntax, preprocessing/macros, semantic model/HIR, compiler behavior, and scoped reconstruction belong to opy-rs.
• DEL/OSTW source/runtime/compiler semantics belong to deltin-rs.
• Wright owns orchestration, diagnostics aggregation, lint/analyze/inspect, source-edit product behavior, LSP/CLI/agent surfaces, and cross-language integration.
• language-provider-protocol is an integration/process protocol, not a language-semantic owner.
• Missing semantics must be fixed in the owning repository rather than compensated for in Wright.

For cross-repository work, use:

owning repo -> owner contract/tests -> consumable release/contract -> Wright integration -> boundary/real-workflow verification

Compatibility and conversion direction

• Target observable/canonical semantic equivalence, not generated-text, temporary-variable, formatter, or optimizer-internal identity.
• Workshop remains the canonical normalization hub for source-language compilation and reconstruction workflows.
• Source-preserving edits follow semantic understanding → validated source edits → original source; full-file regeneration is not the default mutation model.
• Workshop→OPY/DEL reconstruction, where pursued, targets useful semantic reconstruction with explicit information-loss boundaries rather than literal source recovery.
• Known gaps must remain explicit; a supported product surface does not imply unsupported owner semantics are available.

Near-term direction

• resolve workshop-rs#124 where canonical enum-domain ambiguity blocks stronger OPY semantic comparison;
• resolve/design workshop-rs#123 before any OPY workaround for constructs the current canonical WIR cannot represent losslessly;
• continue opy-rs#88/#89 only at the correct owner boundary after Workshop prerequisites are available;
• release the merged Wright ownership cutover through chore(main): release 0.2.16 #233 and re-verify the user-consumable distribution boundary;
• keep DEL expansion, protocol work, architecture cleanup, docs, and UX polish secondary unless current real-project evidence promotes them.

v1.0 product contract

Wright 1.0 is ready when users can rely on it as a stable Workshop tooling platform with:

• cross-platform installation/update;
• stable declared check, lint, analyze, inspect, source-edit/fix, conversion, and agent/embedding surfaces;
• canonical raw Workshop parsing/validation/emission through workshop-rs;
• evidence-backed OPY and DEL/OSTW workflows through independently usable owning implementations without hidden upstream runtime fallback for declared first-party support;
• validated source-oriented mutation with semantic refusal for unsafe operations;
• corpus-defined support claims and reproducible implementation/catalog identities;
• stable CI and machine-readable diagnostic/tool contracts;
• reviewed licensing/provenance/distribution boundaries;
• documentation sufficient for users without knowledge of internal compiler architecture.

Planning rules

• Real user workflows are the primary progress metric.
• Unit tests, snapshots, minimized fixtures, support matrices, and issue state are regression/planning evidence, not substitutes for rerunning motivating real projects.
• Keep near-term work concrete and evidence-driven; keep distant work coarse.
• Do not split catalog-scale work into per-symbol issues.
• Do not let architecture cleanup, protocol work, CLI polish, documentation volume, or dependency churn outrank missing engine semantics that block real tooling workflows.
• Repository code, living docs, issues, PRs, and public interfaces should use concrete capabilities/versions rather than temporary internal phase labels.