Slate v2 Remaining Scope Re-Interview

Read

The remaining True Slate RC scope is clearer now than the roadmap said.

What is settled:

• this is a rewrite, not a nostalgia port
• capability parity matters more than literal legacy semantics
• the best design wins if it is provably safer and more valuable
• every deviation must be justified in proof, PR description, and migration docs

Decisions

1. Live normalization policy

Best rule:

• ordinary live commits guarantee the invariants needed for correctness
• they do not need to force full canonical Slate shape
• heavier canonicalization may happen on:
  • load / replace / reset
  • import / paste ingestion
  • explicit Editor.normalize()
  • app-owned custom normalizeNode(...)

Why:

• this keeps the transaction-first engine stable
• it avoids rebreaking mixed-inline clipboard, refs, and React selection
• it still preserves capability parity through explicit canonicalization seams

2. Non-canonical live shapes

Allowed only when they earn their keep.

Required rule:

• each allowed non-canonical live shape must have a named record that says:
  • why it exists
  • where it is allowed
  • what boundary canonicalizes it
  • which proof rows keep it safe

Why:

• maintainers need auditability
• otherwise “flexibility” just becomes undocumented sludge

3. Collaboration lane

Treat collaboration as an architecture/invariants lane, not a “go build more demos” lane.

That means:

• prove operation semantics and invariants
• do not block the roadmap on more collaboration showcase examples

4. Next mainline blocker

After the current normalization work, the next mainline blocker is:

• broad API / public-surface reconciliation
• plus major file/test deletion closure

Why:

• this is the most maintainers-facing remaining gap
• it is where missing Slate value is still most likely to show up as real cuts
• collaboration remains important, but mostly as an architecture proof lane

Remaining Risk

Normalization is not “done”.

What remains risky:

• broad always-on live inline-container coercion
• default live adjacent-text cleanup beyond the current explicit seam

Those are only worth taking if they survive the full clipboard + range-ref + slate-react runtime proof stack.

Recommended Next Order

1. finish the current normalization family only where the seam is already provably safe
2. move to API/public-surface + file/test deletion closure as the mainline blocker
3. keep collaboration as an invariants/design lane
4. only return to broader live normalization if a better design survives proof

Consequence For The Roadmap

The roadmap should stop saying “normalization is the immediate next move” as if it is the only blocker left.

The honest phrasing now is:

• normalization remains an open risky lane
• but the next mainline blocker after the current safe tranche is API/public-surface reconciliation plus file/test deletion closure