Skip to content

0002 — The .tether store is an SMD superset; round-trip is a data-model property

  • Status: accepted
  • Date: 2026-06-24
  • Deciders: bioedca
  • PRD anchor: §5.3 (interoperability), §7.3 (round-trip); FR-ROUNDTRIP, FR-IDEALIZE
  • Milestone: M0 (declared) → M2 (realized)

Context and problem statement

Tether must interoperate with tMAVEN (SMD-HDF5) and let users hand traces off to the standalone GUI and bring idealizations back — while never losing the trace ⇄ movie link. How do we reconcile a standard interchange format with the extra coordinate/patch provenance the round-trip needs?

Decision drivers

  • tMAVEN reads/writes the SMD dataset/{data,sources,tMAVEN} container [Greenfeld2015]; a hand-off must open directly in standalone tMAVEN.
  • The return leg cannot trust IDs/order (the user may reorder/subset in tMAVEN).
  • Round-trip must survive even when coordinates are absent on re-import.

Considered options

  • A. .tether as an SMD superset — export the exact SMD container with Tether's coordinates as ignored superset metadata; re-import by exact intensity-trace matching, ID/order as a hint only.
  • B. A bespoke format with a converter to/from SMD.
  • C. Store only SMD and reconstruct coordinates heuristically.

Decision outcome

Chosen option: A. The round-trip is a property of the data model: a molecule is re-identified by matching its raw intensity trace, so a reordered or subset SMD still re-imports non-destructively as a new /idealization/{model}.

Consequences

  • Good: clean tMAVEN interop; non-destructive re-import; resilient to user edits.
  • Trade-off: an intensity-matcher must be implemented and tested against reordered/subset inputs (M2 S7).
  • Follow-up: round-trip integrity is exercised by oracle (f) and the M6 consensus/ebFRET round-trips.

More information

PRD §5.3, §7.4, Appendix D.1; ADR-0001.