research-package RP-VE-TCG-2026-DD75

Terminal / Character-Grid Layout Family

RP-VE-TCG-2026-DD75: Terminal / Character-Grid Layout Family

Research question

Can one relational abstraction describe keyboard-first character-grid interfaces (IBM 3270, IBM 5250, DOS text mode, BBS, terminal/TUI, and modern grid applications) well enough that a design system can implement it once and express each style as a profile — without becoming a disguised 3270 component and without sacrificing accessibility?

Scope

In scope: grid geometry, placement, protected/editable runs, traversal, status, action affordances, density, narrow-screen strategy, accessibility, and an optional SequentialReveal presentation behavior.

Out of scope: host data streams, emulation, key-to-action processing, screen transitions, and domain state. Those belong to the consuming application, Limen, and Ordo (CN-VE-TCG-2026-6CA0, rule B-1).

Method

  1. Establish invariants from primary sources and maintained implementations; separate verified facts from background assumptions (GH-17).
  2. Express the family as a relational layout-catalog entry (GH-18).
  3. Define SequentialReveal independently of any profile (GH-19).
  4. Validate against a three-screen reference workflow encoded as data and checked mechanically for the relational rules (GH-20).
  5. Hand the verified relationships and capability gaps to Forma (#37).

Artifacts

Work item Artifact
GH-17 CN-VE-TCG-2026-6CA0, EVR-VE-TCG-001, HYR-VE-TCG-001, JR-VE-TCG-2026-2E0E
GH-18 LAY-TERMINAL-CHARACTER-GRID in the new layout catalog; DF-VE-TCG-2026-2DD5
GH-19 CN-VE-TCG-2026-F9F1 and reference model scripts/sequential-reveal.mjs
GH-20 EX-VE-TCG-2026-5437; fixture content/layouts/reference-workflows/customer-account-inquiry.json; validator scripts/character-grid-layout.mjs

Status

GH-17, GH-18, GH-19, and GH-20 complete as of 2026-09-27. The family is observed: evidence-backed and mechanically validated against a reference workflow, with no conforming implementation verified yet. Conclusion of EX-VE-TCG-2026-5437: supported with gaps (GAP-TCG-07 to GAP-TCG-10). Next: Forma #37 implementation and browser verification; user tests of HY-VE-TCG-2026-9151 and -8750.

Update (WI-0004 to WI-0006, 2026-09-27): Forma #46 implements the family and passed cross-engine conformance, so the family is supported. Reachable-source evidence was added for the 3270 OIA, DOS text mode, and 5250 per-row selection (EV-VE-TCG-2026-1EB5, -D9CB, -8081); GAP-TCG-09 is decided (DF-VE-TCG-2026-DD05); GAP-TCG-10 and GAP-TCG-11 are closed in Forma pull request #47; GAP-TCG-12 (status row versus OIA) is open. The user tests have a preregistration draft (EX-VE-TCG-2026-319C) and have not been run.

Update (WI-0007, 2026-09-28): GAP-TCG-12 is decided by DF-VE-TCG-2026-1320. Grid geometry is the application's presentation space; a profile may declare device status rows after it. The 3270 profile declares one, and the reference workflow's status moves to row 25.

Update (WI-0008 and WI-0009, 2026-09-28): EX-VE-TCG-2026-319C has executable pilot machinery in experiments/ex-ve-tcg-319c, and its operational definitions are fixed by a dated amendment. The pilot (WI-0009) is blocked on participants, consent, approval, a session researcher, and the smallest-effect-of-interest decision. Both hypotheses remain untested.

Update (WI-0010, 2026-09-28): the project owner decided to forgo the pilot because participants cannot currently be recruited, and WI-0009 is abandoned. CG-6 stays a rule on its existing evidence. The narrow-screen guidance stays hypothesis-grade. The machinery remains runnable.

Evidence standard

Claims cite EV-VE-TCG-2026-*. Background that was not verified in an opened source is labeled as an assumption. No user study has been run; all performance claims are hypotheses (HY-VE-TCG-2026-*).