Eve library

OSHUN V1 Layout Density Tiers

Spec

Revision 1

On this page

Status: Accepted v1.0
Owner: OSHUN Design + Product + Engineering
Last Updated: March 26, 2026
Implements: V1-UX-051
Related:

  • docs/releases/v1/scope/minimal-chrome-rules.md
  • docs/releases/v1/specs/surface-redesign-ranking.md
  • docs/releases/v1/specs/experience-scorecards.md
  • docs/releases/v1/design/ux-principles.md
  • docs/releases/v1/design/brand-system.md
  • docs/releases/v1/specs/todos.md

1. Purpose#

This document defines the canonical layout-density tiers for OSHUN V1:

  • contemplative
  • exploratory
  • research
  • operator

The repo already contains multiple local density ideas, especially in Studio workspaces that expose compact, balanced, or dense controls. Those local choices are useful implementation signals, but they are not yet one canonical product model.

This document establishes that model so customer and admin surfaces use density intentionally instead of drifting route by route.

2. Why Density Needs A Canonical Model#

Density is not just spacing. It changes:

  • cognitive load
  • scan speed
  • perceived calm
  • confidence in serious work
  • how much chrome the product can support

OSHUN should not use the same density posture for:

  • a Tara recovery surface
  • an Explore browse flow
  • a Veritas or Nisaba evidence workflow
  • an admin incident or queue cockpit

Those are different jobs and need different visual compression.

3. Canonical Tiers#

Tier Primary job Emotional posture Information posture Typical surfaces
contemplative grounding, recovery, orientation, first-run trust calm, slow, breathable deliberately sparse Tara-led home and ritual moments, onboarding moments, restorative empty/degraded states
exploratory browse, compare, discover, choose open, curious, directed moderate breadth Explore, public welcome, discovery modules, activity and library browse surfaces
research read, verify, compare, annotate, preserve context serious, studious, trustworthy high signal with controlled metadata search results, grounded evidence, Veritas and Nisaba detail flows, notebook and compare routes
operator monitor, triage, approve, configure, recover alert, durable, low-noise high-density and scan-first admin shell, queues, incidents, policy editors, release and governance workspaces

4. Tier Definitions#

4.1 Contemplative#

Use when the product should lower pressure and help the user settle into one next step.

Rules:

  • One dominant action area above the fold.
  • No more than 2-3 meaningful content groups visible before scroll on common desktop and mobile sizes.
  • Generous spacing cadence, with section separation that feels obvious without borders.
  • Copy may breathe, but only in short paragraphs.
  • Supporting metrics and metadata stay secondary and often deferred below the first action plane.

Contemplative surfaces should feel composed enough that removing one content module improves focus, not harms it.

4.2 Exploratory#

Use when the user is sampling options, changing direction, or comparing several possible paths.

Rules:

  • Multiple choices may coexist, but one scan path should still dominate.
  • Above-the-fold breadth is allowed, but it must stay legible in one quick scan.
  • Filters and sorting may be visible, but they must not outweigh results.
  • Highlights, featured collections, and domain cues may coexist if their visual weight is clearly ranked.
  • Modules should feel related, not like independent dashboards.

Exploratory surfaces may be broader than contemplative ones, but they still need calm visual rhythm.

4.3 Research#

Use when the user needs to read carefully, compare evidence, preserve context, or move between primary content and supporting reference material.

Rules:

  • Dense information is acceptable, but only if the hierarchy is unmistakable.
  • A primary reading or comparison pane should always dominate secondary metadata.
  • Evidence, provenance, and source metadata must stay accessible without overwhelming the first reading pass.
  • Dual-pane or split-pane layouts are appropriate when they preserve context.
  • Inline actions should stay quiet until the user reaches a decision point.

Research surfaces should feel information-rich without ever feeling admin-like.

4.4 Operator#

Use when the user is working in a high-frequency, high-consequence environment where scan speed and control availability matter more than visual spaciousness.

Rules:

  • High density is allowed and often required.
  • Multi-panel layouts are acceptable if each panel serves an operational role.
  • Tables, queues, diffs, status clusters, and compact inspectors are first-class layout primitives.
  • Context, state, and alerts must be immediately scannable.
  • Decoration is minimized so the surface can tolerate compression.

Operator density is not license for clutter. It is disciplined compression.

5. Quantitative Budgets#

These budgets define the intended feel of each tier before the grid and typography-specific follow-up tasks tighten the system further.

Tier Above-fold content groups Typical desktop structure Typical mobile structure Default internal spacing feel
contemplative 2-3 single dominant column or one calm split layout single column with one dominant action stack generous
exploratory 3-4 one primary stream plus one secondary cluster, or balanced two-column browse layout stacked modules with one active filter layer moderate
research 3-5 reading column plus evidence rail, compare pane, or notebook sidecar stacked reading flow with collapsible evidence/detail sections medium-compact
operator 4-7 two to three coordinated panels, queue/table plus detail, or cockpit-style multi-region layout compressed single-column with sticky summary/action support compact

Interpretation rules:

  • If a contemplative surface needs more than 3 above-fold groups, it is drifting into exploratory density.
  • If an exploratory surface needs more than 4 above-fold groups, it needs harder prioritization.
  • If a research surface cannot preserve a dominant reading pane, it is too dense.
  • If an operator surface cannot show critical state without scrolling, it is too sparse or improperly grouped.

6. Tier-Specific Layout Behaviors#

6.1 Contemplative Behaviors#

  • Prefer single-column or centered split layouts.
  • Keep line-length and section count low.
  • Use space, not borders, as the primary separator.
  • Avoid persistent sidecars unless they carry genuinely essential continuity.
  • Avoid dense metric clusters in the hero region.

6.2 Exploratory Behaviors#

  • Permit multiple browse modules, but maintain one strongest lane.
  • Use carousels or horizontal rows only when they accelerate scanning.
  • Show filter state clearly, but keep advanced controls tucked away.
  • Let visual cues help orientation, but avoid turning every module into a distinct mini-brand.

6.3 Research Behaviors#

  • Preserve reading continuity during compare, citation, and note actions.
  • Use inline metadata, footnote-style references, or secondary rails instead of heavy top-level banners.
  • Prefer split panes over stacked card towers when context continuity matters.
  • Make evidence density adjustable within the surface, but keep the default state focused on comprehension first.

6.4 Operator Behaviors#

  • Use compact controls, grouped actions, and explicit status language.
  • Keep headers short so the viewport is reserved for work.
  • Prefer table, queue, and inspector layouts over decorative summaries.
  • Allow compact mode and balanced mode, but keep the default rooted in operator-grade scan speed.

7. Mapping Tiers To Current OSHUN Surfaces#

Surface family Canonical tier Notes
Customer home recovery, ritual, and first-run moments contemplative Especially Tara-led or calm orientation sections
Public welcome and domain discovery exploratory First impression should feel broad but not dashboard-heavy
Explore, browse, activity browse, and library browse exploratory Breadth is the point, but hierarchy must remain crisp
Search results, evidence, grounded answer support, notebooks, compare flows research High information value without operator cockpit posture
Admin shell, queue triage, policy, reliability, audit, and release work operator Dense, durable, and keyboard-first by default

This mapping should be used alongside the redesign ranking and experience scorecards when assigning target posture to a surface.

8. Mixing Rules#

Most routes may contain local exceptions, but every route still needs one dominant tier.

Allowed mixes:

  • contemplative route with one exploratory recommendation strip
  • exploratory route with one research preview module
  • research route with one compact operator-style moderation or save action rail
  • operator route with one research-style detail pane

Forbidden mixes:

  • contemplative route dominated by operator-style metric density
  • exploratory route that uses contemplative whitespace but still shows five competing modules above the fold
  • research route that treats evidence like dashboard cards
  • operator route that inherits decorative spaciousness from customer marketing surfaces

9. Tier And Breakpoint Rules#

Density must compress responsibly by viewport.

Desktop and large laptop:

  • each tier may express its full structure
  • operator and research tiers may use multiple coordinated panes

Tablet:

  • exploratory and research tiers should simplify to two strong regions at most
  • operator tiers should retain scan speed, but reduce parallel panel count

Narrow mobile:

  • contemplative stays contemplative
  • exploratory simplifies to one dominant browse lane
  • research becomes stacked with collapsible evidence/context
  • operator must not pretend to be a desktop cockpit; it should become a compact triage flow with sticky summaries and explicit action priority

10. Relationship To Existing Local Density Modes#

Current Studio workspaces already use local modes such as compact, balanced, dense, standard, and spacious.

Those local controls should be interpreted through this hierarchy:

  • contemplative allows only spacious or balanced local expression
  • exploratory defaults to balanced and may expose compact only in secondary browse areas
  • research defaults to balanced and may compress selected panes to compact
  • operator defaults to compact or dense and may expose balanced as the calmer fallback

This prevents local density controls from drifting away from product intent.

11. Exit Standard#

A surface has the correct density tier only when:

  • the tier matches the actual user job
  • the first screen communicates the intended emotional posture
  • information volume feels deliberate rather than accidental
  • scan speed and calmness are balanced correctly for that workflow
  • moving the surface one tier denser or one tier looser would clearly make it worse

If a surface could switch tiers without noticeably changing its character, its layout posture is not defined clearly enough.