Basic layout gap: cannot produce readable 2-row/3-column style tiling for common browser workloads on large monitors

[nostitos]

Problem

AeroSpace still lacks a basic readability layout behavior for common workloads on large displays.

On a 55" monitor, opening multiple browser windows in one workspace often degenerates into narrow, hard-to-read panes (effectively one-dimensional stacking), even after flatten-workspace-tree, balance-sizes, and layout toggles.

This is a foundational UX gap for a tiling WM.

Why this matters

This is not an edge case. A typical workflow is:

• one workspace for browsers
• 4-8 windows/tabs in active use
• need readable, predictable arrangement without manual tree surgery

Today, AeroSpace can tile all windows, but it cannot reliably produce a practical “human-readable” arrangement (e.g. 2 rows x 3 columns style outcome) from normal usage patterns.

Repro

1. Use a wide display (55" TV / ultrawide type setup).
2. Focus one workspace.
3. Open 6 browser windows.
4. Run:
   • flatten-workspace-tree
   • balance-sizes
   • layout tiles horizontal vertical (or equivalent toggles)
5. Observe final arrangement.

Actual

• Layout frequently remains effectively one-axis / skinny columns.
• Windows are technically tiled but not readable.
• “Normalize” actions often appear to do little in practical terms.

Expected

At minimum, AeroSpace should provide a basic readable layout policy for N windows, such as:

• row/column-oriented normalization (e.g. rows, columns),
• or a deterministic multi-lane normalization strategy,
• or built-in policy presets for common counts (4, 6, 8) that prioritize readability over strict tree history.

This is table stakes

Other tiling workflows provide a practical “don’t make this unreadable” baseline without requiring heavy manual join-with choreography.

AeroSpace is excellent in many areas, but this gap makes it hard to recommend for high-density productivity on large monitors.

Related

• Feature proposal: More sophisticated per-workspace normalizations to replicate dynamic twm #260 (advanced/alternative normalization discussion)
• Tiles only happen in one orientation #308 (orientation behavior, closed as duplicate)