An evolving design language

Pixelgrade Pattern Language

A living language for making coherent, original sites across scales. Use it to read a composition, compare recurring solutions, and assemble a working language for the next site.

Move from whole pages to decisive details.

The map keeps every public stage visible. Solid entries are canonical; the others show where the language is still learning, merging, or changing its mind.

One catalogue, one pattern in focus.

15 patterns in view

What do the stages mean?
Canonical
Reviewed, recurrent, and validated through original composition.
Emerging
The invariant is visible, but a contradiction or composition test remains.
Field Note
A contextual observation whose broader claim is still deliberately small.
Merged, Retired, Contradicted
Useful history remains visible when the language changes its mind.

CandidateDetail / System RuleEmerging

Block Overlap

Two adjacent sections should become one composed moment, yet another border or gap would make the page feel assembled from separate panels.

Context

Use an overlap at a meaningful seam: where evidence enters a claim, a card belongs to two fields, or a portrait makes the next passage personal.

Forces

  • Separation protects clarity.
  • Continuity protects momentum.
  • Layering creates depth quickly, but uncontrolled layering damages hierarchy and responsive flow.

Choose one bounded object to cross the seam. Let the following field claim a measured portion of the previous one, then clear its own content by the same logic. Keep every other edge disciplined so the exception remains legible.

Known expressions

The resolution is a field, not a frozen form. Start from what the situation asks for.

  1. 01

    A proof image straddling from a proposition into a results field.

  2. 02

    One card tucked beneath the tail of a hero.

  3. 03

    A portrait crossed by one line of display type.

Watch for

  • Several elements overlap at once and compete for ownership.
  • Body text crosses media or a color seam.
  • Hand-tuned offsets collapse when copy or viewport changes.

Evidence

3 reviewed contexts

Observed in three reviewed contexts, including a native system expression. Its behavior with volatile copy is still under-tested.

Last reviewed:

Test whether the overlap survives long translated copy and a narrow viewport.