Somewhere around the 40th icon in a set, consistency starts to slip.
Not in the shapes. It slips in the values that look perfectly normal one at a time: stroke width, corner radius, cap style, grid alignment, visual weight.
Why your eyes can’t be trusted
Because on any single icon, each deviation is within tolerance.
A 1.75 stroke next to a 2, a corner radius of 1 next to 1.5, a path that ends at 12.03 instead of 12. None of them looks wrong on its own. Only when they’re all side by side does the set start to read as “a little loose,” and by then it’s hard to say which icons are to blame.
And by the 40th icon, you’ve already looked at them hundreds of times.
If you can measure it, don’t look at it
Our icon review is now a script, and it checks three things:
- Geometry: stroke width, corner radius, caps and grid alignment must each land on one of the discrete values in the spec
- Purity: every shape must come from the existing icon library, with nothing drawn by hand on the spot
- Survival at small sizes: at the smallest size, the details can’t blur into a smudge
If it doesn’t pass, it doesn’t go into the library.
What’s left for your eyes
A script can’t tell you whether the metaphor is right, whether this icon and that one are saying the same thing, or whether you can still recognize it when it’s tiny.
⭐ Give the measurable part to a script, and your eyes have room for the half that needs judgment. With icons, the human review was never about stroke width.
