Primary references
These sources support the standards and technical explanations in this guide. Color Pick recommendations and product-specific limitations are identified separately in the article.
Compare two color-token releases, detect breaking changes and probable renames, test theme regressions, and create a safer migration plan.
Compare token paths, values, types, aliases, deprecations, and resolved theme output before replacing a design-token release. Treat removed or renamed semantic and component tokens as potentially breaking, then rerun contrast and component-state audits across every supported theme and brand. A value match can suggest a rename, but it is an inference that requires human review.
A small token-file change can alter many components through aliases and theme overrides.
Renaming a primitive may break a build. Changing a semantic token may compile successfully while altering buttons, links, charts, and focus indicators. A dark-theme override can be missed when only the light theme is reviewed.
Migration review therefore needs structural comparison and resolved visual evidence.
Inventory additions, removals, changes, deprecations, and likely renames.
| Change | Typical risk | Review |
|---|---|---|
| Added | Low | Confirm naming and ownership |
| Removed | High | Find every consumer |
| Changed semantic value | Medium–high | Review all inherited components |
| Probable rename | High | Confirm intent and update imports |
| Deprecated | Low now, higher later | Publish replacement path |
Source-token equality does not prove resolved equality.
Resolve light, dark, and high-contrast contexts in both releases.
Repeat the resolution for each supported brand primitive set.
Compare component-state output, not only semantic tokens.
Report broken and circular aliases before calculating visual metrics.
A migration is safer when measurable behavior is compared before and after.
Recheck body text, secondary text, links, buttons, input boundaries, focus indicators, badges, tables, and tooltips. Report the worst result, not only the average.
State coverage should also be compared. A renamed disabled or error token can disappear from a component while the default state remains correct.
Developers need explicit replacements and release impact.
{
"deprecated": "color.primary",
"replacement": "semantic.action.primary",
"breaking": true,
"reason": "Role-based naming and multi-theme resolution"
}Automated comparison cannot know every design decision.
A tool may infer a rename when an old and new token share the same value and layer. Two unrelated tokens can also share a value, so the suggestion must be confirmed by a person who understands product intent.
Token evidence also cannot confirm that every application has updated its imports, CSS output, visual snapshots, or runtime theme switcher.
Paste previous and proposed token sets, review possible breaking changes, and keep the same workspace for contrast and state audits.
It may not break an import path, but it can produce a wide visual change through aliases. Treat it as a behavioral change and review all resolved consumers.
A comparison can pair a removed token with an added token that has the same value and layer. This is only a heuristic and should be confirmed against source history and product intent.
Usually not. Publish a replacement, migration window, usage search, and removal date that match your release policy.
Build output, aliases, themes, brands, contrast, component states, snapshots, interactive focus behavior, and real product screens.
These sources support the standards and technical explanations in this guide. Color Pick recommendations and product-specific limitations are identified separately in the article.