Rails partials are a great way to reuse view code, but they have traditionally suffered from one weakness: their interface is implicit.
When opening a partial written by another developer, it is often unclear which locals are required, which are optional, and whether all of them are still used. IDEs typically cannot help much either, often showing warnings about unresolved variables because they cannot determine where the values come from.
The problem becomes even more apparent as an application grows and partials are rendered from multiple places.
If a local is forgotten by call, the error only appears when the template is rendered:
undefined local variable or method `missing_local'
If an extra local is passed, Rails traditionally ignores it.
Over time this creates a situation where the real API of the partial exists only in the heads of the developers maintaining it.
Rails Strict Locals
Rails provides a feature called strict locals that allows a partial to declare its expected interface:
<%# locals: (title:, highlight: false) %>
The declaration resembles Ruby keyword arguments and is placed at the top of the template.
A local like title without a default value is required. Locals like highlight with default values become optional
The partial now documents and enforce its own API. If a required local is missing, Rails raises an exception instead of failing later when the variable is accessed. Likewise, if a caller provides a local that is not declared, Rails rejects it.
Conclusion
Strict locals do not fundamentally change how partials work, but they make them easier to understand and maintain.
By declaring the expected locals directly in the template, partials become self-documenting and gain an explicit contract with their callers. Missing locals are detected early, obsolete locals are rejected, and developers no longer have to search through controllers, parent templates, and render calls to understand where variables come from.
An additional benefit is improved tooling support. Once the interface of a partial is explicit, IDEs can understand the available variables much better. Your IDE becomes a helpful companion again rather than a source of noise.
Schneide Blog 