What is the @property decorator, and what problem does it solve compared to plain getter/setter methods?
@property turns methods into attribute access with validation and lazy computation
A property is a descriptor that lets a method be accessed through normal attribute syntax. Its main value is API evolution: you can start with a plain attribute, and later replace it with a computed property without breaking callers, because obj.x still works. It also centralizes validation in a setter and lets you expose a read-only attribute by defining only the getter. The trade-off is that attribute access now runs arbitrary code, which can surprise consumers if it is slow, raises, or has side effects. Properties should be cheap and side-effect-free; anything expensive should be an explicit method.
- 1
Use @property for computed values, validation, and read-only attributes.
- 2
Use @x.setter for validated assignment and @x.deleter for controlled deletion.
- 3
Use functools.cached_property for expensive computations that should be memoized per instance.
- 4
Trade-off: properties hide cost. If accessing the attribute can be slow or fail, prefer a method with a descriptive name.
- 5
Common mistake: using properties for things that mutate global state or perform IO. That violates the principle of least surprise.
- 6
Common mistake: forgetting to make the setter validate and assuming the backing attribute is protected by the property alone.
- 7
Version note: @property is ancient and stable; cached_property arrived in 3.8, and 3.8+ also improved documentation via doc on setters.
Scenario Questions
0-2 years experience
- 1You want obj.area instead of obj.get_area(). Which decorator?
- 2How do you make an attribute read-only with a property?
2-5 years experience
- 1You need to validate that age is non-negative on assignment. How do you implement it?
- 2You add a property that calls a remote API. Why is this a bad idea?
5-8 years experience
- 1You have an expensive computed attribute that is accessed repeatedly. How do you memoize it per instance?
- 2You are evolving a public API from a plain attribute to a property. What compatibility concerns arise?
8+ years experience
- 1Design a configuration class where every attribute is validated, cached, and observable, using properties and descriptors without exploding boilerplate.
- 2Explain how properties, cached_property, and __set_name__ interact with descriptors and slots in a deeply hierarchical model.
Follow-up Questions
- When should a computed value be a method rather than a property?
- How does cached_property differ from property in terms of storage and invalidation?