type-design-performance

Use a readonly struct for small, immutable data payloads to avoid allocation overhead and prevent defensive copying. Use a class when the type requires inheritance, polymorphism, or represents a large object graph where reference semantics are more efficient.

Optimize .NET collections by selecting FrozenDictionary for immutable lookup-heavy scenarios and IReadOnlyList for exposing data safely. Choosing the appropriate collection type minimizes memory overhead and maximizes execution speed during iteration and retrieval operations.

Making a struct readonly in .NET prevents the compiler from emitting defensive copies when the struct is accessed, reducing memory overhead and improving execution speed. It enforces immutability, ensuring the type's state cannot be modified after creation.

Sealing .NET classes by default improves performance by enabling compiler optimizations and devirtualization. It prevents unwanted inheritance chains, allowing the runtime to bypass virtual method table lookups and execute method calls more efficiently.

Span<T> and Memory<T> reduce allocations in .NET type design by providing safe, managed pointers to contiguous memory regions. They enable high-performance processing of arrays or strings without copying data, maximizing execution speed and minimizing memory overhead.

Structs in .NET type design carry limitations including boxing overhead when cast to interfaces and potential performance degradation from defensive copying if not marked readonly. They lack support for inheritance, making them unsuitable for complex object hierarchies.