Span-First C#: Designing Around Span<T>

An in-depth guide to Span<T> and its ecosystem in modern C#, advocating for "span-first" design in performance-sensitive code paths.

Span<T> is a ref struct wrapping a pointer and a length — a type-safe, bounds-checked view over contiguous memory (arrays, stack, or unmanaged) that never allocates a copy. Its heap-storable counterpart, Memory<T>, enables the same zero-allocation pattern across await boundaries. The article covers the full type hierarchy (Span<T>, ReadOnlySpan<T>, Memory<T>, ReadOnlyMemory<T>), ref struct constraints (no fields in classes, no closures, no iterators), and the escape hatches in MemoryMarshal for reinterpreting spans without copying.

The core argument is that parsers, serializers, and protocol codecs should accept ReadOnlySpan<T> as their canonical parameter — array and string overloads then become thin wrappers, since both implicitly convert to spans. A worked CSV parsing example shows the span-first approach running ~4× faster with zero allocations versus string.Split, eliminating Gen0 GC pressure on hot paths. The article also covers common pitfalls (assuming slices are free of bounds checks, treating ReadOnlySpan<char> as a dictionary key, forgetting ref structs can't implement IEnumerable<T>) and is clear about when span-first does not pay off: object graphs, low-volume call paths, broad public library surfaces, and code that must cross await inline.

Sources: raw/span-first-csharp Last updated: 2026-08-01