P3809 R0 Should we make std::linalg reductions deduce return types like fold algorithms?

[wg21bot]

P3809R0 Should we make std::linalg reductions deduce return types like fold algorithms? (Mark Hoemmen)

[ramzanbhutto]

Yes , std::linalg reductions should deduce their return types.
It aligns them with existing fold algorithms (reduce, accumulate), makes mixed-type computations behave naturally (int × double -> double) and removes unnecessary boilerplate.
Only highly sensitive operations (e.g., norms) might justify fixed types, but for standard reductions like sum and dot, deduction is the correct!

[jensmaurer]

This issue tracker is for administrative handling of WG21 papers only; please direct any technical comments to the appropriate e-mail reflector and/or a face-to-face meeting when the paper is discussed.