Normalization by monitor histogram introduces arbitrary scale factors, unless vanadium used same monitor bin size

[SimonHeybrock]

normalize_by_monitor_histogram simply computes d_i/m_i, where i indexes monitor wavelength bins. This means the result depends on the monitor bin size, which is not visible in the final result.

Technically this is not a "bug", if it is ensured that (1) the vanadium reduction uses the same monitor bin size and (2) absolute data values prior to vanadium normalization are not relevant.

Mantid uses an approach that avoids this, I think: https://docs.mantidproject.org/nightly/algorithms/NormaliseToMonitor-v1.html#bin-by-bin-mode.