Why do commercial SPC software makers still use the pure Shewhart model, without workarounds for the 75% or more cases where the data is NOT independent, and/or not stationary, but requires alarms that trigger adjustments for known causes?

Why do commercial SPC software makers still use the pure Shewhart model, without workarounds for the 75% or more cases where the data is NOT independent, and/or not stationary, but requires alarms that trigger adjustments for known causes?