Perhaps the best practical argument for using output sharpening, IMO, is to consider:
the same working image (as processed, adjusted, externally worked on, whatever) might be output in multiple different ways at different times, including at different scales even when it's the same hardware and paper type, or to different pixel dimensions when it's digital.
The printer driver / export settings control the resolution etc, and an output device / paper profile can be imposed... for managing all the output specific aspects and output specific sharpening is part of that.
This all helps some "general purpose" detail processing as seen in Develop, be delivered towards a visually consistent result. The point of output sharpening is that it belongsto a given form of output; not to an image. It adapts itself to whatever makes this form of output different from others; most importantly, to calculate and compensate the particular scaling / resampling factor involved.
If your Develop sharpening is done with a particular output usage in mind, you are IMHO potentially missing a trick.
