I thought the Open Geospatial Consortium's standards would reflect dominant practice, that turns out to have been the wrong assumption [1]
The only geospatial software I've used is Proj4J, which normalizes with a function called toENU.
I guess I've only encountered weird stuff then.
[1]: https://lists.opengeospatial.org/pipermail/coordtran.wg/2006...