Surely this is a difference without a distinction?
Given no two people speaking the same language, then, were bound to write this name the same way (or the next time); and said people worked in at least 3 perhaps 4 or 5 different "lingua francas"... can we not accept that the primary sources are understandably "quantum" and even if the phonetic spelling or enunciation varied over time, or between scribes, it's always the same family? This is a case where the coats of arms and transfer of property etc ought to dominate over the (understandable... but modern, and dangerous) desire to standardize retroactively.
If we believe in our "Use Their Conventions Not Ours" policy rubrick, then we ought to tolerate their inconsistency and not impose a false rigor. This happens also with Vere and de Vere versus de Veer and le Veer-- and a thousand other examples (Avenal, Aubigny, Ros, Tosny, Guines, fitzWhatever, Stuart, de La Tour, Bourbon, Burgundy, Hussey, Grey, Pool, Le Strange etc.. ad infinitum) and we need to be super careful that index-standardization doesn't conflate unrelated families with merely similar names. Don't even get me started on the Irish/Cornish/Gaelic/Breton names; or how we handle basically all pre-modern Catholic FNABs, Islamic patronyms, and non Judeo-Christian LNABs from medieval era (say: Andalusia, Hungary, Kievan Rus, Vikings etc). I say this as traditionalist father of a son with a pre-Norman Irish name spelled the later English way, married to my functionalist Franco-Irish 30th & 33rd cousin... we can't even agree about This Stuff inside the family. I doubt there is a right answer for the internet.
The important thing surely is the "graph" of connected relationships. Not their spelling. Let alone the WikiTree indices.