Thank you to all the learned voices for this delightful discussion of what I naively thought was a simple question. And thank you to Arlo Griffiths who gave a concrete example of the dangers of normalizing editions. He wrote:
. . . I imagine that, Max Müller and his successors in editing Vedic texts actually encountered the same spelling features
in the mss. they used for editing the RV and other Vedic texts, as those that were found by other scholars working on classical Sanskrit texts from the same manuscript catchment areas, but that Vedic texts were edited with a lesser tendency to normalize manuscript
spellings. And the result was that their editorial choices took a life of their own in subsequent Vedic scholarship as it got ever more detached in the course of the 20th
century from the manuscript basis of the textual corpus to which this field of scholarship is devoted.
To my mind this gives further weight to Dominik Wujastyk's recommendation in a recent discussion about manuscript transcription standards to produce a diplomatic transcription
along with an edited/normalized transcription.
I think the problem with normalizing texts was brought up on this list over 25 years ago in a discussion abour critical editions.
Harry Spier