@Bryan Gotcha. Nevertheless I’m still wondering if the complexity issues still arises even if one just wants to reapply the autoranging (basically data is there, everything is in place, just range had explicite start/end values before)? After all data changes should trigger a similar data-ranging pipeline (of course only with active auto-ranging).
For the relatively rare case (IME) you describe, I’d suggest adding some invisible glyph to give the data ranges something to chew on. A kludge? I guess, but a simple one.
Out of curiotisity, is adding data later based on user input really such a rare use case? Maybe I’m the exception here but to me this feels like a common secnario, at least in a server setting.