You can make good chemical and practical arguments on both sides. Chemically, you take a risk if you use a diluted developer (D-76 or Xtol weaker than 1+1, for instance) and barely cover the film with 35 mm. In practice, unless you're doing sensitometry or have a far better eye than I do, you're unlikely to ever notice the difference until you get to dilutions like Xtol 1+3 (and, depending on your standards, maybe not then). The manufacturers have a "minimum active developer" amount for most if not all of the currently available developers, but this level is widely ignored by those who aren't OCD and/or don't consider molecular levels of precision more important than their art (or essential to it).
For myself, I don't generally dilute my stock solution developers in favor of replenishment: Xtol (clone) and D-23. I use Parodinal at the recommended dilutions of 1:25 and much more often 1:50 (which latter still puts about 6 ml of concentrate in a tank with a single 35 mm reel with my Paterson equipment; it'd be 5 ml in stainless). The only reason I've seen to use more developer than needed to cover the film is if I had a single stainless 35 mm reel in a tank that would hold three or more -- which (IMO) is a recipe for flow marks and air bells if you use the minimum developer (too much airspace is as bad or worse than too little).