About agitation: The purpose of agitation is to bring fresh developer in contact with ALL of the film area in the tank, at the same time. If your agitation is too slow, then you run the risk of uneven developing, and if it's too fast, you could see edges of your negatives more developed than the center of the film.
Well, Thomas, agitation during development seems to require a thread of its own! You say: "if it's too fast, you could see edges of your negatives more developed than the center of the film". And I agree, overdeveloped edges of the film (at least 35mm and 120 film i small tanks with reels) is a common problem. The question is: what's the correct explanation for this overdevelopment? One explanation could be that the developing process speeds up during fast agitation. I doubt that's the full/best explanation. Suppose the total developing time is something like 5-10 minutes, agitation with 30 sec interval, or less or even with continious agitaition; now suppose the time you agitiate is 5-15 seconds, not very much development could be expected during this short period relative to the total development time, even with continious agitation. I think a better, or complementary explanation, is that the agitiation, even if fast, is NOT SUFFICIENT to fullfill your correct condition that agitation should "bring fresh developer in contact with ALL of the film area in the tank, at the same time" -- a problem is, off course, how to estimete this "same time", but suppose 5-15 seconds is enough if total development time is 5-10 minutes (or more). If you agitate fast (like Kodak once recommended 5 inversions in 5 seconds!), sure, there will be a lot of turbulent movements around the reel which hold the edges of the film. The result is that fresh development enter the emulsion (and used products that counter act development is washed out from the emulsion) AT THE EDGES, BUT NOT necessarily at the whole film surface – and this "insufficient exchange of chemicals" could happen both with soft and fast agitation, i.e. IF IT'S NOT sufficient to substitute fresh developer for used products evenly over the whole film surface.
The problem thus seems to be: how to achieve sufficient substitution of frech developer over the whole film surface during the agitaion period?
In my experience the solutioin is not "softness" (if you don't take risks with uneven development, probably due to a mistaken belief of getting too contrasty negatives (try a more diluted developer!), rather the opposite.
BUT it should be done also following your good advise of not filling up the tank; that's VERY important in my experience (and experiments). I always now use a tank double the size of the reels and developer reasonable covering the reels, such that when the tank is inverted the film is free from developer and then new developer enters the film surface when going back; I agitate not just by inverting the tank, I also rotate, back and forth, up and down, to insure that the exchange of chemicals is sufficient all over the film surface, and without some regular pattern of the developer flow over the film surface. With this method I have even been violent with critical negatives without unevenness! When I have had problems it always seems to be being too "soft" or in combination with too regular movements during agitation.
This practice, related to the explanation of "insure sufficient exchange of chemicals", seems to work equally good with all sheet films I have tried (4x5, 5x7, 8x10).
My experience with continious agitation in the JOBO rotating drum (just tried C-41 developing with this equipment) seems to give some support for the "insufficient exchange of chemicals" explanation of unevenness. Developing 120 film with big reels (two 120 film/reel) in a big tank taking 3 reels, the tank is rotating back and forth (up till 10 minutes depending on the temp.). But the fact that the film is moving (even if fast!) in the developer (more or less still standing) is not enough for proper exchange of chemicals; but turbulence around the reel's edges makes a difference, and the result is overdevelopment at the edges by this method. By removing the tank and agitating by inversion as above (each 30 seconds or once a minute), and then back to the rotating movement, the result is good even development – and i think this must be judged quite vigorious agitation.
But, if the negatives are good, the developing method is good!