Now I am confused. I have been developing film pretty regularly using 4 ml per roll and getting what I thought were decent results. I do not remember anything in the film developing cookbook that indicated that this was not enough Rodinal to develop a roll of 35mm film. I will have to go back and check this out again.
I am developing 5 rolls of 35mm at a time in a Paterson System 4 tank. This tank needs a minimum of 290 ml per roll of 35mm. I mix 20 ml of Rodinal in a tad less than 1500 ml of solution. No matter how how you cut it that is nowhere close to this minimum 7 ml per roll.
Well, back to the cookbook.
4 ml will develop a film, but it will not develop all films consistently, and some films may be flat overall. More likely, you may find that local areas of high exposure will suffer problems as the developer crashes there. There are different emulsion types that use more or less developer depending on the silver content. And there are exposure differences from one film to the next. If you shot a whole roll of stars against the night sky, then the developer required for that film will be minimal. On the other hand, if you shot an airshow with lots of darkish objects against a bright background, you will require more than the average quantity. My advice is to do your own tests, and see if it makes a difference for you. Over five rolls, the effect of one or two rolls either side of average is diminished. But if you develop only one roll at a time, maybe two, then it is better to use the larger quantity. I cannot say what will work for you, but one cannot make a recommendation to someone who is inexperienced with Rodinal, that is based on living on the edge of what the developer can do. Sooner or later, one is going to have problems with low developer quantities, and it might be the one film where you regret it most.
Another point is that weak developer is more prone to depletion via oxidation. The extent of depletion will be the same per volume of liquid and total tank volume, and will reduce the developer content by a given quantity. The lower the concentration, all other things being equal, the larger the overall effect. With long developing times as are typical for stand development, this creates even more potential for inconsistency. I personally think that stand development is not for beginners, and I avoid it myself for various reasons. That does not mean that it doesn't work, and that it is necessarily complicated. But it does have more variables to consider, and it leaves the door open to inconsistency. Some photographers are happy to tolerate inconsistency, and simply compensate for it after scanning. I personally have an aversion to inconsistency, and will do much to avoid it. It is worth noting how different photographers arrive at wildly differing times for stand development using the same dilutions and films. Furthermore, I haven't seen scientifically backed guidelines for stand development from Agfa or others supplying Rodinal or its clones. For me that places stand development in the realm of alchemy, i.e. proceed with caution.
My advice is based on printing, while for scanning you may prefer thinner negatives and can correct local problems more easily. I rather develop negatives that are good for printing, knowing that they will scan perfectly easily too.
A note on the Cookbook: While it is an excellent resource, there are a few guidelines or methods in there that we don't all have consensus on. Some of the developing times are incorrect, too. Consider the above in light of my personal experience with Rodinal, which may or may not be aligned with the Cookbook's exact quantities, but the sentiment that one shouldn't skimp on cheap chemicals to save a few pennies we certainly share. The Cookbook's section on Rodinal is based on inputs from others (notably Bob Schwalberg), at least that is my interpretation, while Anchell and Troop seem to use different developers for their own work. Bob Schwalberg according to my resources used Rodinal typically at 1:50, 1:75 and sometimes 1:100 with agitation, not stand development. His comparisons with D76 are interesting, and not always borne out by the comparisons done by others. Why that is the case, I cannot say.