The problem you see with DS-10 and many films of box speed 100 or slower is very distorted sensitometric curves, loss of speed, and often with low density. The developer is adjusted to develop slowly, with suitably adjusted level of solvency, so that the developed silver grains consist of compact filaments rather than widely spanning filaments typical of coarse grain developers. However, this strategy doesn't work well with the way slower emulsions are made.
One way to make DS-10 work "better" with slow film is to dilute 1+2 and use at 25C or higher temperature. But I generally find slow t-grain films to work better with developers of DS-12 type, and that's why I recommend two developers for different kinds of films.
Incidentally, Dickerson and Zawadzki published an interesting article on PT a few months ago. The article was about the myth of the "best developer." Their messages was that the best developer depends on what you mean by best. They also pointed out that each developer has its character, and the "direction" of the character is rather invariant with films, although this is a generalized statement from the data for two unidentified films.
I generally agree with them. One thing I like to add in the context of DS-10 is that, when I made DS-10, I chose to make it work best with 400-speed films and didn't care how it fails with slow films. Commercial film developers cannot do this way.