Malzig brings up a good point that when adding chalk (calcite), the calcium does counter the effect of the carbonate. Unfortunately, the formulation that he used was incorrect.
While chalk does contribute 105.8 ppm Ca, it also provides 158.4 ppm of CO3 (not HCO3). The equivalent concentration of HCO3 is 322 ppm. 1 ppm of CO3 is equivalent to 2.033 ppm HCO3. Although the chemical formula for chalk (CaCO3) says that its supplying CO3 to the solution, at the pH of typical drinking water, all the CO3 is immediately converted to HCO3 in solution.
So, the alkalinity contribution is 264 ppm for 1 gm/gal chalk. Plugging that into the RA formula with the 105.8 ppm Calcium addition and the resulting INCREASE in RA is 188 ppm when adding chalk at 1 gm/gal. This compares to the 156 ppm RA increase that baking soda provides when added at a rate of 1 gm/gal.
Regarding the addition of flavor ions to the beer, there is no reason to add them to the mash or sparge water since they aren't a large participant in the mash chemistry. Gordon brings up a good point that not all the ions will make it out of the mash into the wort kettle and adding the minerals directly to the wort kettle makes sure that they make it there. But that brings up another consideration. If the natural water from a major brewing center (ie. Burton, Dublin, Munich, etc) was used in their historic brewing context, then some of those ions that naturally exist in that water would not make it through to the wort. So, there might be some reason to add minerals to the mash and sparge water. I suppose there is the possibility that adding minerals could overdo the flavor effect, but its probably a small discrepancy. Just figured I'd through that out there. ;-)
Gordon brings up another good point regarding the addition of dark grains in a mash. I agree that if you're dealing with low alkalinity water (not soft water as Gordon indicated) and would need alkalinity to keep the mash pH from dropping too low with a dark grist, then adding the dark grains at the vorlauf stage makes sense. I know that Gordon uses RO for his brewing, so he is very adept at this technique (his pile of brewing medals attests). I perform a relatively fast vorlauf at about 15 to 30 minutes and have never done dark grain additions that way. I wonder if that short of a contact time is sufficient to transfer the flavor and color contributions???
For the great number of brewers out there that DO NOT have low alkalinity water, recognize that you will NOT want to do the pale grain mash separately and add the dark grain at the vorlauf. You'll want the dark grain in the mash the entire time to help take out that excessive alkalinity.