My understanding is that DI is in fact quite a bit 'cleaner', from an ionic perspective. Unfortunately, DI will not remove many pesky things we may want to be. "Reverse osmosis is capable of rejecting bacteria, salts, sugars, proteins, particles, dyes, and other constituents that have a molecular weight of greater than 150-250 daltons. The separation of ions with reverse osmosis is aided by charged particles. This means that dissolved ions that carry a charge, such as salts, are more likely to be rejected by the membrane than those that are not charged, such as organics. The larger the charge and the larger the particle, the more likely it will be rejected." DI will not remove non-ionic solids or organics.
There may also be additional anion/cation's left in the water, which is something that is mandated to be measured by anyone who deals with water for the public, per the Safe Drinking Water Act, and FDA USP. A coworker says they add a little barium chloride to a sample to see if insoluable salts form, as a test. I do wonder now if any of the DI products on the market are aware or test for this, since the thread seems to be focused on what is probably unneccesary purity demands?
Oh and my buddy mentioned that colloidal silica is also not removed by DI.
While I would love to go DI only, the fact that DI would leave bacteria, excess cations/anions, dissolved organics and solids and this colloidal silica, along with the recharging, it makes me wonder whether or not it is actually cheaper to dump an extra 400% waste water down the drain? I mean, if you think of what it costs to collect, manufacture, transport the lye and acids, and then after using them to dilute them and dump them down the drain, and the impact upon the water supply system, I don't know.
I found the GE site to be a worthy read:
http://www.gewater.com/library/tp/833_What_Is.jsp
With this one in partuclar germane to the topic:
Advances and Changing Costs in Reverse Osmosis and Ion Exchange Systems
http://www.gewater.com/library/tp/733_Advances_and.jsp
I'm happy to note they do in fact mention the caustics cost into the equation.