There was a lovely article in Diver magazine a couple of years ago by Dr. David Sawatzky, who writes extensively about diving medicine and about DCS. The observation he made at the end of the article really stuck with me . . . He said that the big improvement in DCS was made with Haldane's original observation about ratios, and that everything we have done since is fiddling with the edges.
On the one hand, very few people get bent in comparison to the number of people diving. On the other hand, people still get bent when their profiles were computed by acceptable means, and executed perfectly. Which means we still don't have it "right". Pure dissolved gas models (Buhlmann) may seem reasonable to you, and many people still use them without issue. Bubble-based models may end up being "better", but it takes time and a lot of dives to know for sure. And everything is confounded by individual variability and dive-condition dependent variables as well.