I hear it mostly from dealers and salespeople, and it still keeps popping up on various forums.

Thankfully, the trend is waning amongst manufacturers. Seven years ago, when Meridian and Lexicon introduced room correction to their respective surround processors, all they talked about was correcting in the time domain (e.g., reducing long decay times) instead of improving the frequency response. Soon after, Audyssey was separating itself from the competition by promoting their ability to measure and correct in the time domain (you can even see it in the literature for the SVS subwoofer equalizer that Audyssey made). Nowadays, those companies barely mention time domain correction, instead just talk about overall improvements.

The good news is that the more research that is done into room correction, the less complicated it seems to become. The most recent round of blind testing from Harman demonstrated that listener preference for room correction came down to two main factors: a perceptually flat target curve (as opposed to measured flat) and smoothness across the range (the fewer peaks and dips, the more it was preferred).