der_kluge
Adventurer
I agree with his points, though didn't read his article. (thanks for that executive-level summary) 
For d20 publishers, it's a double-edged sword. If a d20 publisher has a campaign setting, they want to provide source material for it so that it doesn't fade away and die. They key way to do that is by creating modules. AU and Dragonstar are both great examples of campaign settings that licensed out their world IP to other publishers to create modules for them. These modules obviously won't sell to people who don't use those campaign settings, but they help ensure the success of a campaign setting in the long run.
That said, I would also agree that I fall into the camp of not buying any world-specific modules. I would never think to purchase a KoK module, or a Scarred Lands module (or any Scarred Lands product, for that matter), because those things don't fit with my campaign world. I do enjoy a good generic module that I can plug into my own campaign, but those are few and far between.
But, creating a module to be generic is difficult. Let me explain. Once upon a time when I was asked to write for a new d20 company called Thunderhead Games, they said "write a module", and I said "I'll see what I can do." So, I had this idea of having the PCs meet each other on a caravan ride from one outpost city to another larger city. "I need a city", I said, "oh, btw, I need some gods, too". "We have neither", was the reply from on high, "make them up". So, I slapped a couple of city-stat blocks together "Perten" and "Bluffside", and I made a couple generic gods.
Thus, Bluffside was born. So, you see, it's almost a natural evolution to go from a module to a campaign setting. Even generic modules are almost always set in some kind of specific place, but generic in this sense can really only mean "standard D&D fantasy", and that's about as close as you're gonna get.

For d20 publishers, it's a double-edged sword. If a d20 publisher has a campaign setting, they want to provide source material for it so that it doesn't fade away and die. They key way to do that is by creating modules. AU and Dragonstar are both great examples of campaign settings that licensed out their world IP to other publishers to create modules for them. These modules obviously won't sell to people who don't use those campaign settings, but they help ensure the success of a campaign setting in the long run.
That said, I would also agree that I fall into the camp of not buying any world-specific modules. I would never think to purchase a KoK module, or a Scarred Lands module (or any Scarred Lands product, for that matter), because those things don't fit with my campaign world. I do enjoy a good generic module that I can plug into my own campaign, but those are few and far between.
But, creating a module to be generic is difficult. Let me explain. Once upon a time when I was asked to write for a new d20 company called Thunderhead Games, they said "write a module", and I said "I'll see what I can do." So, I had this idea of having the PCs meet each other on a caravan ride from one outpost city to another larger city. "I need a city", I said, "oh, btw, I need some gods, too". "We have neither", was the reply from on high, "make them up". So, I slapped a couple of city-stat blocks together "Perten" and "Bluffside", and I made a couple generic gods.
Thus, Bluffside was born. So, you see, it's almost a natural evolution to go from a module to a campaign setting. Even generic modules are almost always set in some kind of specific place, but generic in this sense can really only mean "standard D&D fantasy", and that's about as close as you're gonna get.