Here's why I'm worried about Core Class Creep: a multitude of base classes dilutes the entire purpose of a class-based game system.
True, if you treat the "non-core" core classes the same as the "core" core classes.
(You know what? That's going to get confusing. For purposes of this discussion, I'm going to use "core" classes to refer to the 11 from the PHB, and "full" classes to refer to 20-level classes from other sources.)
I don't treat core classes and full classes as the same thing. The core classes are still the ones everything is based on/around. I allow full classes in my games under one of two circumstances.
1) The core classes fail to accurately model a player's idea. If someone wants to play a swashbuckler type, you
can mimic that with a fighter/rogue combination, but it's not really entirely accurate. In this respect, I use them like kits from 2E, only balanced.
2) The core classes are inappropriate to model the quirks of a specific culture. If playing in a renaissance-era game, a swashbuckler makes
more sense than a fighter. If playing in certain sorts of less advanced campaigns, the ancestral speaker from the alternative clerics article in Dragon (see my own horn, see me toot it

) might be better, culturally, than the normal cleric. In these sorts of instances, I use full classes to
replace core classes.
In neither instance, as you see, is there a great deal of diluting going on, because the full classes are used to address specific needs, rather than simply diluting the pool, so to speak. Obviously, not everyone uses them this way, and my point isn't that everyone should. Rather, it's that full classes can provide options and opportunities without weakening--or even, in a specific campaign, expanding--the class system.