Isn't the codebase just the entire Mozilla codebase? Firefox is built inside the main code tree as far as I know. Forking the code and throwing out the parts that aren't used would only make the development process that much more complex, and would not do anything but save a bit of diskspace and bandwidth for developers.
Reducing the memory footprint significantly would require rewrites of the entire rendering engine, which was never the goal of Firefox. Others ARE working on that, however, as seen with the optimized Gecko that's Coming Soon.
The main driving force of Firefox (or at that time, Phoenix) was to streamline the browsing experience. Mozilla was big, slow and bloated (both code-wise and interface-wise). This meant developing a better interface, and upgrading and optimizing the GUI toolkit. Later on, things like extensions were added, but mostly Firefox has always been about the interface.
And the main thing to realize is that the interface is the program. The Mozilla suite may be just as capable as Firefox, but nobody save a few stubborn power users are going to use it unless they feel comfortable with the interface. The real accomplishment of Firefox was that they took an open-soruce project, which are infamous for being lacking in interface design, and made it usable to everyone. And it's all thanks to focusing on the interface.
I can only wish more open-source programs will follow their lead.