Apple has a well established history of quickly (some might say ruthlessly) depreciating and moving OS components and APIs forward. The upside of this is that developers don't have to experience the anguish associated with supporting several versions of sub-system for long. OS upgrades for the Mac are cheap and easy, so users tend to run closer to the front line. This is especially true for developers.
The downside is that you're frequently caught off guard. The change from Apple X11 to XQuartz is, definitively, a yak shaving exercise. Bleh.
The approach isn't without trade-offs. In other words, I try not to don my rose colored glasses prematurely, but I've spent enough time on all the major platforms to recognize that there are many advantages to the approach Apple uses here.
It is my understanding that the X11 included with OS X was actually a snapshot of XQuartz at the time of OS release, so the config data is in line with what XQuartz expects. In the Unixy bits, OS X upgrades are a lot like other Unix-like operating systems in that package operations remove the application, but not the configuration.
After installing XQuartz, gnuplot continues to run just like it did under previous versions of OS X. I simply ran the XQuartz installer and got back to work.
Good luck
Apple didn't do the best job of keeping up with some OSS components included in OS X. X11 is a great example of such a component. Outside of OS X, lots of changes were occurring with X windows servers. By including X11 (or even providing their own package), Apple was actually in the way. XQuartz dates back to 2007.
There were five releases in 2010 and 2011. By stepping out of the way, OS X users will have a more up to date version of X11 on their machines. That is a good thing™.