Back in the day, we licensed/forked Hyperic code and built JBoss Operations Network version 1.x. It was very successful and we learned alot from that effort.
Since then, we have open sourced our new code as "RHQ" along with new "Jopr" plugins - both of which form the basis for our commercial version of JBoss Operations Network 2.x. RHQ, Jopr and JON 2.x are not Hyperic HQ/Spring Source. They are very different code bases and I would not characterize what we have now as a "hyperic fork" or "hyperic is in RHQ".
There is still some legacy code in the UI layer (e.g. the graphs on the monitor page and some other jsp pages are still being used from the old fork). But even the UI is mostly new built-from-the-ground-up code utilizing JSF, Seam and RichFaces technology (things the old code did not use).
Another example of the difference is the agent which is completely different and looks nothing like the old agent - in terms of its implementation code, configuration, user interface, plugin API, everything.
One thing that the old code and the new RHQ/Jopr code use that is the same is SIGAR - which is a very nice native library used to monitor platforms via a cross-platform Java API. But our new integration with SIGAR is, again, very different (for example, we can now run an agent on any platform that supports Java - even if there is no native SIGAR support - unlike the old agent which did not run on a platform unless it had a SIGAR native library ported to it).