Actually, I feel PCGen has done a very good job of dealing with multiple sources and their conflicts as far as identifying which are interdependent and which are exclusive; if you spend any time manipulating sources in the Advanced Sources list you very quickly see just how much effort went into that.
To the extent that file duplication is inherent in the sources, and the very real caution with which AM and his fellow designers should approach the alteration of source material provided by the Game Designer itself---say Pathfinder---your point is valid, and clearly the PCGen team made a real effort to be consistent with the Game Designer's vision while permitting the GM to be as crazy as we like to be.
Where your agrument breaks down---and I cannot speak at all to AM's comment about a quantitative view of "issues", as he is the expert there, clearly---is within the many data files themselves, within the sources loaded by the game's user interface.
If you've spent any times inside them---and I have spent a lot of time there, just so you are fully warned before you go off again---you find that very little consistency exists across the numerous files and filesets. To speak to AM's point about the long history of PCGen and its nature as an open-source project, it is clear that the more recent efforts **have** made attempts to exert standards and practices, and to ensure as much commonality as possible. One of the issues seems to be that successive file "Tsars" want to go in different directions, so that "the wheel" keeps getting reinvented; another is the scale of the files themselves, which makes it close to impossible to make wholesale changes in a single iteration of the PCGen engine--this second issue results in several "started-and-stopped" efforts at clean-up.
If you had spent any real time in the files, you'd have known those thing, and have known about the efforts to fix them, and recognised that it was a real problem. As you are clearly one of the "privledged few"---I see your name on the output files when I'm repairing them---one might have expected you to answer my comment with the same calm that AM did, but add some salient facts; like **why** there are so many starts and stops, and **why** the project seems unable to settle upon a standard for data files; after all, the choices of Java brought with it the absolute requirement for complexly-formatted data files.
But none of that answers some of the low-level "whys". For example, why is it that any given file might require two or three entries for the same item, in different lists? Why is it that a single data point---say, for example, a Weapon Bonus---has to be repeated two or three times in a single data entry? All of those redundancies create opportunities for error and anomaly, and make altering the data files, even for experienced editors, an onerous and error-prone task.
Good version control doesn't just protect a file from being edited by more than one person at a time, nor track progressive edits. It controls the process by which a project advances the technology employed and the technology being attempted---so that if a new technology should prove to trouble-probe, or unusuable, it can be "backed out" without the whole project collapsing. The data files supporting PCGen---and without which the programme is useless---are not well managed, however you want to split hairs over how you define the work of version control, and whatever the reasons.
Certainly the deluge of supplements and "campaigns" and etc... coming out of Gaming Companies is a huge burden, and quite probably beyond the resources of an all-volunteer effort; much less going back and enforcing one of the several "formats" envisioned for the data sets and the way they are loaded into the java engine. I do see that. Better than you, clearly. But that there is an explanation does not mean there is no problem.
