Ken,

Whether it should be is another story because Joint Programs with wings seem to have a tendency to come under USAF control -- to the detriment of guys on the ground
Well, I disagree. The purpose is to ensure commonality so the guys on the ground don't have to carry one set of gear to link with Army assets and another set of gear to link with Navy and a third to link with Air Force. Of course, cost containment is also a reason. I also reject the assertion that USAF "control" automatically results in the detriment to the guys on the ground. I see similar statements all the time - statements that are apparently to be taken as objective fact but are really debatable.

And speaking of "control" we need to establish what that means. Control can be any or all of the following:

1. Control over requirements, development and acquisition
2. Administrative control
3. Operational control

Predator has always been intended as a Joint Force commander asset, not exclusively an air component asset. Predator utilization in theater is not up to the "Air Force" - the priorities are set by the joint force commander and, as is the case with all air assets that operate above a certain altitude block, coordinated through the CFACC.

but I also strongly believe the Army should own and deploy Warrior as it sees fit and you did not add that caveat.
Using our current doctrine, I'm assuming here that by "Army" you really mean the "land component commander" (LCC). I have no problem with that but the CFACC must still be the agency for deconfliction and coordination. For all the same reasons the LCC needs a measure of unity of command over all ground forces be they Army, Marine or Coalition, the air component needs the same over air assets.

Having experience with close air support with and without an overarching 'joint' air effort, I am in no doubt that the CAS is better without total consolidation and a Theater ATO. No doubt what so ever...
I wonder how long its been since you've had first-hand experience with CAS? I'm guessing not in the past several years - a lot has changed.

True statement -- it also, as does your opinion, neatlyelides the fact that all those ROVERs may not need a video of an area 200 km away while the boys in blue search for 'strategic' targets; that is no guarantee of coverage when and where needed, merely access to what ever the ACC or JFACC deems important to them at the time. Which may or may not be the pressing concern of the Theater Commander, much less the Ground Component Commander...
Well, rover (the newer versions at least) can also receive the imagery from a manned fighter's targeting pod - indeed from almost any AF platform with a sensor. Can the same be said for Apache or Kiowa's? Again, this is the kind of integration the AF is way ahead of the Army on and is a big area FCS is supposed to address.

And again, the "boys in blue" (by which I assume you mean the CFACC) are not out there executing their own missions willy-nilly, but implementing the joint force commander's intent and the JFC wants coverage of strategic targets too. If the CFACC is not providing adequate support the the CFLCC, then the CFLCC can take the issue up with the JFC who is the decider. The idea the CFACC can or is executing missions contrary to the JFC's wishes, or is going off and doing its own thing to the detriment of the land component is ludicrious, especially if you're talking about predator which is a JFC-owned asset. The vast majority of predators and, indeed, all tactical aircraft, are used to support the ground forces, but the JFC also uses predators for strategic reconnaissance and the Joint special ops commander has some too for their specific missions.

No question that centralization increases efficiency -- there is also little doubt that it adversely impacts effectiveness.
Not in all or even most cases. Did Goldwater-Nichols adversely impact effectiveness? Centralization often increases effectiveness. Again, what is more effective, a UAV C2 architecture that's common over the whole force, or one where each service develops its own incompatible system? A common standard architecture means better integration across the whole force.

And this doesn't mean that each platform has to be the same. If the Army wants a UAV with greater endurance, or different weapons options, or whatever, that's great. The problems and ineffectiveness creep in when the services want to use proprietary datalink, C2, etc. In other words, centralization is not an either-or concept.