We still have the primary problem: If it hits a town (due to any error, such as an unforeseen destruction of a component) the human casualty would be a disaster. It's not even just the $$.
A barge would not work because it is mobile as well.
That's why they have range safety. http://en.m.wikipedia.org/wiki/Range_safety.
Basically a remote self-destruct trigger with multiple redundant data channels and possibly a self destruct on communication loss.
You will get some debris but the damage is lot more predictable than a full stage coming down.
Isn't the issue that 'range safety' isn't really a landing feature? Look at the shuttle strewn out over texas. Range safety is fine for launch (when you have control over initial trajectory and ___location); or for a test-flight that is retsricted to and actual missle range (or other restricted airspace).
The Shuttle had to reenter over populated areas (and didn't hit anyone, anyways). The SpaceX first stages will be returning from over water, and targeting a beachfront pad, so if it's coming towards land too fast for comfort you can just drop the debris into the water.
The way Columbia disintegrated somewhat imitates what range safety would do anyway. The whole idea is to blow the thing into little pieces which are collectively relatively harmless. You'll note that nobody on the ground was hurt by Columbia's debris.
Beyond that, the idea with range safety is to blow it up before it gets to a populated area. Not only to you break the rocket into little pieces, but you do it so that they fall down into the ocean, or on empty ground. Columbia's landing profile didn't allow this (even if it had range safety, which it didn't) but F9-R's certainly ought to, since it's returning over the same empty ocean that it launched over.
The space shuttle has range safety, but only during launch. The shuttle launches over the ocean, which allows for 'range safety' to be implemented, on the SRBs and EFT.
Since the shuttle's re-entry flightpath is not constrained to restricted airspace, destroying the shuttle orbiter deliberately would only be "range safety" in the loosest sense.
Empirical data seem to suggest Columbia began to self-destruct over the CA/AZ border before landing in a debris field scattered over (populated) west Texas. As seen from an actual destructive episode, the debris path/cone was extensive, and unlikely to be constrained to air-space designated for such purposes.
The issues about how to mitigate this problem comes down to some of the things noted in at least one other comment here--basically SpaceX would limit the re-entry flight-path to open-ocean during approach.
Even then, the question is about ballistic "backstop"/shadow of its landing site ("the beach").
These are pretty basic questions about the size and nature of the landing facility and the ballistic match of un-guided (and potentially un-aerodynamic) debris based upon whatever the realistic assumptions of velocity/altitude and response time of the system are.
The SRBs and external fuel tanks had range safety, but as I said, the Shuttle Orbiter did not. Any "self-destructing" that Columbia did was due to extreme forces induced by the atmosphere.
The willingness to fly a manned reentering Shuttle Orbiter over populated land without range safety should not be misconstrued as the willingness to drop an unmanned first stage down with engine power without range safety.
They use range safety for launches at those same locations with planned flight paths over the same ocean. I see little reason for them to not use range safety for landings.
The orbiter (re-entry vehicle) is well know to lack range safety, unlike the full space shuttle (launch system). That is clear from my earlier comments. That is not a point on which anyone is dis-agreeing.
With that out of the way, lets look at the issue at hand.
Two issues comprise range-safety and both are at play: (1) is the egineering; and (2) is the flight paths. Proper range safety requires both (1) and (2) combined. The orbiter-as-re-entry vehicle lacked both (1) and (2). Whist the shuttle launch system had them both.
Certainly space X could engineer (1) and (2) using similar techniques at launch with no issue.
The open question is simply providing for range safety for the re-entry/recovery portion of the flight. The shuttle providese little to no road-map in that regards.
Just the opposite: it illustrates some of the difficulties.[0]
Assuming an engineered solution is present (ie, pt1 above) what would the limitations on the flight-patch (ie, pt 2 above) need to be in order that the comination (1,2) together would qualify as "range safety" in the legitimate sense.[1]
The ~rough~ answer seems to be (2') needs to be kept over water/open ocean.
So, my question is more about ballistics math: what is the envelope of precision needed to keep something either (a) in the ocean; and/or (b) out of harms way if the event is triggered closer to land.
The answer to that is something the engineers at SpaceX have surely considered.
I don't know what those calculations show; it (surely) can be safely done up to some threshold.
The question then simply is "what is the threshold"?
_____________________
[0] The scale of the debris cone from a columbia's ("natually occurring") event @ 100,000+ ft is illustrative of a couple things. None of: China Lake, White Sands, and Barry Goldwater etc alone could ~readily contain such an event. We know this because all were proximate to the debbris path (ie, western CA, Southern AZ, South/Central NM).
Some outside observers' current best guess about mitigating this danger is to have the returning boost stage initially aim for water just off-shore, which leaves time to destroy it in the air if it winds up off-course. If things look good, they can then use the landing burn to divert from just off-shore to just on-shore.
FWIW, they haven't announced anything specific about range safety management, but the "Grasshopper" landing test rig was seen rehearsing such a horizontal diversion on one of its last flights.
(As to the barge, its being mobile is a feature, not a bug: they could tow it to whatever spot at sea needs the least fuel to get to after stage separation. The problems are that the barge itself may not stay level during landing, and that the rocket exhaust from even a single throttled-down engine is still likely to burn a hole in a barge.)
The empty weight of a Falcon 9 first stage is somewhere in the neighborhood of 14 tons. It would be vaguely comparable to a DC-9 crash with little fuel left in the tanks. Not quite as bad, as it's lighter (DC-9 is about 22 tons empty) and considerably less dense. So it would be a "take out a couple of houses" event if it hit the wrong spot, I think, but nothing much worse than that.
A barge would not work because it is mobile as well.
This is indeed rocket science.