The actual problem is one of power and intention. Yes, version control would make earmarks obvious. Yes, it would make tracking contributors (lobbies) easier. Yes, it would make tailoring tried and trusted legal documents easy.
All of these are reasons why version control will never be applied to the law. We want these things, but we are not the customer. The real customer is actively trying to prevent these things from ever happening.
The reality is that these tools are probably already being used for these exact purposes - in private, and for personal gain.
A separate problem, however, is that this is essentially a process improvement, that doesn't generate much advantage to the person who makes it happen. A politician would need to use significant personal resources to push this type of reform through the system, yet the 'payoff' is probably minimal compared to substantive policy or constituent service activities. By raising the issue from process improvement to substantive issue, the chances of implementation improve. Discussing and spreading the idea is valuable in that light.
On a side note, once these types of distributed systems begin to take root, it becomes quickly apparent that co-location in Washington is of at least questionable value. (Look at the attendance records of members at Congressional hearings.) Virtually everything that Congress does can be accomplished without the need for the massive physical and personnel support system found at the Capitol. It would be interesting to see how government would be designed given the communication tools available today.