3. To "globalize" an asset, for each block a proof-of-stake calculation is run for addresses controlled by a running instance of counterpartyd (against some random function derived from BTC block hash, for instance). A "globalize" command can be called by the user to denote the asset to "submit" to the PoS function. (I'm no good at the maths, so looking to existing PoS implementations to figure out how to do this). If no asset is selected to globalize, the client automatically chooses the "first" one (in lexicographical order, for instance).
The whole idea is good, though this might imho cause that an asset lower in priority, will be never globalized. We need to think of some way to compensate for that.
As I mentioned, allowing users to "pool" their "asset-globalizing" power to support projects that otherwise wouldn't hit the markets. Not in the NXT sense of "send all your money to me so I can mine PoS blocks", but something that better fits the "trust, but be trustless" model of XCP. And without requiring a centralized pool.
Suggestion: now would be an opportune time to get that 'dev' branch of the github set up to test ideas such as these. Alterations to assets are definitely not ready for prime-time yet.