the seller would participate in the swap/repo if R_b>=1+r_b and r_r<=R_r-1, only if his gain on one leg is greater than the loss on the other leg. No need to gain on both legs.

In fact, suppose the real interests rates of Repocoin and Bitcoin were r_r and r_b. Currently, Alice holds B bitcoins and Bob holds B*P repocoins (P is the current price of repocoin in one unit bitcoin). Alice and Bob make a one week swap deal, and will get B * r_b bitcoins and B * P * r_r repocoins after the deal respectively. In a equilibrium market, the new price of repocoin one week later will be P * (r_r / r_b).

However, in my opinion, the real interests rates of both coins can't be independently revealed by the market. Even a well functional market can only reveal the relation between r_r and r_b, more specifically, r_r / r_b. When I am putting or accepting an order with 3 quantities: B, r_r and r_b, another order with 0.5B, 2r_r and 2r_b is almost same to me. The only difference is the latter has a leverage of 2 times.

Regarding the modified proposal, I believe it is not a repo deal any more, it's more like a fx swap deal. As a result, Rr is not real repo rate of repocoin, but a function of the rates of both coins. If so, how could we apply these numbers to adjust repocoin's PoS interests rates?

Thank you for your excellent work.
Seems examples can't run properly in python2.7, but OK in py3.3.

Thank you so much for your kind reply. It's quite clear and I get it now.
BTW: I wonder is there any way to unlock a wallet?

I copied my wallet to another computer, then generate a new address, and received some coins with the new address, the balance added. But I can't see any change in my original computer and balance. Why? Doesn't the wallet recognise the new address generated by its copy?
Thank you for any help.

It's wired since PPC is much newer, and its blocks are much less than BTC.