Absolutely. I've bought 30 bitcoins on TradeHill and had my transaction push the price by 35 cents. I didn't place it high enough at first and had to repeat it two more times to fill it! That's obviously an extreme example, but that is the trade-off with a smaller exchange.

The priority scheme stops the attack. The idea is that any attack would have to consist of transactions that legitimate transactions could easily out-priority. So while the attack might, at worst, flood the block chain with junk, it can't stop legitimate transactions in the short term.

If someone attacks the network with low-priority transactions, just make sure your legitimate transactions exceed their prioritity, which they probably will anyway. Even if they don't now, they will eventually as your coins age. An attacker can't have an unlimited supply of old transaction outputs. You can't, in general, use 100 old bitcoins to fund 100 transactions. Once the old coins fund one transaction, they are no longer old.

The current scheme is not perfect and may not even be great, but it does seem to be adequate. And it can evolve, and has evolve, as needed.

It does seem that if a smaller exchange is above the Mt. Gox rate, and the Mt. Gox rate itself is dropping, it's very likely that exchange's rate will move down as well. Similarly, if a smaller exchange is below the Mt. Gox rate, and the Mt. Gox rate is increasing, it's very likely that exchange's rate will move up.

Two things are built into the protocol:

1) A miner has absolute freedom to include no transactions or whatever valid transactions he wishes in the block he mines.

2) A client has absolute freedom to accept and relay whatever valid transactions it wishes.

So if the priority scheme needs to change, it can. If the expected transaction fees need to change, they will (and they have). It won't break the protocol or cause any incompatibilities. Except that people who don't follow the new majority rules may have a hard time getting their transactions accepted.

Right now, there are still quite a few client operators and miners who will relay and include in a block all valid transactions. So pretty much any valid transaction will currently get into the block chain eventually. (Especially if you make a special effort to connect to and relay to these clients/miners.)

There are two things TradeHill doesn't do that I would really appreciate. I'm curious if other people would appreciate these features as well and also if there's some unusual reason they're not easy to do.

1) Orders to sell bitcoins I don't yet hold or buy bitcoins with funds I don't yet have:

Say I want to sell all my bitcoins if the exchange rate hits $13. But I also have a lot of triggered buy orders. So I don't know how many bitcoins I'll have. Right now, I can't place a sell order for more bitcoins than I currently hold. I'd like to be able to place a sell or buy order with an "up to" amount, and it will buy or sell as much as possible up to that amount. This way, if bitcoins are at $10, I can place a buy order at $9.50 and a buy order at $9 and a sell order at $13 to sell any bitcoins I might buy if those buy orders trigger and the price comes back to $10.

Similarly, say bitcoins are currently trading at $10 and I currently hold no bitcoins. I might want to place a buy order at $9 and a sell order at $9.50 to sell any bitcoins I might buy should the price drop to $9 and come back to $9.50.

In reverse, this means I can place sell orders as the market rises and use those funds to fund a buy order that I place now. If the buy order triggers, it will buy with funds I have now plus funds acquired after the order was placed should any sell orders trigger. This would also mean I could place an order to sell bitcoins before I transfer them into my account or buy bitcoins before I transfer dollars into my account and the order would fill as it was funded.

2) Stop loss / take profit orders:

Say I hold 50 bitcoins and bitcoins are currently trading at $7. I can certainly sell them now for $6-$7, but I don't want to sell them. But I don't want to keep holding them to $5. What I'd like to be able to do is place a stop loss / take profit order that will sell my bitcoins if they start trading below $6.

In reverse, if I have $100 and bitcoins are trading at $10, I can certainly buy them for $10-$12 now. But suppose I'd prefer to wait for the market to drop, but if it goes above $12, then I want to buy before it gets too much higher. I'd like to place a triggered order to buy at market rate before the market gets much above $12.

So, are these features doable? Am I the only one that would use them?

Shifting the risk is valuable for the customer as well. If one party to a contract is afraid the other party will default, that makes them more reluctant to enter into the contract. To induce them to enter against the reluctance, the other party must make a greater offer.

Imagine you want bitcoins and you have PayPal. By your logic, the fact that PayPal lets you reverse the transaction, putting all the risk on the other party, should be great for you. But in fact, it's awful for you. Because you can easily cheat the other party, there won't be any other party. You would have to offer someone a massive inducement to take that risk, that is, you would have to pay extra for the extra risk you present.

Because bitcoins remove the chargeback risk, the customer need not pay the merchant to take that risk, saving the customer money.

As a general rule, anything that makes an agreement more beneficial to either party benefits both parties. The party that doesn't benefit directly can make a slightly worse offer and the other party will still benefit from the contract, thus forcing a split of the benefit.

Good, then difficulty will drop and there will be a mass influx of miners.

Why, does something stop people from saying "I agree to compensate you for your losses"? And, in any event, these are not verbal agreements, these are written offers and then acceptance through an action.

That's certainly true. But I'm not sure what that has to do with trust and contracts.

So you're suggesting that they could kill bitcoin by making everyone involved in it rich?

In the deflation case, you borrow the right to 1.05 rc next year, and can invest/trade/whatever precisely that amount of value. You pay back 1.05 rc next year plus interest. You pay back precisely the same real value you borrowed, plus interest. It is no harder or easier.

You can only have this kind of deflation in a world where 1 rc today is worth very slightly more than 1.05 rc next year. So you are borrowing very slightly more than you are paying back. The difference is precisely the opportunity value of having the money sooner.

That's not correct. You are forgetting that those investments would have a higher return as well because a deflationary currency changes the time value of money.

That's correct. So the assumption that 1 btc is worth 1 rc now but will be worth 1.05 rc in a year changes the time value of money dramatically -- it means that 1 rc today must be worth more than 1.05 rc next year.

They already will have. Otherwise, that condition wouldn't exist. You can't change one assumption while keeping all the others the same, the world doesn't work that way.

You got a reductio ad absurdum because you assumed both inflation and deflation at the same time.

Deflation simply means that more of the present value of money comes from its expected future value. But it is still worth precisely what it is worth, and that includes the present values of its future value. A borrower is borrowing the present values of those greater future values and gets to spend them, providing the notional surplus that compensates for the deflation.

Say I borrow 1 bitcoin for a year. I pay back 1 bitcoin plus some interest. You're thinking "I only borrowed 1 rc and I have to pay back 1.05 rc and that's before the interest -- unfair". But it is just as valid to think "I borrowed a bit more than the right to have 1.05 rc in a year, and I pay back precisely 1.05 rc in a year, before interest -- more than fair". I am paying back almost precisely the same thing I borrowed, plus interest of course.

And you'll note how perfectly balanced it is. The thing I'm borrowing more than I'm paying back, before interest, is the extra opportunity value of having that 1->1.05 rc for a year. And that is always precisely what the lender contributes and the thing for which he is paid interest.

All of them as a set conflict. Which one you change so that they don't conflict is up to you.

For example, you cannot assume that I can get a 10% rate of return by doing nothing but would be willing to settle for a 5% rate of return on an investment. Those are conflicting assumptions. It is very unlikely you will ever see both of those conditions in the same world and the same time, and if you do, obviously very strange things will happen.

If you wish to assume that 1 bitcoin will be worth 5% more real value in a year than it is worth now, you cannot also assume that the present value of that greater future value is more than 1 bitcoin now. (How can something have a net present value of more than 1 bitcoin if anyone who wants to can buy it now for 1 bitcoin? That makes no sense.) Otherwise, you are assuming an economy that is concurrently experience both predictable long-term inflation and predictable long-term deflation.

So I should, say, buy 100 bitcoins for $1,200 and then sell them for $900 -- less than I bought them for?

If only that chart were to scale and had numbers on it.

You've calculated the consequences of a set of conflicting assumptions. That's kind of like figuring out what happens if both "X=3" and "x>5" are both true. Even if you do the calculations correctly, it's hard to figure out what your results apply to.

Correct. Since 1 bitcoin includes the right to have 1.05 rc next year, the present value of one bitcoin must be at least as much as the present value of 1.05 rc next year (in any comparable units).

If your assumptions lead to the present value of 1.05 rc being less than 1 bitcoin, then you have conflicting assumptions. Something that includes something else and then some cannot be worth less than that something else.

Right, so you're doing it wrong. If 1 bitcoin will be worth 1.05 rc next year, then the present value of 1.05 rc next year cannot possibly exceed 1 bitcoin. So your assumptions conflict. (Most likely because you aren't being clear about whether your inflation percentage is in real value or in bitcoins.)

As a general rule, you want to avoid the case where two parties have the ability to spend the same coins unless one is just acting as an emergency backup to prevent loss of the coins. There are just too many possible race conditions.

The problem is that the network doesn't think of coins as belonging to an account. So even if you and I both have access to the same account's 100 coins, if you and I both go to spend 5 bitcoins each at around the same time, we will likely conflict.

You already have a contract. If someone makes an offer and you accept it by providing them something of value, you have a contract. That's how the vast majority of contracts are formed:
http://en.wikipedia.org/wiki/Offer_and_acceptance

See, for example, Carlill v. Carbolic Smoke Ball Company:
http://en.wikipedia.org/wiki/Carlill_v_Carbolic_Smoke_Ball_Co

You haven't fixed either error:

1) The 1.05 is the future value of the future value, you're supposed to subtract the present value of the future value. Even if 1 bitcoin today will be 1.05 units of value next year, 1.05 units of value next year isn't worth 1.05 units of value  now.

2) If a person can turn 1 bitcoin into 1.05 units of value next year, then the present value of 1.05 units of value cannot possibly exceed 1 bitcoin.