Often in those situations, cigarettes come into the economy as a ration (or a purchase limit) that people get whether or not they are smokers, and the economy revolves around getting this commodity that everyone gets to the minority that need it. Pretty quickly, you end up with non-smokers trading cigarettes with each other, because you can always get whatever you want through paying a smoker in cigarettes.
Bitcoin could benefit from that if it encouraged a very large drug market, because people need their drugs. If an easy, comfortable bitcoin drug market could get a lot of volume, it would become a lot easier to buy a cup of coffee or a car with it, and the prices would rise and fall based on the relationship of the ease of purchasing drugs with bitcoin vs. the ease of purchasing drugs in dollars. Maybe money laundering could keep bitcoin stable, too. In any case, to be successful, IMO bitcoin will have to be driven by something other than belief in bitcoin.
The use of the USD? It can be used to pay US taxes.
You could say a commodity is like a currency many people can produce.
Considering how commodities are often produced, the term "Bitcoin mining" is shockingly apt.
See http://www.npr.org/blogs/money/2011/08/19/139791374/the-frid... and http://www.npr.org/blogs/money/2011/09/06/140211340/swiss-to...
cube13's characterization is a simplification, but I think it still captures their intended uses. Bitcoin's deflation, and the Swiss franc's high value compared to other currencies, harm their ability to act as they were intended to. Right now, I agree with Duffy's original point, which is that bitcoin is more of a commodity than a currency.
A square is a rectangle, but not all rectangles are squares.
Scott is on the right path -- there's some sort of threshold where a substance is more valued as a transaction medium than for its intrinsic value.
I personally believe that it will never achieve anything other than niche status as a currency for trading illegal goods; it will be used for other types of transactions, but it will be a very small number of transactions when compared to contemporary financial transaction systems like, e.g. credit cards, or paypal. I say this only because in practical terms it is not very easy to use, and furthermore, its benefit of being pseudoanonymous is also is downfall; very safe for the recipient to receive bitcoins but very risky for the sender to send them since the sender has absolutely no insurance or any recourse in the transaction. There is no authority to appeal to for transactions gone bad, and even if you could sue over bitcoin transactions and the courts would understand them, the pseudoanonymity would make it very difficult to know who to sue.
- Easy of use: this is mostly a matter of software, and I see no reason to believe it has to be difficult to use. With current proposals such as specific tags for websites and such, sending money can be made as easy as inserting someone's email address, the amount and hit "send."
- Unsafe for the sender: nothing in the design of bitcoin prevents the creation of escrow services. It's simply not an issue to be fixed by the currency but by higher level services, just like with other currencies.
- Pseudoanonymity: it's mostly a red herring. When I'm buying from a random Chinese 'shop' through Ebay I won't be able to sue him either. When I'm buying from a reputed shop in my country, I don't need to ask its bank to know who I need to sue.