Are you aware of the "Ignore Boards" feature?

When you visit another person's house, make an ass of yourself, and they tell you to knock it off - that's not censorship.

If somebody comes into your house and tells you what you can and can't say - that is censorship.

Learn the difference.

There's already precident for this: 1/1000 of a dollar is a mill.

Since people tend to gravitate towards shorter words instead of longer words my guess is "mill" and "mike" would be more commonly used than their three syllable equivalents.

In the long term, learning how things work does result in a no-hassles, easy to use system.

The best learning experience is to install Linux From Scratch.

Then once you've gotten the hang of LFS and have a decent understanding of how all the parts work together switch to something a little more automated like Gentoo.

That is exactly what mixing should focus on.

I can take 73.26 and split it myself in to smaller sizes in a way that looks like a series of purchases on my own node with no cooperation needed from anyone else. What I can't do is anonymously combine all my dust addresses into an address large enough to be useful without outside assistance. A client can be very careful to use different incoming addresses for every receipt and to never link addresses but at some point the user is going to want to spend an amount larger than the balance of any single address. The only way to avoid this situation without compromizing anonyminity is to have the ability to securely combine small addresses into larger ones.

That's why I think mixing should be focused on transactions which have many more inputs than outputs. https://bitcointalk.org/index.php?topic=93390.msg1036811

I have a policy of not applauding the efforts of companies which enjoy state subsidies, including barriers to entry, unless their efforts are lobbying for the removal of those subsidies.

You don't need an external mixing service to anonymously break large bitcoin balances into small bitcoin balances - you need one to combine small balances into larger balances.

So a better use case would be several people sending their small balances to a mixing address to obtain larger balances. If everybody takes out bitcoins in increments of BTC1, for example then it's even harder to figure out which inputs map to a particular output because all outputs are the same.

Why are random output values better than identical output values?

A planned economy controlled by the intellectual elite didn't work in the USSR or anywhere else it has ever been tried.

The ideas in the Venus Project are taken nearly word-for-word from the 1917 Russian Revolution propaganda but with computers this time, as if that's going to make a difference.

I wish more companies would do this.

These are some thoughts I had while considering the problem of anonymous coin mixing. Not all of the features exist currently in bitcoin but I believe they are possible:

N agents meet over an anonymous communication channel. They collaborate to create an N-of-N multisig address (M). They also agree on a time limit (L) in the form of a future block number and a payout amount (A).

Once those preconditions have been met the begin constructing a transaction (T) which takes M as an input and several outputs of size A. They nominate output addresses in a manner which can not be traced to any particular participant.

When all N agents are satisfied with T they sign the transaction but wait to broadcast it until M is sufficiently funded.

Now it's up to all participants to fund M before the time limit expires. In order to prevent cheating, the transactions they use to fund M expire if they are not included in a block prior to L and are invalid unless they are included in the same block as T. Each participant should include additional input beyond the outputs they have specified as a contribution to the transaction fee. They can attempt to contribute less than their share but that makes it unlikely that T will be processed.

If M is not sufficiently funded prior to L then mixing fails. T never happens and everybody gets their coins back because all the transactions which have M as an output are invalid.

What's the point of mixing if it isn't secure against that scenario?

If you don't like including protection from governments in your threat model then just replace court orders with sufficiently-motivated elements of organized crime, or malicious operators.

Mixing isn't the problem. If Instawallet keeps records of its users it can map the mixed outputs to the original inputs and provide that information to third parties. Accessing Instawallet through Tor doesn't help with that.

It doesn't matter how trustworthy the operators of Instawallet are, they can be forced to keep those records against their will. https://en.wikipedia.org/wiki/Hushmail

So the only way to have secure mixing is if it's not possible for even the operators of the mixing service to determine which outputs correspond with the original inputs. I don't know if it's possible to do that or not; if I knew any cryptographers I'd ask them.

The way in which the operator of TORwallet has partially compromised his identity confirms for me this entire approach is wrong; it's a mistake to trust a third party to protect your anonymity.  Even when the owners of a mixing service have the best intentions they can make mistakes and if they can be located they can be threatened.

I don't know enough about cryptography to say how (or if) it can be accomplished but mixing coins needs to be done in a way in such a way that it's not necessary to trust the mixing service.

Crop insurance companies didn't start up this year. If they've been doing proper risk management they'll survive one bad year.

I agree with that.

It would be a mistake to assume insurance is impossible in principle merely because you don't know how to do it.

The form of the equation you should be looking at is P=M(V/Q), and I'm not assuming V=8760.

M is set by protocol so the factors to consider are P, V and Q. P is the price of goods in BTC. People who want the value of bitcoins to increase want a lower P.

If a large business started accepting payments in bitcoin Q would certainly increase. If immediately spending, or converting the bitcoins to another currency, tends to increase velocity depending on where it was to start with. Holding on to the bitcoins tends to reduce velocity.

Which effect dominates depends on the relative change of both P and Q. If Megaupload increased V more than it increased Q the exchange rate (USD/BTC) would go down.

Right now my best guess is that velocity today is such that anything Kim Dotcom is likely to do would increase the exchange rate, but that's not the same as a guarantee, and certainly it's not true to say any significant increase of bitcoin as a payment method will make the price go up.

For their sake I hope your customers know not to hire you if the value of what they want you to protect is more than BTC10K.