I think 1MB is good. Keeps the cup-of-coffee-buyers and third-world-small-bag-of-rice-purchaser off the blockchain.

Go to the Digital Goods section and buy an MSDN admin account. Activate your 3 year free $150/month Azure credits. That should be enough to pay for a couple of 1-core virtual servers to host 2 full nodes. (Make sure Vod does not catch you doing it, though.)

Good thread indeed. I've always been more worried about potential weaknesses in ECC rather than somebody's ability to "decrypt" SHA256.

You are probably not a geek, but I understood what was said. There were obtuse references to geeky stuff that many on the forum would get.

Please post an example. Maybe their off-topic comment were entertaining enough to others.

Is anyone currently NOT having problems?

I guess the assumption that everyone is a Compounding Moron^TM is false. I would not be suprised if many depositors have taken into consideration the risk of default/ponzi and already have in place some strategy of risk reduction based on systematic withdrawals.

Some still say Oanda is a bucketshop, albeit a respectable one.

Rolling back trades is not the issue I have with Mt Gox. Just the sheer size of the rollback and the fact that they got themselves into a state where this is the only best solution.

I am not a trader, I develop trading systems software. I've worked on systems that cleaned up the mess when trades get busted, so my attitude is less cavalier than yours. Admittedly, I've dealt only with institutional and larger hedge fund trades that don't typically involve lightly traded securities, so these massive rollbacks could be a more common occurence than I thought.

That many people here are extremely naive ...

1) thinking that rolling back trades will somehow "make things right",
2) assuming that all BUY trades involved only people buying the stolen coins, and therefore it is rightious to reverse all BUY trades,
3) assuming that Mt. Gox is a closed system and that nobody did any offsetting trades on Trade Hill or even offline as a result of the trades that cleared in Mt Gox (e.g. arbitrage)
4) what about people who sold into the panic at $10 but after reversal the market re-opens at $5? Was it fair to reverse their trades?

Those of us who have worked on real exchanges know that no matter how you do a rollback some innocent party will get screwed in the end.  Usually it is the small guys who get screwed, never the GS, JPMs or the Smith Barneys. Looks like the same thing will happen here. The 500K BTC guy would get most of his coins back but the small timers who traded on good faith during the plunge won't.

Remember: not every trade involved some lucky dweeb buying "stolen" coins at 0.1.

Hopefully few innocent people will be affected this time round.

If Mt. Gox wants to be taken seriously this should be the first and last time they do this kind of bullshit.

I don't think it matters either way. Any properly written parser would use explicit conversion and not rely on conversion to the default types provided by JSON specs.

A more extreme and simpler analogy ... If I get receive change from a store which later turns out to be stolen currency I shouln't have to give it back.

Further more, it is not clear yet that all sold Bitcoins were stolen. Some people cold have panicked and sold  into the plunge with "legit" coins.

Probably by running a basic brute force attack using dictionaries and rainbow tables.

I once worked at a place and we had a security officer who periodically attacked our network passwords and forced us to change those he managed to break. It was pretty annoying of him.

Hmmm ... if I had been playing market maker and during the selloff bought BTC at $5 in my Mt Gox account and sold BTC for $6 in my TradeHill account (hoping to pocket the $1 difference), is Mt Gox also going to roll back my Trade Hill trade? Or am I supposed to just suck in my losses because one leg of my trade happened to be fradulent?

The above situation was hypothetical, at least for my case. I am just pointing out that by rolling back trades there are unintended consequences because people might have placed other trades elsewhere on good faith that their Mt. Gox trades had cleared and will be honored.

Can someone try to crack user 16139 please?

I would like to know how strong the password is. I believe it is pretty strong but I could be wrong.

They tried that with some commodities futures exchanges. Remember in the 70's silver locking limit up when the Hunt brotheres tried to corner the market? The price protection rule could not keep up with the market and the first trade of the the day would lock the limit and nothing else got traded for the entire day. This repeated day after day until the price reached market value. Only then did the market unfreeze.

Sounds like a good approach for a hoarder, but what about someone who frequently needs to spend BTC?

If everything were in a state of equilibrium, then what the OP stated would be true. As things currently stand, the Bitcoin economy is expanding and things are certainly not in equilibrium.

It may still be cheaper to produce a Bitcoin vs. mining, but if I had an immediate need for 50 BTC to buy something from Silk Road I would certainly not attempt to order parts from Newegg, build the miners and then wait several weeks for my 50 BTC to trickle in. Instead, I would first try to buy it locally (possibly from Craigslist), failing that I would go the Mt. Gox which would take 1 week max.

Yes, there still is some relation between the cost of mining 1 BTC and the price of 1 BTC on the open market, but it is just one of the many factors that determine the open market price. While there is a tendency for both prices to converge over the long term, in the near term you cannot really say what will happen.

There lies the biggest flaw of open source.

The first ASICs would probably be GPUs modeled after the 5970 with the unneeded modules stripped out and extra SPs packed into the freed up area. Also, going down to a 20 nm process would allow for more transistors and less power. I could see a small US-based startup company with 2-3 million USD funding, 2 chip designers, 1 board designer and a couple of lab techs. coming up with this optimized GPU/ASIC in about 9 months. You basically need to figure out how to do all the brain work yourself and outsource the fab and packaging as efficiently/cheaply as you can.

An end product with 4 times the hashing rate of a 5970 and consuming 25% the power is very conceivable. That would just be the first cut.

The barrier to entry is not so high so you would probably see more than one company do this.