As I see it, the tricky bit without an exchange is (1) establishing a "market price" in the first place, and (2) coordinating the fiat side of the transaction with the cryptocurrency side (the same problem exists between two different cryptocurrencies with different networks) so that the two either succeed or fail together.
There is also localbitcoin
some people used mtgox to trade, but I can not understand why since their fees were high, trading engine slow and buggy and well there are better alternatives
This is all so clear now, it seems. Further, Bitstamp was bitten by exactly the same bug (the malleability bug which was apparently what demonstrated Gox's purported cluelessness, and which I highly doubt was the cause of MtGox's apparent failure).
Unless there was an independent audit by a third party, why should you trust that they won't fall victim to exactly the same issue?
This is not factually correct. Bitstamp and MTGox's issues were unrelated, and the transactions that caused problems for Bitstamp were never seen on the network prior to a couple hours before MTGox's press release (and in volume, until after it).
In both cases there was an issue with the service having transaction issues due to the malleability issue. Bitstamp suspended operations to fix the issue. How can you possibly say they are unrelated?
Now supposedly Gox was robbed (over months) using this exploit, but that simply makes no sense at all, and is people grasping, in the absence of anything else. It seems more likely that someone got the private keys on holdings wallets, whether insider or hacker.
Because they were unrelated. Bitstamp was spending its own unconfirmed change, and those spends got hung when the change was mutated by the attacker— with an attack that didn't exist while MTGox was processing withdraws. MTGox happened to never spend any coins that weren't at least 6 deep in the blockchain, even it's own change.
> and is people grasping, in the absence of anything else
It's actually MTGox's own claims. Their actual issues in this space were because they reissued payments without conflicting the original payment— allowing both to go through. No one else, as far as I can find, was performing reissues at all— much less in such an unsafe manner.
Where did they claim this? They did claim that they were suspending withdrawals due to that issue, but in actuality they had not done any withdrawals for months. It was a convenient cover at a convenient time, which they tried to extend out with the "until this problem is fixed in the protocol" nonsense.
They certainly had made withdraws within months. Here is a random address receiving payments from MTGox (the ones with the 0.001 fees) https://blockchain.info/address/1AacEsKeXqnUqtYQWKDHJV3JtpJk... if the API were still up I'd show you how to query the vin scrippubkeys to have mtgox themselves identify the txn as theirs— but it's not up now.
Only ways to get bitcoin:
1. Mine them (very very hard and getting harder)
2. Get paid in bitcoin for product/service
3. Buy them on an exchange for fiat
4. Be given bitcoin by early adopters
Until its possible to get paid (salaries), pay suppliers, pay taxes etc the exchanges are needed to convert traditional currency to bitcoin and vice versa.
Anyways the writing was on the wall for a long time, i stopped using them over a year ago, bitstamp are professional while mtgox were amateur cowboys with no support.