Sorry for the delay, here is the update on the current situation:
Why is trading still paused?
A new team member has discovered some bugs in the trading engine. They need to be fixed to avoid exploitation and maintain integrity.
We are re-writing the whole engine as we speak. Among other things, the new engine will propely indicate buys and sells in the market history.
Why is the site going down so often today?
After moving large amounts of altcoins to cold (secure) storage, our wallets were left with a large amount of small unspent inputs.
These inputs are generated by people who mine directly to our deposit addresses, using small automatic payments from pools.
When a withdrawal fails because too many unspent inputs must be combined, the coindaemon throws an error: transaction too large.
What are we doing to prevent this? We will manually consolidate the inputs by sending coins to ourselves.
This can take some hours, but its not a major issue. After this, the coindaemon will be able to choose less inputs for withdrawals.
Our system is configured to stop on any unexpected condition, at all times.
This is why the server is going down - we have to manually restart it after a coindaemon error.
This is the same protection that has prevented people from exploiting bugs - its important to "crash" on any unexpected condition.
We carefully inspect all error messages before restarting the Node.js process. Having this protection will make us alot better in the long run.
Also, because of the Mt.Gox situation, withdraw volumes have increased on all exchanges.
Because of this, we have to re-fill our hot wallets, multiple times per day. Usually its the other way around.
Another effect of increasing withdrawal volumes is that the coindaemons run out of "large" unspent inputs and start combining the dust from mining deposits.
What happens to stuck deposits, withdrawals and other support requests from the last pages on this thread?
Less than 0.5% of users have a stuck transaction. Mostly because of e-mail delivery issues (working on them aswell).
All recent support requests will be handled in 4-24 hours, starting now (already handling some).
If your confirmed withdrawal failed without an txid, dont worry. We will manually resend these later.
We are hiring more people - another Node.js/crypto developer and 3 additional support team members from different timezones around the world.
If you happen to be (or know) an expert in Node.js, feel free to contact us.
We will start moving our infrastructure to a production-ready datacentre as soon as the dust settles.
Websockets-based realtime UI is ready and will be deployed next week.
While the markets are paused, everybody can still cancel orders and make withdrawals.
We expect to resume normal operation in 24 hours or less.
Why is trading still paused?
A new team member has discovered some bugs in the trading engine. They need to be fixed to avoid exploitation and maintain integrity.
We are re-writing the whole engine as we speak. Among other things, the new engine will propely indicate buys and sells in the market history.
Why is the site going down so often today?
After moving large amounts of altcoins to cold (secure) storage, our wallets were left with a large amount of small unspent inputs.
These inputs are generated by people who mine directly to our deposit addresses, using small automatic payments from pools.
When a withdrawal fails because too many unspent inputs must be combined, the coindaemon throws an error: transaction too large.
What are we doing to prevent this? We will manually consolidate the inputs by sending coins to ourselves.
This can take some hours, but its not a major issue. After this, the coindaemon will be able to choose less inputs for withdrawals.
Our system is configured to stop on any unexpected condition, at all times.
This is why the server is going down - we have to manually restart it after a coindaemon error.
This is the same protection that has prevented people from exploiting bugs - its important to "crash" on any unexpected condition.
We carefully inspect all error messages before restarting the Node.js process. Having this protection will make us alot better in the long run.
Also, because of the Mt.Gox situation, withdraw volumes have increased on all exchanges.
Because of this, we have to re-fill our hot wallets, multiple times per day. Usually its the other way around.
Another effect of increasing withdrawal volumes is that the coindaemons run out of "large" unspent inputs and start combining the dust from mining deposits.
What happens to stuck deposits, withdrawals and other support requests from the last pages on this thread?
Less than 0.5% of users have a stuck transaction. Mostly because of e-mail delivery issues (working on them aswell).
All recent support requests will be handled in 4-24 hours, starting now (already handling some).
If your confirmed withdrawal failed without an txid, dont worry. We will manually resend these later.
We are hiring more people - another Node.js/crypto developer and 3 additional support team members from different timezones around the world.
If you happen to be (or know) an expert in Node.js, feel free to contact us.
We will start moving our infrastructure to a production-ready datacentre as soon as the dust settles.
Websockets-based realtime UI is ready and will be deployed next week.
While the markets are paused, everybody can still cancel orders and make withdrawals.
We expect to resume normal operation in 24 hours or less.
Support Team Member?
maybe that happened with the 7BTC buy order on Mint, it went offline they bought it and put it back online and sold it on mintpal or cryptsy... you never know.... 
right?
