 Show Posts
|
|
Pages: [1] 2 »
|
|
nix930 - that does sound pretty bad. Why don't you create a new thread with an appropriate title and you might get some more responses?
|
|
|
|
Bitcoin transactions cannot be rolled back.
What you are talking about is ONE bitcoin exchange, and even then, they didn't 'adjust bitcoin prices'. They simply removed some of the trades that occurred during an intrusion into their database. After that, bitcoin prices were subject to market forces as usual.
Exactly. Because of the decentralized nature of Bitcoin, a mass rollback is not possible. I think the 'flash crash' has been misrepresented in various articles in the media - people don't realize that it only affected internal transactions on one exchange. |
|
|
|
|
The zero trades I saw were on Britcoin, by the way.
|
|
|
|
|
It's good to see people suggesting languages that push the boundaries like Haskell and Erlang - although I'm not sure I'd recommend them for a novice programmer as the learning curve is probably a bit too much to start with. Once you're ready to have your mind seriously expanded, check them out...
|
|
|
|
It looks like a trade of amount 0 has gone through and has caused the problem. This is the culprit: {"date": 1311104701, "price": , "amount": 0}I also noticed a trade in the order book earlier today where one of the quantities was zero so looks like the validation may need tightening up somewhere. (Edit: that's a guess, I haven't read the code in any detail) |
|
|
|
Just a heads up - the ticker API seems to be broken as it is returning: {"ticker": {"vol": 561.62370667, "buy": 8.4035, "sell": 8.65,
Warning: Division by zero in /var/www/britcoin.co.uk/htdocs/api/ticker.php on line 72
"last": }} |
|
|
|
|
This sounds a little bit like HashCash, which was around before Bitcoin and I believe influenced it.
|
|
|
|
please don't remove support for openid! if myopenid.com is down then it's really easy to set up an open id at a different provider, and open id is so much more flexible/secure than having a logon with a separate password, risk of password database compromise, and no support for two factor authentication (which I use with Verisign PIP).
I agree, it would be a shame to see it go. Perhaps you could leave it as an alternative option for more technical users? That would probably save you the hassle of migrating all the existing accounts as well. |
|
|
|
|
TradeHill seem to use a captcha to prevent logins being bruce forced. It seems a bit of an odd approach to me ... wouldn't rate throttling on the server be a better solution that wouldn't inconvenience the users every time they log in?
|
|
|
|
HSBC-to-HSBC clears the same day without having to pay for FPS; was his point I think.
You (usually) don't have to pay for Faster Payments - if both banks support it, a normal transfer should go through within an hour instead of the usual 3 working days that BACS takes. |
|
|
|
At the moment the trade page is displaying: Warning: fopen(http://...) ... [details removed] ... Probably worth mentioning that you shouldn't really be exposing error details to users... Edit: seems fixed now. I have removed the error details in case they contained anything sensitive. |
|
|
|
Why should I read it? I know that it's wrong.
Controversial, I know ... but maybe you should read it before you decide whether it's wrong or not. Seemed a very well thought out comment to me. |
|
|
|
I know what my client is doing. Do some research before you expose your own ignorance. The packets aren't even encrypted. You can use a sniffer and look at them yourself if you don't trust the code. Nice, why not make it personal. Did you actually read the link? I don't see that packets being encrypted has any relevance. |
|
|
|
Guys, the hate isn't necessary. Full Disclosure vs. other methods is a (computer) age old debate that is like arguing Right Wing politics vs. Left Wing. No one is right, no one is wrong. They are opinions on how to handle these situations. If MtGox wants people to minimize impact of disclosed security vulnerabilities, they need to fix them promptly. I assume they are doing so.
Absolutely. And what people don't seem to realize is that Full Disclosure is infinitely better than No Disclosure, which was another option open to the OP. |
|
|
|
|
Presumably this has happened because the order book has stopped reconciling, but they carried on taking orders. So the question is - when it starts working, how are they going to match up all the overlaps fairly?
|
|
|
|
Free bitcoins! What are the chances that loads of people will exploit it, and they'll end up doing another rollback  |
|
|
|
I agree that would be great, but I don't know if it's possible. The current clients are basically doing password cracking/hashing - that probably wasn't one of the "good causes" you had in mind.  |
|
|
|
BTW that last post was irony, not spam |
|
|
|
|
