Auction ended. Final result:

Slots BTC/Slot Person
1 1.30 Neo501
4 1.25 Stunna
2 1.25 MariaQin
1 1.25 victorhing

Done.

The idea that bcrypt is somehow extra strong is AFAIK entirely a myth. bcrypt is based on a fast Blowfish-based hash function comparable to SHA-256 and other cryptographically-secure hash functions. It makes the entire process slow by hashing the password many times. But this is exactly what any decent key derivation function does.

The forum uses sha256crypt (which has an extremely similar interface to bcrypt) with 7500 iterations. If SHA-256 and bcrypt's underlying hash function were exactly the same speed, this would be equivalent to a bcrypt cost of about 13. I prefer SHA-2 because Blowfish (and especially bcrypt's Blowfish-based hash function) are not as widely used or studied.

That's basically what the payment protocol does. You click a link like:
bitcoin:1NXYoJ5xU91Jp83XfVMHwwTUyZFK64BoAD?r=https://www.coinbase.com/r/invalid
And Bitcoin Core negotiates a transaction with the URL at the end via HTTPS. It's very similar to IP transactions.

This is reminiscent of Satoshi's old proposal for strengthening IP transactions, though it relies entirely on HTTPS and doesn't actually use the provided address for verification, so the security is probably better in general but worse in rare cases where the URI was gotten securely but HTTPS isn't secure.

I also preferred the overall look of the old UI, though it had far fewer features.

That GUI is nowadays called wxBitcoin because it used the wxWidgets GUI framework. After Satoshi left, no one working on Bitcoin wanted to work with wxWidgets, so Wladamir wrote a new GUI using the Qt GUI framework called Bitcoin-Qt.

I think that marcotheminer should return the account now since it was probably hacked, but everyone should give yussuf89 negative feedback for being unable to stand behind his account's actions unless he pays 50-75% of the loan principle. (This is just my opinion -- I'm not going to try to enforce it.)

An interesting thing about that screenshot is that it must have been made on a different block chain. The dates don't match up with the current block chain.

That also doesn't strike me as entirely fair. If nothing can be gotten from the person who actually did the stealing, then the person who lost the stolen item and the person who bought it should somehow split the cost IMO.

I think that the account was hacked, though this can never be known for certain.

The IP that posted the loan request from yussuf89's account also posted this:
https://bitcointalk.org/index.php?topic=960603.0
Someone with the same IP was trying to sell another account here, presumably hacked:
https://bitcointalk.org/index.php?topic=925024.0

yussuf89's password wasn't changed/reset recently except for when marcotheminer did it, so the attacker must have somehow gotten his password. Keylogger or phishing, maybe.

Zeki's IP matches yussuf89.

Since ownership of the account is contested, I'm thinking that I should not restore the account and let people give trust ratings depending on what happens. For example, if marcotheminer needs to eat the entire cost of this when he's mostly not at fault, some people might want to give yussuf89 negative feedback.

No one knows what AT is because everything you've written about it is either very long and technical or assumes that the reader already knows what AT is and what it's useful for. This applies to everything at ciyam.org/at. You need to clearly summarize its uses and give a basic outline of how it works in a few paragraphs, or else pretty much no one is going to dive any deeper into your documentation. (People who spend a lot of time on something very often make this mistake of assuming that everyone is interested and knows what you're talking about.)

About categorization:

Firstly, an "altcoin" is any non-Bitcoin currency, even if it uses the Bitcoin block chain in some way. And for something to be considered to accept or send bitcoins, it must be able to accept or send Bitcoin transactions directly from/to an unmodified Bitcoin Core client without any intermediate step.

Project-related topics belong in the non-altcoin sections if any of these are true:

1. People can make use of the project directly using bitcoins without any explicit conversion step, and the end result of user action is not just another cryptocurrency (or something similar). For example, if you have some sort of gambling application that uses a sidechain/altcoin/etc. to function, but people can immediately send bitcoins to it and are get bitcoins back, then that's OK. The site can also use an altcoin as "chips" if the site has an integrated and very simple way of exchanging between these chips and bitcoins. Smart-property systems are OK (whether Bitcoin-based or not) as long as the site or software that the forum topic is about can accept bitcoins directly in exchange for the smart property. A site where you pay BTC to get a Namecoin domain would be OK. But altcoin or colored coin exchanges and topics about individual altcoins or colored coins themselves would usually not be OK (even if based on the Bitcoin block chain) because the end result of user action is a non-Bitcoin currency.
2. The main purpose of the topic is to discuss the creation/extension of something that would meet the above definition. (This probably belongs in Project Development.)
3. The main purpose of the topic is to discuss a reasonably-possible extension to the Bitcoin protocol/software or general discussion about ways of using the existing protocol. For example, discussion of merged mining is OK; discussion of Namecoin is not.  (This probably belongs in Dev&Tech.)

I don't know enough about AT to say whether it belongs outside of the altcoin section or not. I don't even know whether it has a currency component.

You should create a topic in the Russian section to discuss whether a new moderator is necessary and, if so, who the new moderator should be. (Do not take a vote, though.) If there is consensus that one or more specific people should be made moderators, then I will (almost always) do it.

I think that the system was working basically as expected, but I made an adjustment which will reduce evil accumulation in some cases.


It links to this topic:
https://bitcointalk.org/index.php?topic=74046.0

There are only 75 IPs in the database that can by themselves cause a fee requirement of more than 0.01 BTC, and 19 above 0.05 BTC. Fees are often very small. See the inputs of these transactions, for example:
http://blockexplorer.com/tx/f01405a1b0b0350d1217a2471aac65378fcc7c73b0488ae59a66c963f02480e9
http://blockexplorer.com/tx/6109d4c7d4f6d5c7abb867b21aeeaa45289ff7045a4d5e244ba6370c246493ca

(It was my intention when originally designing the system that 1 unit of evil would require about an hour of work and about a day of waiting to get the fee via faucet sites, but it's probably easier than that now.)

It's a test to see who will try to delete posts that they're not allowed to delete. You failed.

It's a bug (or maybe an intentional optimization) that always exists in SMF AFAIK, but it's a very minor issue, so I never fixed it.

That's part of the anti-adblock system. There are a random number of fake posts before and after the ad area. They'll be hidden unless you turn off CSS. Disabling CSS is extremely unusual, so I don't consider it to be an issue.

The number is the unix timestamp when the page was generated.

It makes me really sad when people lose money like that, but I'm not sure how I can make things any clearer. Perhaps in future rounds I will require bidders to be at least Jr. Members so that they're more likely to know how the forum works.


Unrelated: I created this test page to help people design ads:
https://bitcointalk.org/ad_test.html
Download it, open it in a text editor, and look at the comments. (This is a little difficult to do with normal topic pages due to the anti-adblocker obfuscation code.)

This also happened previously; you just didn't notice it because the link wasn't so long.

Whenever your session code is in the URL, all links are broken. Otherwise, your session code will be sent in the referrer if you click one of the links and the destination is HTTPS. (Links are also broken with some JavaScript things like quick edit when your session code is not in the URL.)

I changed the link to a shorter one so it's less annoying.

No. There will be no more features added after it is released.


Yeah, it'll work like it does now. You'll post in Meta with your request and I might add it (or have it added) if it's a good suggestion. Epochtalk will also exist as a separate software project that other sites can use, so you could probably also post an issue on github or something if your suggestion is relevant to more than just bitcointalk.org.

OK, I removed the CSS that attempted to limit the width of tables and the CSS that removed the breaks between the "show unread..., show new..., ..." links, which I guess are what most people disliked about this.

I'm not going to work on this mobile stylesheet much. I am very much not an expert in Web design. (Take a look at blockexplorer.com. )If anyone knows what CSS would look good, tell me and I'll try adding it.


That's intentional. It looks weird (especially when there is no » because you're on the last page) , but it should make it easier to click.

BTW, my addition of next/previous arrows applied to most pagination on the site at once.

Yes, it's large text instead of an image in the mobile stylesheet to make it easier to click.

I don't know. It's triggered by this CSS media query:

What does everyone else think of it? If everyone hates it, I'll make it not so weird.