Ask HN: Why haven't we built a decentralized exchange for bitcoins? | Hacker News Reader

Ask HN: Why haven't we built a decentralized exchange for bitcoins?

6 pointsEGreg17 comments

It seems to me that a lot of the latest bitcoin crash is attributed to MtGox being a centralized place and unable to handle so much trading activity.

The question of centralization has come to bite us multiple times over the last 10 years, as the internet has centralized around AOL, then Facebook, Google etc. Every time there is a single point of failure and various problems can arise from the disconnect in policies of the provider and the customers.

I wrote about this issue 2 years ago, and I believe eventually we will find better decentralized solutions to a lot of the challenges today (social networking, trading, etc.)

http://myownstream.com/blog#2011-05-21

But if BitCoin was a tour de force for decentralization of currency, why haven't we had a prominent decentralized exchange to go along with it?

View on HN

Bitcoin itself is unable to handle the rate of transactions. Currently, the Bitcoin network can verify about 0.75 transactions per second, because of the block sizes.

This was what broke in the 0.7/0.8 fiasco. 0.8 would have increased the block size and upped the transaction rate, but broke backwards compatibility and blah blah blah.

You may want to read further about the Buttercoin project. The promise is an open source exchange software that runs on commodity servers. It's still in the early days, but seems pretty possible. If this project were successful, then multiple exchanges could sit on top of this software and in effect decentralize the exchanges.

It's not a decentralized exchange. It's an open sourced exchange. It will be up to entrepreneurs to take the buttercoin code and set it up as the engine for their exchanges. The majority of the work will not be in the systems, but in setting up the exchange as a proper legal entity. There is significant cost in operating in the US, because you have to register as an MSB at the fed level, plus once with each state. Fees run in the thousands per state, per year.

So I think the OP's question was why there aren't decentralized exchanges. I took that to mean why there aren't P2P style exchanges. I'm not sure what the answer is. I think people have been tossing that idea around for a while but I'm not sure if they have started code for it. There are many exchanges now, but for some reason people choose to use MtGox, which is AWFUL.

I agree with this - I think it's the fact that MtGox is the main exchange that's wrong. There should be multiple exchanges with proper backing.

multiple exchanges could sit on top of this software and in effect decentralize the exchanges.

This model leads to failure. Each exchange would have a small order book which would discourage anyone from using it, ensuring that liquidity never reaches critical mass. Small exchanges cannot afford security audits so they are guaranteed to be hacked. Small exchanges also cannot afford FinCEN compliance, so if they touch USD then they're illegal.

Agree that lots of small crappy exchanges with incorrect legal status will pop up, but that doesn't mean that at least one or two or three large-ish buttercoin ones won't start up. I think that will happen. So your point is partially correct, but it doesn't matter, because larger ones will pop up. There is nothing to stop Gox from migrating to buttercoin either.

I think the first credible implementation that doesn't have any major red flags in terms of trustworthiness or being run by amateurs will quickly siphon off volume from Gox.

What are the main challenges that explain why it hasn't been completed yet? Seems like a clear win over centralization, kind of like git/mercurial vs svn.

> What are the main challenges that explain why it hasn't been completed yet?

Because its pretty hard to define even what "decentralized currency exchange" means, and, particularly, to identify a mechanism by which you could do order matching in such a system so that wouldn't leave you with just a bunch of disconnected exchanges that are far less efficient at market clearing than a centralized exchange.

buttercoin literally just started 2 days ago (on 4/13). It's coming along quickly, a lot of developers have shown interest. It might be a while before we see an exchange based on this project. There are other high performance exchange projects starting up too.

I don't think buttercoin vs any of the existing exchanges has any proper analogy that would match your git/mercurial vs svn example. The only diff between buttercoin and those is that it will be open source. It doesn't actually decentralize. You might even see MtGox take buttercoin and use that to run their exchange.

I don't know this space well, but if I understood the SVForum EMTech talk I attended last week, that's what Ripple is doing.

https://ripple.com/

How does ripple work? Everyone is charging 2% fees...

OK, how do you build a decentralized currency exchange? No one knows.

Well if you have a decentralized ledger of transactions, it can't be that hard to also add another ledger where people transfer bitcoins one or more virtual "channels" where others could buy them? These channels wouldn't actually correspond to any servers but would just be virtual bitcoin users?

Bitcoin is a decentralized exchange for bitcoins - it's built into the system.

No.

Um, yes.

When people say exchange, they don't just mean exchange, like change hands, they mean something akin to the NYSE.

Reply on news.ycombinator.com