Thanks everybody for the thoughts. I'll answer questions individually when I have a moment -- weekends are for the family so I just get a little time in the morning before everybody wakes up.

I see what you mean about the centralization issue.

But to be clear: The only person that would need to get an API key would be if another trading site wished to open and trade with people on this proposed site.

Perhaps the Bitcoin community would want this to stay with each site as an island unto itself and I can appreciate that. But when I look at the markets we know -- US equities and the internet advertising market -- and I apply the principles of those "exchanges" onto BitCoin, those are the differences I see. And the assumption is, Bitcoin will look more like that as it grows up.


Here's the inherent risk of creating such a platform and the reason why direct access to it would have to be carefully granted:

If somebody came along as a new trading site operator and got permission to trade with counterparties on other trading sites thru said platform, they could quite easily just invent users and add account balance with a SQL script and then "trade" on the open market and at the end of the day SiteA, SiteB, and SiteC would all settle their business for that day with this NewSite and they'd all transfer thousands of dollars into the site and then -- poof. Gone. They cash out and have just robbed the platform.

It's the same reason very few agencies (relatively) can trade directly onto, say, the NYSE or Nasdaq.

So if people are reflexively against the kind if community-policing you'd need of new trading sites, then maybe we just keep it as every-site-as-an-island.


Truth of why I'm considering this is that it scratches a hobbyist itch (I'm just generally more interested in this market than building another ad tracking system). And because even at the low surchage per-trade we'd charge, I think the Bitcoin market (black friday nothwithstanding) is going to blow up and we could make a lot of money. And if it doesn't, since we already have a lot of these types of tools developed to make it easy to build this site, there's not all that much risk.

Though to answer one question upfront: I don't think we'd be able to base this business onshore in the US. But we have a brilliant attorney so we'll see what she says.

And it's worth noting that she's got veto on all of this. If she doesn't feel she can shelter us and our existing business, well, we won't risk jailtime over this idea

I've been watching the BitCoin community -- and this forum -- closely for several months.

I'm a software developer. Myself and 2 partners run a consultancy. We've done a lot of general work but our speciality -- and what we bill as much as $200/hr to do -- is building highly scalable tracking systems for the advertising and affiliate marketing industries. We build systems that can track 100 billion impressions a month with the reliability you need when people are spending 6 and 7 figures on advertising campaigns. (Though nobody we've ever worked for bills on impressions -- nearly everything is billed on Conversions or Sales or Clicks -- everybody wants accurate stats down to the impression).

We've also worked recently for a smaller High Frequency Trading firm. And we've done a fair bit of PCI Compliance (credit card acceptance compliance) in the last year because those changing rules made that a lucrative market.

In other words, we're not weekend hackers.

I personally have taken an interest in BitCoin because I see it as an interesting technology and perhaps the first bit of people-powered-economics since bartering gave way to gov't issued currency.

Here's why I'm posting today:

I've kicked around the idea of building another trading platform and website. We have a lot of custom code that we own the rights to that we reuse from project to project and we're all quite good at what we do. And we look at the general state of BitCoin trading platforms (like MtGox) and we see a very underdeveloped system. We see a community flying by the seat of their pants -- which is quite understandable. What we go back and forth on is if we see a real opportunity to provide a useful service and hit at least a break-even point on our investment.

The first thing we would build would be a very simple trading platform (similar the API's that the other sites are currently using). This would be its own system, a first-class project in its own right. It would perhaps have its own simple readonly website that publishes a "ticker" of activity on the platform.

Then we'd build a site with functionality similar to TradeHill, MtGox, etc, etc, that would merely be granted the first API key on the trading platform.

The fact is we've got a number of home-brewed tools that help us produce reliable, secure software. And that seems to be a novel idea in the world of slapped-together "weekend projects" the existing sites seem to be.

If you're interested in the technology, among other things, we've got a customized version of Tsung we use to stress test systems, and we'd be able to build a blazingly fast trading platform. We use Python for most our web development but we also commonly use Java for things like stats aggregation daemons. We use either Percona MySQL or PG SQL. We have a fantastic python static analysis tool that helps us find security flaws in our code, as well as a continuous integration and security-flaw framework we built on Selenium that hits with a test suite after every deploy. And of course we practice TDD on all core systems and use a CI server to run all our PyUnit (or PHPUnit or JUnit) tests on each SVN commit and we always know who broke the build.

SO my question is to anybody who is still reading: 

Is there a receptive market to this concept? Should we invest that much time in the bitcoin community? Are we going to get a frigid response from people who prefer their current favorite trading sites?

And FWIW, we wouldn't wish to put them out of business. Ideally, though, all those sites would be sharing a central trading platform so people can Sell on siteA to buyers on siteB and siteC. And the platform itself would earn a share of the commision that's large enough to support profitable operation but small enough to leave the trading site operators a comfortable profit margin.

However to let that happen we'd have to have a rigid process of granting API keys because we'd have business each day that would rely on the integrity of each trading site to settle.  That is, if you sell $100/worth of bitcoin on SiteA to somebody that buys on SiteB, siteA will "credit" your account $100 in their system and SiteB will deduct $100 from the buyer in THEIR system and at some point throughout the day or perhaps in batch at the end of the day, actual cash will need to move from SiteB to SiteA.  Of course, ideally, the actual sum changing hands will be a fraction of total trades that day -- that is, if there's a transaction later in the day where somebody on SiteB sells $90 to somebody on SiteA, then at he end of the day only $10 needs to be settled between the two sites. And that settlement, of course, would also use the trading platform in some way.

Consider this a first, open proposal.

Thoughts? Comments? Criticism?