Okay, given the number of members, I guess it wasn't around yet when I was really searching for one over a year ago. I've already converted stuff to C# now. It would have been nice back then. That is, if it's really good.
Without a doubt, that's why members of tickzoom all agree that we keep the tickzoom user base restricted. So only a limited number of new users get admitted for each release. For example, when beta finishes shortly, only 4 get admitted. Wayne
blouelou: I've just played with it a bit and have not used it in production. I found it nice, because it seamlessly combines a lot of functionality and at the heart has a tremendous and well tested OMS. As of the last time I checked, you could rig backtesting capabilities in by creating a custom data feed that just read your binary data file or whatever and executed your orders on a simulator. I think it's a good platform to extend for your own general purposes and if you do want to do something from scratch, borrow/translate code pieces from it (like the OMS). I've talked with a guy from MC and he's told me that their business model will be similar to MySQL, where they open source the real engine of the application (the OMS, which is what their main clients are after, since the competition is expensive), while custom modules will be for sale (data feeds, brokerage connections, additional strategy programming languages).
JSystemTrader has been developed for at least 3 years that I know of. The discussion groups at the current location may be newer however. Marketcetera has been developed for at least 2 and half years. Other Java based projects have been in development even longer e.g. ActiveQuant: http://www.activestocks.de/ Not sure where you were looking.
No problem, you might want to look into Marketcetera's CEP engine which is very fast, mature and stable. Their engine can process 500,000 event a second on a dual CPU 2GHz Intel based hardware, with engine latency around 10us according to their website. Which , on a back of the envelope calculation, would be much faster than TickZoom. So if I really wanted speed, I think Marketcetera may be a better choice than TickZoom.
Hi bluelou, Thanks for a quick review of Marketcetera platform. I work at Marketcetera and wanted to point out couple of things: 1. Marketcetera open source platform is actually gaining good adoption within hedge funds and buy-side segments. Retail traders obviously would like more out-of-the-box (broker-terminal like) functionality- but quants, traders and technology developers at buy-side firms like the openness and extensibility of Marketcetera platform. Marketcetera packs in FIX connectivity to multiple brokers, ability to automate strategies in Ruby or Java, integrated market data and complex event processing and other proprietary and external libraries. Many firms have taken the platform and added support for asset classes like Futures. So, the real attraction right now is the ability to build upon the solid foundation that the platform provides - as opposed to out-of-the-box bells and whistles. 2. The other aspect is performance and scalability. Hedge funds and buy-side firms deal with a large universe of stocks and related options. This requires processing hundreds of thousands of quotes and thousands of trade messages per second. The platform has been able up to measure up to these numbers and that is another reason for the rapid take up. Marketcetera is an open source product and interest and feedback like yours actually goes a long way in improving the product, So thank you.
Chinar, I wasn't looking for Marketcetera to bake bread for me. But it seems that bar-based event handlers (e.g., tick bars) and the ability to trade futures "out-of-the-box" isn't asking for too much. Does your firm plan on addressing the active trader community. The traders who aren't retail but aren't institutions either? Why else would you be on ET? I don't need Marketcetera for charting, manual trading, or even strategy development tools (though this last one would be nice). But, I'd think that you could broaden the appeal of your product by including bar-based event handlers (I don't think Esper can generate tick or volume based bars) and the ability to trade futures. -Lou
Last I checked, Marketcetera could not use a quote feed from FIX, and this was a show stopper in my eyes (or, I couldn't find it in the docs). It does seem like an interesting project with lots of potential, and I have hopes of seeing good things from it in the future.
Interesting discussion so far. I think the big draw for a MarketCetera type framework for me is the reduction in time it takes to implement something. Sure its possible to do a custom framework for yourself but, for me, that isn't the best use of my time. My time is better spent on coming up with new models & algorithms rather than building something that may end up very similar to a product that is very similar to one thats already tested & proven. For better or worse, my strategies aren't overly complex that they can't be handled in existing platforms like MarketCetera, albeit with a bit of changes. I do see value in having a deep understanding of the framework you're using and you would definately get that from creating something from scratch. I don't know that you need to do write the whole thing to get a deep understanding of the framework. It seems to me it would be a lot quicker & more efficient to spend a month or something learning the ins & outs of marketcetera instead of spending 6-9 creating your own app. Any opinions on that idea???? Eric
I don't know if this will be useful...but I justed started trying out Marketcetera for a linux deployment and here are my comments thus far: 1. For Linux, they only support Ubuntu 8.04: not necessarily bad but a limitation 2. For the Linux installation there are a few undocumented quirks: this should improve with time 3. You really do have to understand the architecture of Marketcetera and how components relate to begin using the system -- there is a steep learning curve to start. 4. It is open source and very flexible...it is easy to add missing functionality or even replace current modules with your own implementation 5. It takes away the need to architect your own trading system