same observation here. The developers seemed very cocky and arrogant while maintaining and servicing a severely lacking product. Whenever someone challenged them they either ignored such person or shot angry responses back on their own forum, not the greatest way to give a good impression about their service. In the end it always came down to the same point, whatever retail targeted off-the-shelf product there was, it was lacking beyond the absolutely most basic needs. For a serious quant trader there really is no way around developing your own thing. It impressed so many people that I get uncountable offers from hedge funds but I still stuck to my own fund that I started for reasons related to wanting to safeguard my IP.
I would not consider anything that even mentions the term "web". The full deltix product is very capable and I would use it if they would not charge a leg and arm for their product. It only proves there is nothing for free in life and I learned it the hard (working) way that there are no shortcuts. Actually having coded up my own platform from scratch made me become so good in programming (I have traded for over 13 years at institutions and do not have a CS background) that I now run technology and fully understand it that not even experienced bank IT coders would touch without having to spend a considerable amount of time to feel comfortable with.
As this is a trade link related thread I find it telling that there is not a single response from the trade link guys despite them closely monitoring threads and mentioning their product whenever the opportunity arises. I think I made fair and most above truthful comments re trade links' shortcomings and had hoped to get feedback from them but nothing so far....
I looked at Marketcetera. Great looking product that seems to never be updated. The user community is practically non existent. At least with Tradelink, you have a lot of users and activity in the user group. Maybe the other one you are thinking of is Algotrader? It too appears to be something to watch. Not quite ready, but could be a good one. I'm not a Java person, but if I were, it would of been what I chose over Tradelink.
Not sure to be honest, I have never really looked too deep into Java platforms, I had a different name in mind but I just can't recall it. Sorry.
we won't have time to respond to everything, we're going to try to address the most specific post by AI and avoid the cross currents. if we miss something, try the tradelink users list. a tip here for getting responses.... be socially courteous, contribute things back to the project, help other people... and that will greatly increase your odds someone will help you. "1) I have not seen even the capability in Tradelink to merge multi symbol data into one stream." The multisym implementation of HistSim provides this. It's turned on by default in all the platform apps, usually called portfolio sim. Last year we also added a single sim interface for people who have extremely simple strategies and want the extra speed from avoiding the merge. "2) The profiling and backtest platform is different from the live trading one, which I personally really dislike. I can turn a switch to change from my historical data feed to real-time data and the quote streams flow through the architecture in the exactly same way." You could easily merge two of these apps together into one app, if you prefer. Since many retail platforms operate this way, you're probably not the only developer who would be excited to work on this. 3) They over advertise, they say they have a lot of adapters and interfaces to brokers but when really looking into those they severely lack. Its easy to say "hey we offer some sort of fix connectivity to IB" but when you look closely its nowhere close to even be usable for testing purpose let alone for trading purposes. There is no IB FIX connector, so you must've been misinformed. If there is one out there it's definately not in the tradelink project. The IB API connector is extremely stable and one of the most popular connectors. Sometimes third parties (like broker dealers) contribute connectors to tradelink that have not been built by people familiar with tradelink, or not tested well. these are labeled as beta. if you have other suggestions for doing things better, all ideas are open for discussion in open source. "4) Gui? No gui, no charts, no data grids, at least not what I would call it as such. So, if you get to run a backtest after all the hard work to get it up and running how do you profile the results? You still end up writing pretty much everything on your own. " If this is referring to tradelink we're not sure what to say other than perhaps you haven't used tradelink in a couple of years. Tradelink includes a non open source add-on tool so you can build tradelink strategies in minutes (yes from a GUI) without having to watch any tutorials or even really know .net or programming. Charting is definately present both in the platform and the SDK, and included in at least 3 of the platform apps from my memory. charting is not at the level of esignal, but it's very functional and allows multi intervals, manual and automated drawing (iow indicator plots), screenshots, and probably some other things I'm forgetting. I know charts don't support oscilators, but that is fairly easy to add if one wanted this. We receive contributions back from users sometimes with chart improvements, and we also hear about people using 3rd party charting libraries with tradelink. such is the beauty of open source! regarding backtesting results, there are something like 30-40 of basic summary results stats (winners/losers/sharpe/etc) included, in addition to csv export of fills/pl so you can do excel or google spreadsheet pivots or generate r charts. if you add something, contribute it back. the "main guys" or primary guys that amazing industry refers to... they manage the QA for the project and do the checkins. previously we used to have lots of developers with individual commit rights, but what happened was that people would build things because they worked great in their own environments or for their own particular application or concept but broke others, sometimes in unexpected ways. now we have more of a linus-torvalis-like system where we let anyone submit patches but they have to meet some standards (like unit testing) to get commited. eg someone recently submitted a brute force optimizer they wrote and didn't have time to maintain seperately. if you don't want to wait for this to be integrated, you can start using the patch right away or speed the integration yourself. with open source, the ball is in your court. even more so the platform continues to improve even when you're not working on it, which is not the case when you do everything yourself.
With all due respect but as usual you spin your story in a very good light. But one thing I like to set straight before responding to your replies. I laud your effort and the open source and community approach. I do not think I ever criticized the product overall and I did contribute ideas some time ago. My issue lies with over-promising in the VERY SAME WAY than you just did, someone reading your posts gets the impression it all works fine, no errors, no bugs, please check your own forums yourself, your claim that the IB API is stable is anything but true. There are constant issues with it as documented in the message trail. Those working on the adapter cannot be fully faulted for that because IB does not have the best track record when it comes to pumping out clean and bug free new API versions. 1) I admit it was some time ago when I checked but I am absolutely positive multi-symbol merge historical streams did not exist 2010. I asked about it and discussed it in length, there was working being done but it was unusable and incredibly slow. Please can you inform the community of the speed an average platform processes quote/tick data per second on the multi-symbol engine? Can you provide couple examples, examples make everything very clear and people can judge for themselves whether it fits their needs before committing tens if not hundreds of hours finding out later it may not suit their needs. 2) "you can EASILY merge the two"? Come on, please let's be a little more truthful here. I understand you try to build a user community and potentially like to build a commercial model around it but I find it quite irresponsible to say a testing platform and a live trading platform with entirely different code bases (yes they do share common interfaces) can be easily merged together. Thats simply not gonna happen EASILY. 3) Anyone can do a simple search on the TradeLink google discussion forum for "FIX". Yes there is a half-developed FIX adapter but its essentially un-finished, unusable. Again I am not criticizing that its not done I am criticizing that when I discussed it and asked questions on the forum the adapter was over-advertised as being usable when it was in fact far from being that. 4) Ok, mate, with all due respect, if you call the crude windows a GUI then I give that to you, sorry, my apologies, I stand corrected. But I am talking WPF grade charts. A GUI in a trading platform is something where you start and stop back tests, where you set ALL parameters, where you receive ALL statistical analysis analytics, where you potentially can zoom through trades on a chart. If this is something that was added in the past 7-8 months then yes I am happy to stand corrected but before that it was DEFINITELY not there. Risking to repeat myself, I am not saying it is supposed to be there, I am saying that it should not be called GUI because most people understand something different under GUI. You mentioned yourself one needs to use a different tool to look at fills/trades, the data is dumped into CSV. All great, but just mention it openly that most of the result data from a back test is dumped into uncountable different destinations and not brought back to a unified entity to which a GUI relates, because thats what Dashboards and GUIs are exactly designed for. All I am saying is if the product can be advertised and described more transparently, including openly talking about its lacking features and weaknesses, then more people are able to estimate whether its worth contributing or not. That is all I am saying (even I said a lot now ;-)
There seems to be a bit of smugness from those in charge of the project. I've seen and experienced quite a bit of rudeness from the same people as well. Haven't had a message deleted, but I've basically been told to shut up. However, I have never run an open source project, so perhaps it comes with the territory. Unfortunately, the alternatives are slim pickings. Unless you want to write your own platform, there isn't a lot to choose from.
most of this stuff is going to only apply to sdk folks. "1) I admit it was some time ago when I checked but I am absolutely positive multi-symbol merge historical streams did not exist 2010. I asked about it and discussed it in length, there was working being done but it was unusable and incredibly slow. " That's about half right. multi symbol was always there but it got huge speed bump in 2009 or perhaps very early 2010. it wasn't called multisym until single sym was added. "Please can you inform the community of the speed an average platform processes quote/tick data per second on the multi-symbol engine? " google tradelink speed "Can you provide couple examples, examples make everything very clear and people can judge for themselves whether it fits their needs before committing tens if not hundreds of hours finding out later it may not suit their needs. " google tradelink project or tradelink quickstart 2) "you can EASILY merge the two"? Come on, please let's be a little more truthful here. Based on your comments I can tell you are an experienced and accomplished developer. from a programming point of view, Tradelink makes that easy. more below. the harder part is usually the community part. massaging the community to get the help you need, but also learning from when you don't get a response. also I certainly don't claim we know how to do the community part right, it's hard. if you're a pure open source user, we always encourage people to start discussions and get involved. don't depend on the "primary guys" to do anything other than their best to ensure your contributions don't break others contributions. "I find it quite irresponsible to say a testing platform and a live trading platform with entirely different code bases (yes they do share common interfaces) can be easily merged together. Thats simply not gonna happen EASILY. If the tradelink SDK was just interfaces I would agree with you. because the SDK kadina and ASP apps are NOT entirely different code bases. merging them into a single app, that is not a difficult project (for people like you that have "the skills" and also the desire). since many of the biggest contributors including myself came from big hedge funds where we had platforms that supported large varieties of trading applications and strategies... we've always tried to stick to this intention even as the platform has evolved since 2007 (open source since 2008). so we should have something to offer here, but obviously you have to judge for yourself. 3) Anyone can do a simple search on the TradeLink google discussion forum for "FIX". Yes there is a half-developed FIX adapter but its essentially un-finished, unusable. Again I am not criticizing that its not done I am criticizing that when I discussed it and asked questions on the forum the adapter was over-advertised as being usable when it was in fact far from being that. Ok, except I'm not exactly sure what your criticism is. I think FIX template was added in late 2010. In tradelink we mark our FIX connector as a template, to differentiate it from named connectors that are already turnkey for specific brokers. I would agree that apart from tradelink.... FIX in general is over-advertised as being a turnkey standard. 4) Ok, mate, with all due respect, if you call the crude windows a GUI then I give that to you, sorry, my apologies, I stand corrected. But I am talking WPF grade charts. We would love to have more UI focus. But I think it's unfair to suggest to people we don't have any charts and UI and so forth. We've had charts since 2007, maybe not amazingIndustry-riffic charts but they are quite functional. you can see prices and days and everything and draw on them. "A GUI in a trading platform is something where you start and stop back tests, where you set ALL parameters, where you receive ALL statistical analysis analytics, " Again maybe this is a several years ago criticism, but you can actually set all parameters outside of code. Sounds like you could probably do much better GUIs than us. As I mentioned previously there are 30-40 stats included, and easy export to R/excel for custom no programming stats. Sounds like you could probably snaz this up a lot and maybe make it more intuitive to use. It sounds to me as if you could even do it pretty easily, but what do I know about that. We'd encourage you to add this stuff in an SDK friendly way, so it can be unbundled and incorporated to as many of the apps and use cases as possible with min effort (and to simplify merges such as you suggest above). but really even if you don't do this or if a contribution isn't incorporated in the trunk, it will still be used and appreciated and all that gooey warm stuff. hopefully this was helpful. i probably won't comment too much more here but will try to read the entire stream.