Why does this nonsense resurface every now and then ? And why is it so woefully misinformed ? And why does anybody with any common sense expect IB to rewrite TWS in C#/C++/Python or anything else without providing any real evidence of any problem that would be solved by doing so when the costs and risks of doing so would be substantial. The "stop creeping featureism, I don't want features, I want stability" brigade should be having fits at any such suggestion and the risk to stability that it implies but I'll bet there are some of the Java bashers/rewrite TWS crowd amongst them too. Lets take a look at some of the so called "arguments" - 1. The myth of slow performance. If anybody can demonstrate inadequate performance then they should go ahead and do so. I've yet to see it. The reality is that response time in transmitting an order or getting a price/bid/offer update is less than an order of magnitude less than data comms latency and is therefore irrelevant. If there was a performance issue, then you can see it in CPU utilization with task manager in Windows. But nobody ever cites CPU utilization as an issue - basically because it isn't. 2. Memory. 50 - 70 Mb is not excessive on a modern PC. Memory now costs less than $100 per gig. This is a non-issue. 3. Stability. Current JVMs are quite stable and reliable. It is absolutely not true that they are buggy to the point of compromising the stability of TWS. Anybody banging on the stability issue is years behind current reality. Anybody with evidence (rather than hearsay) to the contrary please step forward. 4. Java is dying. Absolute rubbish. Just look at the number of jobs advertised for Java developers. As with many things in life using Java is a compromise but as far as I can see, it is the right compromise. IB should be applauded for cross platform support and I expect it will not drive customers away by abandoning it. If you want cross platform support then there are limited choices available. C++/QT is suggested and excellent as QT may be there are very few programmers with QT experience. This is a very serious consideration for any organisation especially in this type of environment. What if key developers leave ? Where are you going to find replacements that can hit the ground running ? Even very good people will take some time to learn. Python/QT has the exactly the same problem with the added issue of a question mark over performance. Java is backed by SUN, IBM, Oracle and so on. This is important to IT managers. There is no other cross platform environment that enjoys this level of support. The reality is that the only realistic options are Java, or being locked into Microsoft. IB has made a correct decision with Java.
Rather than adding a non java TWS or even new sophisticated features to TWS.... I'd rather IB invest time/effort/people/tools in a good quality management and regression testing process to make sure that for each new built.... new features work as they are advertised and old features are not broken.
Just provide an ATS API without the having to have to go through the TWS. This is a very resonable request. ktm
IB is an incredibly well-run organization. They're not going to bite the hand that feeds them. When/if the time comes to migrate to standalone, they will do so. If clients demand it, they will provide. The probs with java aren't anything that more proc and RAM won't fix.
I'm not sure what terrible problems folks are having with Java, but I see very few 'Java' related issues. The charts are a bit clumsy and sluggish, but the platform itself , seems to run fine. You do need a decent machine, well configured. Easily done for well under a grand these days.
Don't they have now? FIX API. I used to think of the same thing, but now I prefer to bipas a frontend software. If you write ATS built on FIX from scratch, the overhead task you must achieve is way too big. Need to code an auto-reconnection using heatbeat or filtering FIX message or a ceratain risk management etc, needless to say the account, oder, log screen.
Personally, I don't like Java, but I understand some appreciated the platform independence benefit. However, platfrom independence is not that important in fact simply because the defacto standard of trading world is Windows. X-Trader, easyscreen, TradeStation, esignal,.... every software is written in C++ or C# now. When TWS version beta is created, it was Java boom, and considered cross-platfrom is something cool. That's it. The truth is that numbers of users who appreciate TWS Windows version is more than the numbers of who appreciate Java version.
It doesn't exhibit this type of behavior for me. (JDK 1.5, Linux), ie no mouse clicked event is sent to either column. In general Swing allows the application programmer complete control over mouse events, should they wish to alter default behavior.