Today possible find plenty of commercial and few free(or de bene esse free of charge, if will want) automated trading systems for Forex. Their rapid grow is explained by increasing speeds of data processing and similarly by a difficult mathematical formulas, applied at an analysis. I will not jut out into reasonings about the questions of efficiency of one or another strategy or software, its realizing.
It is ungrateful pastime and there is a risk to be reputed a man, using black PR in the aims. My reasonings will carry rather a technical tint and potentially can help you at a choice or planning a new software. In latter days we had to do some the work with set of automatic systems of third-party developers and I marked on your own some vexatiously mistakes. I will make attempt tell about most considerable.
End users
Textual programming language now almost always present in entry or middle level trading systems. He can be used for strategy coding or/and for extending of existing functions of application(e.g. – creation of new indicators). This is wonderful possibility! Question is for whom. In many cases developers forget who will be end user of their product. That’s why we have systems where strategies are written on C or C++. I personally don’t see problems here but thing is that I am earn by programming, not from trading. Trader is not a software developer and he don’t have time to dive deep in shading of C programming. No time, no desire, no mood.
Certainly, when manager will try to sell you that type of system, for sure he will tell about high performance and such kind of things. Yes, this is a true. C++ unambiguously faster than scripting languages. But reality is that in 80% of cases you will not have and advantages of more faster system. How many orders do you plan process simultaneously? How many accounts to manage? 1, 10, 100, 500? We successfully did applications on Java which handling thousands of orders on usual computer. Yes, if you need to work with 10 000 of orders – superfluous productivity will not injure, but, probably, out-of-shelf system will not meat your requirements and you will use something custom, created especially for you, application.
If you unable to develop strategy personally, you need search for external consultant. And here you will be surprised by level of prices! Good C developers expensive. What does remain except to appeal for help to the company what developed and sold you software? How do you think why they used C or C++ 
?
My personal opinion is that using scripting languages like Ruby, JavaScript, Perl inside regular trading system is absolutely normal solution. In many cases this languages is only high level wrappers around native application libraries and play role of glue. As for extending system itself by adding new features – here better not sacrifice speed and use native language of platform. If it C – better write on C, Java – on Java and so on.
Original source here
