Unfortunately in its desire to be everything to everyone the existing architecture of the current JForex API incorporates way too many moving parts that time and time again are proving way too difficult for the team maintaining it to manage in a consistent fashion - bugs and feature regression hit the same areas of the API month to month over and over with the lack of any comprehensive regression testing suite making it hard to understand what is and what isn't 100% reliable and intact from release to release.

For individuals like myself who require a simple and secure two way connection for retrieving quotes and issuing orders, the FIX option appeals as it's an open and mature scheme largely standardised and unproprietary in nature therefore allowing for more portable solutions. As you continue to expand the portfolio of supported securities and attract even more and more people interested in automating their trading strategies, it might be time to waive the 100k restriction on FIX and enhance the ability for experienced developers to marry sophisticated solutions with reliable (and open) interfaces for market data/order execution.

Alternatively, another good option would be to have the current API split out into ESSENTIALS (simple two-way connection, account login/info, quote subscription, order request/reporting, etc) and ENHANCEMENTS (everything related to indicators, everything related to charting, everything related to the terrible local cache, historical tick/bar reading, back-testing, etc). This way the bare essentials can be made available as a simple Maven dependency SEPARATE to its superset which would be the regular Maven stuff currently available. I'd leave it to the good people there to decide whether the Essentials module would deal with strategy execution (personally I consider it an enhancement as does FIX) but the idea here is one could include the Essentials module and use it to connect, get data and put in orders without worrying about what is going on with Renko Bars or wondering if the last release regressed OrderInfo / AccountInfo type things while trying to fix bugs or enhance things in the Local Cache or Strategy Container or any other non-essential area.

While some may feel these points warrant little regard, in time I have no doubt that any opportunity missed to put things on a better footing right away will cost far more later than it will getting on the case immediately.

Keep up the good effort.

+1

Hi!

We understand the frustration that you experienced due to recent problems with JForex API/SDK. We are now in the process of some major internal software updates (it's been going for few months now) that caused a lot of unexpected situations and produced a lot of bugs. Currently we have solved most of those problems and the situation should become more stable now.

The idea splitting out SDK into two different releases is impossible to implement in current system's architecture (in terms of both API and implementation), we also don't consider it a good idea as it will only complicate the API and increase the codebase, thus requiring more resources to develop/support/test. So you should not expect that kind of changes to happen anytime soon.

FIX API and business questions can be addressed to [email protected] as these topics are out of JForex API support team's competence.

We also want to mention that we are aware of different problems with JForex API and working to make it better.

Hello,

I must agree that it would be very nice to use FIX as long as the API is almost unuseable right now.
Just look at all the bug reports I've send:
1
2
3
4

These are all somehow related when using threads.
There must be a more stable API, since more experienced users need them urgently.
Even if those errors are on my side, then create an API which is hard to misuse(Scott Meyers prays on this).
Why not opening FIX as an option?

Also, sometimes the API-Team forgets about issues or ignores them, see this as an example.
Support promised to update the javadocs, but so far nothing happened(at least I cannot see it).
The next user asking this question will run into the same problem; it would take 2(!) minutes to update the docs...

So, please try to improve on these matters

Thanks,
Juergen