Are you interested to participate to a public team project using the new mql5 "Shared Projects" ? - page 2

Alain Verleyen
42596
Fernando Carreiro:
The reason I mentioned it was to demonstrate that either:
... not many people are aware of it, or
... there is not much interest in shared projects.

:-)

6 people interested on 8 votes (+me). That's a good start.

Alain Verleyen
42596

So I will ask to people interested either to contact me by PM or the write a post here.

Also the people who have idea can either post here publicly or by PM if they prefer.


I have a lot of ideas, but here is one. People who are interests in Neural Network, Deep learning or similar things are using Python libraries. The idea would be to start a small EA project, but using mql5 libraries, so they would need to be ported from Python, that part is not a small project though. Probably we would need to something smaller to start.

All ideas are welcome.

Fernando Carreiro
6949
Alain Verleyen: 6 people interested on 8 votes (+me). That's a good start.

However, I do see a major problem for which I don't think MetaQuotes will do anything to impose it!

A shared project, by its nature, should be considered "Open Source" and for private use only, not for commercial use. The problem, will be that there will always be some "smart alec" who will repackage it and submit it as their own product for the Market.

How will MetaQuotes handle that, or how to we impose control on that?

Fernando Carreiro
6949
Alain Verleyen:

So I will ask to people interested either to contact me by PM or the write a post here.

Also the people who have idea can either post here publicly or by PM if they prefer.


I have a lot of ideas, but here is one. People who are interests in Neural Network, Deep learning or similar things are using Python libraries. The idea would be to start a small EA project, but using mql5 libraries, so they would need to be ported from Python, that part is not a small project though. Probably we would need to something smaller to start.

All ideas are welcome.

Actually, I was thinking about something along those lines too - Deep learning, in particular! But we could however start with the C++ libraries of TensorFlow instead of Python!

However, the lack of multi-threading in MQL5 (or MQL4) bothers me! When is the "Service" thing going to come out for MQL5?

Alain Verleyen
42596
Fernando Carreiro:

However, I do see a major problem for which I don't think MetaQuotes will do anything to impose it!

A shared project, by its nature, should be considered "Open Source" and for private use only, not for commercial use. The problem, will be that there will always be some "smart alec" who will repackage it and submit it as their own product for the Market.

How will MetaQuotes handle that, or how to we impose control on that?

You will never avoid such possibility. A project can be private but with trustworthy people participating. For sure a shared projects will certainly not be public if it's about a profitable EA.

Anyway, it's not the time to worry about that. Let's first see the ideas and the trustworthy members.

Alain Verleyen
42596
Fernando Carreiro:

Actually, I was thinking about something along those lines too - Deep learning, in particular! But we could however start with the C++ libraries of TensorFlow instead of Python!

Yes why not.

However, the lack of multi-threading in MQL5 (or MQL4) bothers me! When is the "Service" thing going to come out for MQL5?

No idea sorry. It's "planned", but I didn't read any more information.
Alain Verleyen
42596

By the way there will be soon new communication tools for team group. I don't have much more information though.

Fernando Carreiro
6949
Alain Verleyen: Yes why not.

I don't think we can implement proper AI, Deep Learning, purely in MQL5 without multi-threading. It would have to rely on external implementations of major parts, such as running as a Windows Service and then provide Inter-process communication between the two.

It may also be too big a project to handle! Maybe we should start with something smaller!

Alain Verleyen
42596
Fernando Carreiro:

I don't think we can implement proper AI, Deep Learning, purely in MQL5 without multi-threading. It would have to rely on external implementations of major parts, such as running as a Windows Service and then provide Inter-process communication between the two.

Who said there is no multi-treading possible from MQL5 ? Anyway you are right it's something to think about seriously.

It may also be too big a project to handle! Maybe we should start with something smaller!

I do agree. An idea ? :-D

Fernando Carreiro
6949
Alain Verleyen: Who said there is no multi-treading possible from MQL5 ? Anyway you are right it's something to think about seriously.

Yes, I know there are ways around it, but is not a very clean, elegant way of doing it!