Wishes for MQL5 - page 12

 

um... But mysql <=> mql... has anyone thought of that? :)

 
Red.Line писал (а):

um... But mysql <=> mql... has anyone thought of that? :)


What's there to think about 2 completely different programs
 
And silence. Highlighted the question in red. Maybe it will be more noticeable.
 
Prival:
And silence. Highlighted the question in red. Maybe it will be more noticeable.

If I understood correctly, you want the EA's ex4 to be impossible to decompile, but still be able to run it anywhere. How do you imagine it?
Either I got you wrong or you have a wrong understanding of cryptography.
 
amirul:
Prival:
And silence. Putting a question in red. Maybe it will be more noticeable.

If I understood correctly, you want to make it impossible to decompile the Expert Advisor's ex4 but still be able to run it anywhere. How do you imagine it?
Either I got you wrong or you have a wrong understanding of cryptography.


No, you don't.

1. Let's say to enter a competition (or sell an ekpert), I build in restrictions on opening hours, account numbers etc. and it works. But decompiling and removing this protection from there was difficult (preferably not doable).

2. Intercepting passwords and account numbers is quite possible, the Trojan simply needs to be made special targeting it. And then to make a couple of tricks. The electronic key is needed to clearly confirm that the transaction is done on your behalf.

If you read it and find it important, just highlight the issue and do AP maybe developers and ponder, or only one me needs it?

 
Prival писал (а):

No, it's not a bit like that. Let's say for a contest, I build in restrictions on running time, account numbers etc. and it works. But decompiling and removing that protection from there was difficult (preferably not doable).

Restrictions can be inserted even now. After successful decompilation it is not difficult to remove the protection, hence we need to make it impossible to decompile. By decompiling I mean in this case, to restore the meaning of instructions in their binary representation (and not the restoration of the source code with all the variable names and comments). But the terminal itself needs to "understand" the meaning of written instructions. If any terminal can "decompile" them, anybody can.

The examples of DVD, HD-DVD and Blue-Ray are illustrative in this respect. Huge companies spent huge money to develop such protection so that encrypted data was stored on the disc and only certified players could decrypt and play it (of course, the certificate was issued to those who would not play counterfeit, etc.). All these protections at the moment are opened. They were opened simply because the players already had the code to decrypt the contents of the discs, the only thing needed was to understand HOW they did it. In my opinion, such "protection" as any other kind of security by obscurity is harmful, as it gives a False sense of security.
 
That's the thing that needs to be understood ("this is what needs to be protected"), I think it is possible. Suppose I built in my Expert Advisor logic variables related to checksums of eX4. At any change of checksum it starts trading incorrectly, and if I don't change it, it simply doesn't trade, because time ran out, for example. To understand the logic and idea of an Expert Advisor, at that if you get machine codes, IHMO, it is very difficult. The defences you mention were cracked because they perform the same operation. By understanding it, this protection was bypassed. The experts are always different and issue a command, even if binary, to buy and sell, at different times with different currencies, etc. Here, even in the open code you will not immediately understand everything, even if you have a comment.
 
Just for the record, not all of the protections are cracked, there are some that even theoretically cannot be broken. And there's strict mathematical proof of that.
 
Prival:
That's the thing that needs to be understood ("this is what needs to be protected"), I think it is possible. Suppose I built in my Expert Advisor logic variables related to checksums of eX4. At any change of checksum it starts trading incorrectly, and if I don't change it, it simply doesn't trade, because time ran out, say. To understand the logic and idea of EA, at that, if you get machine codes, IHMO is very difficult. The defences you mention were cracked because they perform the same operation. By understanding it, this protection was bypassed. The inspectors are always different and issue a command, albeit binary, to buy and sell, at different times with different currencies, etc. It's hard to understand everything in the open source code, even with comments.
Of course, compiling the EA to the level of machine code will increase its size and complicate the decompilation process, but it will still be possible. Another matter is if the MQL will have a messaging-signals exchange system, i.e. the Expert Advisor owner will not sell the Expert Advisor itself, but the signals to a certain circle of "subscribers". This option will require additional costs from the owner for providing signals to clients, but this is a different matter.
 
Prival:
That's the thing that needs to be understood ("this is what needs to be protected"), I think it is possible. Suppose I built in my Expert Advisor logic variables related to checksums of eX4. At any change of checksum it starts trading incorrectly, and if I don't change it, it simply doesn't trade, because time ran out, say. To understand the logic and idea of EA, at that, if you get machine codes, IHMO is very difficult. The defences you mention were cracked because they perform the same operation. By understanding it, this protection was bypassed. The inspectors are always different and issue a command, albeit binary, to buy and sell, at different times with different currencies, etc. In this case even in the open code you will not immediately understand everything, even if you have comments.

You mentioned "major interests". If there is enough will, people reconstruct the principle of operation even of operating systems - it would be understandable each particular instruction. I repeat: if the meaning of the instructions is clear to the public, it will be available to everyone.

As for unbreakable protections. Here it all depends on what you understand by "protection". Encryption is one thing when two parties share a secret and quite another when only one party has a secret. In the latter case, protection is impossible as a matter of principle.
Reason: