|
PureBytes Links
Trading Reference Links
|
At 8:12 PM -0600 11/26/01, Mike Eggleston wrote:
>Is there an accepted or standard method of forcing all inputs
>to be the same (in the indicators where that's wanted)? Something like
>a single input at the top that is fed to a function that returns
>the rest of the common values for the indicator to function? Sounds
>like a lot of work to do it that way, though.
The problem is that all systems and indicators are separate programs
with no common storage. You can buy a "Global Variable" add-on that
will let you, for example, have all the inputs in only one program
(usually the trading "Signal") and have it store the values in global
variables that can be read by the other Indicators. It is harder than
it sounds, however, since you then have to make sure the programs are
run in the proper sequence.
Programs also run differently for historical testing than in real
time. For historical testing, each program is run from first to last
bar, then the next program is run from first to last bar, etc. In
real-time all programs are run in some unknown sequence as each
real-time bar is completed. Needless to say, this greatly complicates
passing information between programs.
>Oh... so many questions.
>
>I understand that the password alone on ELA's is not sufficient and that
>the source protect of ELS is a byte-compiled version of the EL without
>the EL source as baggage.
The ELA format has the source present in the compressed file. The
password mechanism is very simplistic and not good protection. The
ELS password scheme seems equally simplistic but I don't know if it
has been cracked.
The "Export to ELS protected" format does not contain the source and
should be a lot more secure. Any scheme can be cracked but the effort
of cracking this one vs. the benefit probably makes it pretty safe.
>Now... If I can call a dll (DefineDLLFunc:) from the EL, will the
>code still work to call the external dll from the byte-compiled ELS?
The method of calling an external dll should work in any version.
>Has anyone published or reverse-engineered the ELS byte-code format?
>
>Does anyone have a favorite method of either encryption of the EL/ELS
>or a favorite method of authenticating a valid customer is using the
>ELS? I have learned of CustomerID available within TS6, but if the ELS
>were decompiled then the if statement testing for a valid CustomerID
>could be removed. Frankly anything can be done to the system once
>decompiled.
Decompiling an ELS Protected file is possible but more like a
disassembly process. It is obviously possible but making sense out of
the resulting stuff would be very difficult and time consuming.
>The setup... Well I did mention that I was working with a trader
>to convert his system into something sellable, but he's really
>worried about 'non-authorized' use of the system.
Tell him not to worry. In all likelihood, the system will not be
worth stealing. And if someone steals it, they most likely would not
know what to do with it. I have hundreds of trading systems on my
machines that I got from various lists but have never had the time to
study them, to see if they work or how they work. I am sure there are
a few real gems in there but it is a needle-in-a-haystack problem.
Selling systems is more about marketing and support than about the
quality of the system. If someone really wants to use a trading
system to make money, he or she will gladly pay the trivial price to
get the advice and support of the author. One mistake in using the
system would cost far more than the system sells for.
Bob Fulks
|