Skip to content

Spread discovery and backtester code

10 messages · mail at chrisbird.com, Whit Armstrong, Daniel Cegiełka +3 more

Original
Jeff Ryan
# 
I'll be the first.

Did you just spam R-sig-finance with closed source code, distributed
as a binary, that USES the community's hard work, without giving back
anything?  That requires a password???!

This is a (very) likely a violation of the GPL in letter and, and
certainly in spirit.

As someone who works rather hard for this community, I will be the
first to request you take your efforts elsewhere. Not a single person
here has 'time' to look at it.

Jeff
On Tue, May 22, 2012 at 2:40 AM, mail at chrisbird.com <mail at chrisbird.com> wrote:

  
    

      
      
    

  


  


      

    

    

      
      

        
      

    

    

      
      

        
      

    

    

      
      

        


  

        

      Whit Armstrong
    

    

      
      #
    

  


  

      
Why don't you throw this up on github.

You'll get more respect from the community that way.

Most people will be reluctant to run code that they cannot inspect for
themselves (myself included).

Also, nice move on the tslib submodules!

-Whit

        
On Tue, May 22, 2012 at 8:46 AM, mail at chrisbird.com <mail at chrisbird.com> wrote:

            

      
      
      
    

        

      
      
    

            

      
      
      
    

        

      
      
    

  
  


  


      

    

    

      
      

        
      

    

    

      
      

        


  

        

      Joshua Ulrich
    

    

      
      #
    

  


  

      
Wearing my devil's advocate hat...

        
On Tue, May 22, 2012 at 7:34 AM, mail at chrisbird.com <mail at chrisbird.com> wrote:

        
<snip>

            

      
      
      
    

        
I hope you can understand how some people may be reluctant to make
this assumption. "Chris told me to ignore the license and assume it is
GPL" probably isn't a strong legal argument.  It could be different
now that you've said it in a public forum, but I'm not a lawyer.

            

      
      
      
    

        
While Jeff's response may have seemed inappropriate, your post was
highly unorthodox and it wasn't immediately clear you were trying to
contribute to the open source community.  The "requires password" and
the ".zip" (binary) extension were red flags.  I hope you can see how
this could give people the wrong impression.

If you're looking for feedback on your code, you really need to supply
the code, not just the program/package.  github, R-Forge, or
code.google.com are good for this.

I could only find one other instance where you thought someone
responded to you "in a manner which deters people from making a
contribution" and it seemed like you were very quick to assume that
Daniel was being rude when he was just being brief (and English isn't
his native language).

            

      
      
      
    

        
Best,
--
Joshua Ulrich  |  FOSS Trading: www.fosstrading.com

            

      
      
      
    

        

      
      
    

            

      
      
      
    

        

      
      
    

  
  


  


      

    

    

      
      

        
      

    

    

      
      

        


  

        

      Me
    

    

      
      #
    

  


  

      
I have not looked at chris's code in detail, but I downloaded the .zip and
there was no password protection and the R package it contains  consists of
a dll (his source i would presume, but not shared here) which has been
wrapped up in R functions.

Lets assume for a moment that the sourcecode on which the dll is based is
completely Chris's brainchild and work. In that case, wrapping it up in R
and making it available here, or anwhere else, is certainly NOT a violation
of GPL, because the dll is (assumption) not a *derivative* work of a GPL
product (xts, other packages, C source that is GPL, etc...) but his own.  In
that sense, i don't understand the rather frosty reception.

However, as has also been pointed out, distributing binary version of
something is not going to earn anyone any kudos in an open source community.
as simple as that.

Now if my assumption is wrong and the dll is indeed a *derivative* work of
something that itself is licensed under GPL, then and only then this whole
thing would be problematic and violate GPL. But the easiest fix would be to
include the dll's source in the package and voila! you are good to go.

The author would know best ...

-----
http://censix.com
--
View this message in context: http://r.789695.n4.nabble.com/Spread-discovery-and-backtester-code-tp4630831p4630919.html
Sent from the Rmetrics mailing list archive at Nabble.com.

  
  


  


      

    

    

      
      

        


  

        

      Jeff Ryan
    

    

      
      #
    

  


  

      
My apologies for starting this.

I didn't mean to imply that Chris was intending to do anything wrong.  In
fact, from a technical GPL perspective, I don't really care too much.  My
focus, and my immediate reaction was for one reason only, discourage
future use of the list to distribute opaque code.  Binaries, passwords,
etc are that - no question about it.

Chris has done a great job in responding to my somewhat less than friendly
initial reply, in terms of passing along more details as well as
explaining his position, and fixing some of the issues that weren't so
clear at the start.

The point of this list, community, et al, is open source finance, within
an R context.  It seems (now) that that is exactly what Chris is doing.
For that, I want to chime in with the others and say thanks for making the
effort in sharing your code and ideas here, and I think many would be
willing to help in the process of getting it into more hands and in front
of more eyeballs.

Jeff

        
On 5/22/12 10:47 AM, "soren wilkening" <me at censix.com> wrote: