Re: [JXTA discuss] JXTA Language Bindings

Shane Isbell wrote:
>  
> Hi all,
> 
> I took a look at the latest title for the jxta-ruby project (on java.net 
> <http://java.net>), which I believe Henry has changed to 'Ruby Binding 
> for JXTA'. Because of the nature of the project, there may be some 
> confusion. My original intent was to use jxta-ruby as the basis for 
> another mobile project. By focusing on constrained environments, my 
> intention was not to get too bogged down in getting all the JXTA 
> functionality working in Ruby, just what is needed for devices.
> 
>  
> 
> I think that this may not have been the best approach because I am 
> seeing three different projects emerge. 1) JXTA language binding 
> project, 2) JXTA-Ruby and 3) JXTA-Ruby for Constrained Environments, 
> which should be a subproject of JXTA-Ruby. What I would like to do is to 
> step back and re-approach the problem.

I am not sure I understand you correctly, here is what I have in mind:

Produce different language bindings based on JXTA-C: In JXTA, language 
binding seems to imply to a native implementation of JXTA protocol with 
the language. The way I think of Jxta-Ruby project is to do just that. 
Jxta-Ruby is going to do that for Ruby, if possible, I would like to see 
Python, Perl, C++(may not be necessary if C API is good enough to C++ 
developers) and C#.

Should those project be subproject of JXTA-C? I don't know. As each 
project may have it's own pace, I tend to have separate projects depends 
on JXTA-C.

> With tools like SWIG, we can 
> create a general binding that could be used for multiple languages. I 
> would like to break the bindings off into a separate project, say JXTA 
> Language Bindings. We could then generate the infrastructure for groups 
> that want to build up the application in any of a dozen languages, from 
> C# to Perl to Ruby. After that, it makes sense to work on JXTA-Ruby 
> using the underlying bindings. I did take a look at SWIG-1.3 and got it 
> working on some test files so a JXTA language bindings project would be 
> feasible. Thoughts?
> 

This is what I really like to see happening. The goal I have for 
Jxta-Ruby project is to provide the binding, if not SWIG, do it Ruby's 
way is good too.

If we go for SWIG, maybe we can use JXTA-C as the umbrella project.

Cheers,
Henry