Future direction and vision of the J2V8 project

[aschrijver]

This issue is a continuation of the discussion that was started here: #201

I have included a small summary of relevant comments and my response:

@drywolf in #201 (comment) - I am also a big advocate of separation of concerns (talking about the class-sharing parts)

[...] I would say that J2V8 should stay focused on delivering only the low-level primitives to communicate between the JVM and Node.js/V8 and everything else that provides higher-level API / programming features should be kept as separate projects.

@mizumi in #201 (comment) - I agree that a separation of concerns - both within this pull request and without - is important. We really don't want to bloat J2V8 with unnecessary stuff.

@irbull in #201 (comment) - Maybe we should start to consider how we can structure J2V8 in terms of a core, and a set of libraries.

And from #328 :

@irbull in #328 (comment) - We considered (and sometimes still do) moving J2V8 to the Eclipse Foundation. That's the reason for the EPL. I also just used the standard copyright header that was used in other Eclipse projects for the same reason. I am happy to re-evaluate that, but I will need to get some outside council as I am not a lawyer.

Yesss!! I am 100% for this direction.

To be frank I was worried on the state and direction of J2V8, and the PR's hanging around for so long (mentioned in the 'cons' section of my evaluation). So having as many potential useful features merged was a good thing. But J2V8 would be so much more helped with a razor-sharp focus and purpose!

In terms of good project structure and community approach I have an excellent case study for J2V8 and that is Vert.x.

Vert.x also made the move to the Eclipse Foundation! ... and in such way that they reaped most of the benefits without most of the strictures. Only vertx-core is under strict EF control, the rest is freedom.

Please read this issue I created for a similar purpose for the Dat Project:

• Vert.x: A case study of successful project reorganization

I intend to use Dat technology and are altogether more worried about their development approach than I am with J2V8. They practice what you could call 'programmer anarchy' and are not yet that interested in restructuring.

The Vert.x way would fit very well though on a Java project like J2V8, much more than in node world with all the micro-module development going on.

One important thing to add is that - looking at the vertx.io landing page - if you had such good design it would be an excellent location to more prominently attract attention to commercial products and professional services on top of J2V8!

[irbull]

Regarding the chat, I've opened #331. I want to go through the rest of your discussion and I'll comment here.