It seems to me there may be many other good ways to implement GoldSpike. 
Whether it's the configuration format, the way rails maps to a servlet 
engine (or no servlet engine at all, as in the Grizzly Rails module), or 
how it retrieves, allocates, and manages resources (including JRuby 
instances), there's a lot of variability here.

It also seems to me that GoldSpike has become somewhat set in its ways. 
It has had a 1.0 release already (somewhat prematurely, in my opinion) 
and although it seems to be working really well, a 1.0 release always 
seems to kill new innovations.

I propose a friendly fork of GoldSpike in the same repository, to allow 
more drastic changes to be experimented with, and potentially full 
rewrites of key pieces to happen. This would continue to live under the 
jruby-extras umbrella, and would be regarded as a research project. 
There's always the possibility the changes would be excellent, and the 
new GoldSpike would become *the* GoldSpike, but at the very least it 
seems like we need more diversity in this area.

Remember folks, the OSS rule is that your implementation is worse than 
the one coming tomorrow, and more diversity is a net gain for the 
project as a whole. We need to foster more diversity in the area of 
rails-in-a-WAR-file.

So who would be interested in making the fork and running away with it? 
I know a few of you have had good ideas that maybe didn't fit well into 
the current GoldSpike design. If you have commit rights, I highly 
recommend you pull off a branch and try your ideas out, and invite 
others to join in or do the same.

- Charlie
_______________________________________________
Jruby-extras-devel mailing list
[email protected]
http://rubyforge.org/mailman/listinfo/jruby-extras-devel

Reply via email to