Re: Funding LuaJIT 2.1 and beyond
- From: Richard Hundt <richardhundt@xxxxxxxxx>
- To: luajit@xxxxxxxxxxxxx
- Date: Fri, 8 Jun 2012 02:38:38 +0200
On Jun 7, 2012, at 11:33 PM, Mike Pall wrote:
> This is a follow-up to the call for sponsors in the LuaJIT Roadmap
> 2012/2013:
>
> Any and all ideas for the continued funding of LuaJIT, and in
> particular the initial funding for the work on LuaJIT 2.1, are
> welcome!
I'm not sure how much capital is in the pot here and how hard it is to get at,
but from what I can tell, the European Union seems to support private
researchers as well:
http://cordis.europa.eu/eu-funding-guide/checklist01_de.html#03
I think that selling a GC is hard, sure, but what excites me about this project
is its potential as a platform for running dynamic languages, in the same way
that the JVM isn't really about Java anymore. Don't all start shooting now, I
really like Lua, but for me LuaJIT's VM is too special to only run Lua. The
only reason I can think of that a Ruby implementation on LuaJIT might not be a
couple of orders of magnitude faster than the reference implementation is the
GC.
These guys [1] built LLVM into a realtime stream mining framework, so that they
could run a home grown functional query language on it. The reason they gave
for choosing LLVM was that they wanted JIT compilation.
So given that I've invested a fair bit of my time in trying to leverage LuaJIT
in this way, if it came to crowd funding, I'd definitely be making donations
toward the GC, but perhaps there's also an angle in this kind of thinking which
a corporate sponsor would buy into.
Cheers,
Richard
[1] http://www.celeraone.com/
Other related posts:
