harmony-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Stefano Mazzocchi <stef...@apache.org>
Subject Re: Harmony goals and priorities
Date Wed, 11 May 2005 23:57:00 GMT
Bob Griswold wrote:

> I am excited about this project, but I am also skeptical about its chances
> for success. This is not easy stuff, and as long as we have major vendors
> out there who are investing a great deal in duplicate work (Sun, BEA, IBM,
> HP, etc.), I doubt this project will do much more than be a cool academic
> exercise. We need to get the major vendors and investment (and their
> existing teams) behind this project, or a project like this, for this to
> stand much of a chance of success. I¹ll be excited if BEA opens up the
> JRockit source base and backs this project, or if IBM does the same with J9,
> or ideally, both.

Bob, thanks much for your comments.

As much as this sounds weird, I completely agree with you: without code, 
we are just a bunch people whining.

But let's look at history for a second: do you remember what happen to 
the mozilla codebase when they dumped it into the open? It got rewritten!

Was it that bad? maybe, but maybe not. The need for 'redesign' comes 
from "getting it right this time" (and people already expressed that 
overhere) but, most importantly, because it's very hard to understand 
somebody else's code, especially when built by big groups over several 
years and behind closed doors! (read: without an organic incremental 

Say BEA donated JRockit and IBM donated J9, both at the same time.

What would we do with it? how much likely it is that people here would 
understand it well enough or be fearless enough to get there and modify it?

I said this once before and got cited by many: "good ideas and bad code 
build communities, the other three combinations do not".

I'm not saying that if some of those vendors showed up, we would turned 
them down, but I have the impression (well, let's say a little more than 
that) that no big vendor will show up until we manage to convince them 
we can get some traction ourselves.


Well, let's glue the pieces together... if you have a (small and lean) 
JVM that you want to see it becoming the seed of Harmony, donate it to 
us, we'll give you commit access and we'll start from there.

Is it in java? even better, a lot more people will be able to help out.

Academic exercise? potentially, but also a way to make seeds grow. And 
no matter how small the seed, it doesn't really indicate how big, strong 
and long lasting the plant is going to be ;-)


View raw message