[kaffe] Re: Future directions for Kaffe

[Riccardo]

Hi Jim,

yes, it has been quiet here lately. To be honest though, I like how kaffe  
is (I still perceive it more "open" as OpenJDK) and I like our old-style  
site and CVS. I use both sourceforge and savannah, but our own stuff has  
more charm.

Personally I am pretty busy and I am not so expert to delve in the innards  
of kaffe. With somebody as a guide though I can accomplish some tasks.  
Currently with Guilhem I have pending some AWT event fixes and more  
importantly PPC jit.

Before releasing, Swing and GTK peers from classpath needs to be fixed  
absolutely!

For a 1.1.8 release, which is due since ages, I'd just like to see a "fix  
it at best" of classpath so that swing/gtk peer runs, a check that  
interprerter and jits are compiling an working on most of "our" platforms  
and maybe the fixes with our xawt if we still have time.

For the future I see following points:
- using a premade classpath seems a decent idea, if it alleviates other  
problems
- I would not spin everything off into pieces. It would just rot, makes  
things more difficult to install and bring more instability. Maybe just  
using a ready-built classpath is enough for te beginning
- we are still meaningful for some platforms like 68k, PPC, BSDs...
- as Tim wrote there is no point in being just a "slow" and "buggy"  
OpenJDK clone. Although we need compatibility, diversity is interesting
- slowly we may improve anyway

as far as the site goes, I like its spartan look, I like our CVS and our  
autarchy and thank you for keeping everything up. Still it clealry needs  
updates! If we get 1.1.8 out maybe there is also an incentive in writing  
things up.
I vote that if a wiki can be integrated on our server, it would be a nice  
thing to do (although due to spam it needs accounts probably). Some stuff  
like Hall of Fame, Hall of Shame, compatibility lists are best kept on  
there probably and generally help to keep pages were information is  
generated from the community up to date.

I'd prefer not jumping into the future, but morphing gradually. We cannot  
afford to big breakage an lots of refactoring is not fun, also it gives us  
time to reflect on what to do - we are few and busy.

I'll try to keep an eye and help people out.

Have a nice day,

   Riccardo