The future of L4
guillaume.fortaine at wanadoo.fr
Wed Aug 2 18:23:43 CEST 2006
I believe that the Mozart-Oz language could be a marvellous start point
: object-oriented, concurrency, logic verification, security (planned)...
What do we need now ?
A good kernel design : security, asynchronous IPC, threading, real time
A compiler : parser and optimization ( i think a sort of Elsa &
Mozart-oz front-end for llvm )
A driver development environment : I believe the most time consuming part
Forget Glibc .... or a poor fork of this => a new library design will be
a good challenge too ;-) !
Powerful tools to reengineer legacy programs :
1 year of work and 100 very good programmers at least, and I think we
will be able to enjoy it ... ;-) !
I am contacting some of the hall of fame, preparing my weapons, no
problem. Makus & Neal were too weak because too few, but thanks to their
effort, they show us the good way : we can revolutionize the world :
linux was only a start point, it's time to port open source philosophy
and community to a "military grade" quality ;-) !
Nowadays, is GNU really GNU is not Unix ( I would say GNU is only Linux
... ). Open source life behind a legacy OS ( 30 years ) design : Unix (
& C language too...)
Hey guys, 15 years of development and I send you a mail with a LiveCd :
very impressive, all my hardware is detected ( sound & video ok) =>
better than M$ ...
But the security and the " raw" performance are indeed disappointing (
debian.org hacked 2 times in 2 years => 2000 developers behind ... )
kernel.org shut down one week ( ext3 ... )
Why an asynchronous IPC and a massively multithread model ?
Asynchronous, like in real life : each one takes it times ;-) !
Massively multithread : performance, of course :-) !
I believe that security could be "easily" implemented with a good logic
framework, we have marvellous tools to do that.
More information about the l4-hackers