Where now?

Laurent Martelli martelli@iie.cnam.fr
29 Dec 1998 01:31:12 +0100


>>>>> "Matt" == Matt Miller <idiot@slack.net> writes:

    Matt> Now that you have stated the goals, where does work
    Matt> commence?

    Matt> At what level do you want to start building?  Prototyping
    Matt> something on top of a current system?  Low level kernel
    Matt> work?  I know all of these have "started", but nothing seems
    Matt> to be moving...  where can i apply my time to have the most
    Matt> impact?

I suggest that we concentrate our work on Hihl Level Abtstractions
implemented on top of an existing system (unix/posix). For the
following reasons : 

  1 - The mainstream hardware architecture may change before we end
  coding low level kernel stuff, :-)

  2 - We will avoid taking into consideration low level constraints in
  the design of the system, which must hardware independent. 

  3 - Developpement speed : It will enable us to run the system on a
  wide variety of hardware with very few work. I guess that a pair of
  very gifted hackers could develop a kernel in a few months, but
  low-level drivers would pump all our stamina.

  4 - Low level stuff is not uninteresting, however I feel there's not
  much to add to the known techniques. There's much more to be
  improved in global high level design than in low-level
  architecture. 

  5 - Developpement speed : I guess that a pair of very gifted hackers
  could develop a kernel

Anyway, I have to set me up a CVS tree, and look at what you've done
already.

Laurent.