Next Familiar release...

Things have come a long way since familiar .4. At Usenix a number of us
chatted about the next release, while hacking massively all week.

Our general feeling was that we have enough more working that we should
try to package up a release as soon as possible, rather than to continue
to add new goodies (to the base system).

Familiar unstable has come a long way in the recent past, with much improved
package management, a new window manager (ion) about to replace blackbox
that also removes the need for c++ libraries in the base load, replacement
of the remaining utilities that required Xt, Xaw, Xmu; and Alex has a
much smaller build of python. RandR is working and stable.

Together this should free up significant space, when we finish
the packaging/dependecy work. And the new menu system allows one click
installation of applications over the net all the way to a usable state.
So we're mostly focussing on packaging and cleanup now, rather than new
code.

This implies that folks working on applications should try to get them
packaged up sooner rather than later, so we can "cook" a system a bit
after we think we are done (it always takes longer than we want).

Work on sysset to get system configuration under control for some of the
software we have in hand would also be grealy appreciated. For example,
Tom Lawler has packaged Samba (client and server): this needs some
configuration scripts to really be useful. Edward Mueller is pretty
busy with his new job, so help is really needed here.

There are a number of open issues that still make setting a schedule in concrete
(to the extent this is ever possible in an open souce project) a bit difficult.

These include:
        1) what kernel to use. Jamey's currently taking 2.4.6 for a spin: we'll
see. We'd like to get the fork used for the BackPAQ back into the mainstream.
2.4.4 had a problem that prevented using the HP VGA card we'd like to get
fixed. Hopefully we'll know in the next couple days if 2.4.6 is a good
candidate. Tracking the mainstream is a long term necessity.
        2) Power management got broken inadvertently: we need to get this fixed.
We have a clue: turning off the screen is causing audio to stop playing.
Clearly the screen power off code is messing with registers it isn't supposed
to; this may or may not be part of the power problem.

For the moment, we'll likely punt putting in users, though this will be
at the top of the list for .6. Running everything as root is not a good
way to go long term.

Comments?
                                        - Jim

--
Jim Gettys
Technology and Corporate Development
Compaq Computer Corporation
jg_at_pa.dec.com