libxml v1 dependencies (was: Re: multilib fun - devel packages)

[Jeff Spaleta]

On 12/15/05, Jesse Keating <jkeating at j2solutions.net> wrote:
> I'm not suggesting we move it as a target issue, I'm suggesting we move
> it as a legacy software w/ no good replacement.

Are we sure there's isn't a good replacement?  I think thats an open question.
grisbi in Extras might be a reasonable replacement.

> But by then, we may be to
> the stage of Fedora Linux being pushed as a disk set made up of core
> +extras.

Here's me.. holding my breath. There are much larger issues with
moving to a single cd Core than whether or not a single financial
application is in Core.
Until there is a clear signal that the definition of Core has changed
direction I'm not going to be working under the single CD goal
assumption.  Someone is going to have the balls to move KDE and some
of that java crap over to Extras if you have a chance in hell of going
to one CD.  I'll pay more attention to this argument as a serious
statement of policy direction once I see KDE move over.

And I think the self-hosting restriction on whatever that 1 -cd "Core"
is going to be too tight a constraint and ultimately kill the goal of
a single cd that is in anyway usable to a mere mortal by itself.   I
think the 1 cd goal will require a competely different restructuring,
and the brickwall between Core and Extras buildsystem is going to be a
big problem that is going to get in the way of that re-structuring. 
Quite frankly i think everyone will have dvd burner drives by the time
Core is ready to make the infrstructure and policy jump to a single
cd.... which would be satisifyingly ironic.

> We really need to combat the
> Extras is Second Class stigma going forward.

You don't do that by pulling all functionality out of Core just to get
down to a single cd. Core has to be designed with some usage pattern
in mind,  If Core usage case is going to be condensed to "web
browsing" fine, lets state that and pick the applications based on
that. and we live with the size of the distro that results.  As it
stands I'm working under the assumption of "knowledge worker" as a
design goal for Core and I think that includes financial software more
strongly than say  a bittorrent client or the gimp or emacs*

-jef"and by balls I mean moist warm Schwetty balls"spaleta