Re: Discussing issues

On Tue, Apr 15, 2003 at 12:23:59PM -0400, David Dawes wrote: 
> Rather than us trying to guess what you mean by "platform", it
> might help if you provided us with a more detailed description of
> what you have in mind.
> 
> I suspect that it is differently focused than what XFree86 has
> usually considered its domain, but I can only guess unless you
> explain what you mean in more detail.
> 
> XFree86's domain has traditionally been the X infrastructure, which
> does include all of what you mention above, and implemented on a
> wide range of platforms.  I'm trying to understand what else you'd
> like to push down into "X", what parts you'd like to eliminate,
> and how that fits in with the needs of all XFree86 users (not just
> those using it on the popular platform of the day).

Most of the stuff Keith is working on falls into this category:
RENDER, Xr/Xc, fontconfig, Xft, Xcursor, Xrandr, XFIXES, and future
plans for double buffering, magnification/thumbnailing, and
alpha channel in the server.

Then the stuff from freedesktop.org and other ad hoc specification
writers: icon themes, desktop menus, XEMBED, XSETTINGS, MIME system,
EWMH, Xdnd, UTF8_STRING, etc.

Software from freedesktop.org: D-BUS, pkg-config, etc.

Then the stuff from X.org: STSF, MAS, etc.

Then there's some random stuff under other umbrellas: gstreamer, arts,
CSL for sound, for example.

Future scope is essentially: whenever having multiple implementations
of Foo in the toolkits/desktops causes significant user/developer
problems, ideally we refactor it so Foo is down one more layer in the
dependency chain:

  GNOME     KDE   Enlightenment   Java/Swing    OO.org, Mozilla, WINE...
    \       |           |               /
     \      |          /               /
       --------------------------------
                     |
                   X Platform                   <--- refactored stuff 
                                                     moves to live here

To get GNOME and KDE to refactor stuff, they need to have faith in an
organization there that's going to end up owning it.

As far as elimination, I don't see the need to delete anything from
existence, but I do want some way to "bless" something like the font
subsystem, which simply means the key core developers feel like they
agree on it and it was selected via a fair process.

Provided we keep back compat, which is essential, I don't think this
impacts users who don't want to use the popular platform of the day.
If you want to use twm then great. You don't have to use the
complicated desktops. And in practice people have been doing things 
like extending xterm to use Xft. That's great too.

Or does "popular platform" mean Linux vs. BSD? I don't think this has
any impact on that whatsoever, GNOME, KDE, etc. are portable all the
way up to the top,
e.g. http://www.gnomedesktop.com/article.php?sid=1053&mode=thread&order=0&thold=1
That isn't going to change.

Havoc