Charlie Kravetz wrote:
On Sat, 5 Dec 2009 09:05:45 +0100
Steve Dodier <email@example.com> wrote:
2009/12/5 Pasi Lallinaho <firstname.lastname@example.org>
J. Anthony Limon wrote:
I just thought I would open the door for some brainstorming in the area
of Lucid and beyond! I have some thoughts I'd like to extend everyone's
Do we really need it? Who really uses it? How stable is it anyways?
I feel gnome-app-install does more harm than good in the XFce desktop.
Firstly it does a poor job of representing the total software in the
repositories. Secondly, we almost *always* send people to Synaptic or
apt-get to install software. Thirdly, I've found it to be HORRIDLY
On my system I've ---purge autoremove'd it. A nice side effect was that
my XFce menu looks a lot nicer without that wide entry at the top.
I agree you on this. I don't really know gnome-app-install since I
always use apt-get or Synaptic myself. Maybe we should just seed
Synaptic as the default application/repository manager in Lucid?
Both Synaptic and gnome-app-install are being replaced by the Software
Centre (not sure if it's the exact name). This new app brings a few
dependencies but it's likely that removing gnome-app-install and synaptic
will make enough room on the CD for it.
As long as it doesn't pull mono and gnome* I'm all for giving the new app a
Maybe we could ask the desktop team what they think will be ready for Lucid
in the software centre, and whether they think they'll be able to replace
synaptic in this release.
I know this app has some serious features that alternatives do not, but
is consistently a source of problems and bugs, primarily in the area of
super high CPU usage and memory leaks, ironic given the nature of the
On my system I use a mixture of xfce4-taskmanager and htop, I'm not sure
if this would be satisfactory on the majority of people's desktops but I
am of the opinion that GSM has to go.
For now, I don't think the Xfce components can deliver the same amount
of features and, regretfully, quality. I also like htop, but we can't
consider it as the main application for system monitoring, as it's CLI
and many people fear command line.
Gnome system monitor monitors system load, network load, ram and swap usage,
and HDD usage. It may be doing too much for one's needs, but when you want
to know if some app is using all of your bandwidth, it's cool to can check
in the system monitor without having to go in command line.
While xftaskmanager may be more appropriate for your needs,
gnome-system-monitor is in my opinion better for end users.
Is the plan to stick with Totem for Lucid? It's kind of stagnant issue
but it's also a difficult one to address with the next release being LTS.
Agree. There is loads of *decent* video player alternatives. I've never
liked Totem. It sounds it is from the stone-age. I'd really like to see
something else already in Lucid.
I can name only one player that also uses a decent backend and that is
written with a proper GTK+ GUI. It's Parole, and I'm looking forward to it,
but considering that it's rather new, we can't expect it to be as integrated
in the desktop yet (for instance, does it already manage to find missing
codecs for the user?).
Whats the point of a player with tons of features like audacity, mplayer or
vlc if it crashes miserably when you launch a file or if the GUI is
difficult to use because of some particular skin, or very debatable keyboard
shortcut choices? I'm all for keeping Totem for the LTS, and testing Parole
from the very beginning of Lucid+1's release cycle (ie. before alpha 1, and
until beta 1 at least, so we can report bugs to Ali and see what's missing
from the Xubuntu point of view).
This seems to be an issue entirely out of anyone's hands unless they
want to try making one using xfce libs.
This is about all I can think of right now, but I do know I am missing a
couple things which I will bring up at another time. I feel this is a
good start to a brainstorming. Also, nobody has any intentions of
adopting Pulse Audio into the Xubuntu system, right?
My personal experience is that PA is only bringing in problems, but if
we can get those sorted out, I can live with it. I hear PA can do
wonderful things once it works.
We probably want to ship Exaile as our default media player for Lucid
also, but I want to finger at the really bad quality of media players in
general in Karmic. Most of them do not work for me at all (read: they
crash constantly or leak into memory).
I was also disappointed by some last-minute performance regressions in
Exaile. And the disappearance of some gstreamer codecs in the 64 bit version
didn't help (yeah, now Canonical sells them for real money... there is no
explanation over why they aren't packaged and free of charge anymore).
Exaile 0.3.1 uses GIO, and has a few mem usage optimisations. It also
contains some of the missing Exaile plugins/features, and it should be ready
for Lucid. I didn't look at the progress done by other players developers
recently (except Parole, but as I said back then I think music players
should have a music-oriented GUI, which eliminates media players in my point
of view). If there are things that you consider wrong in Exaile, it's the
time to tell them, because it was clear to me that the switch to Exaile
would be done in two parts, and that the goal was a seamless music player in
the LTS. I'm still available for working on Exaile.
Concerning PulseAudio, I think Lucid may be the right time for a switch. In
fact, it looks like PA and Wine work fine together, and Wine's new sound
architecture should be much more friendly with PulseAudio (I don't know when
it'll be released as stable though). PA solved a number of recurrent
problems I had, but it also introduced new ones.
The first problem is that, whatever people say, PA is NOT compatible with
the Gstreamer mixer plugin. This results in sound being muted after login,
because xfce4-volumed does something perfectly valid and normal, it fetches
the track's volume, and somehow it seems that PA doesn't like that and mutes
the volume instead. Switching to PA means writing a PA backend to
xfce4-mixer and xfce4-volumed. Trying to write such a backend (and a new
frontend) is very high in my To Do list, but I can't tell when I'll be free
from the school project I work on at the moment.
I would also like to say that latest releases of xfce4-power-manager are
getting more stable, and that it now works rather fine for me. But if you
experience bugs with it, please report them directly to bugzilla.xfce.org,
Ali (the upstream dev) is responsive and will very likely look quickly into
them. It's important to have it work seamlessly in Lucid, since it's a LTS.
But I'm quite happy of the switch to xfpm, I must say.
One last thing: do we still want to look into a replacement for Firefox?
That'd require some testing, to draw a features table to tell what's really
badly missing in Firefox's competitors, so we can see if it can be replaced
in the LiveCD. In my opinion the best competitors are Midori and Chromium,
and I have a soft spot for Chromium.
I do NOT want to look for a firefox replacement and the issues it will
bring into an LTS release. That belongs in the regular release, perhaps
lucid +1. Lucid as an LTS needs to be as solid as we can make it. It is
not the release to test what we can in, but rather, the release to fix
what we can in.
I have to agree with Charlie here. Changing the default browser to
something not Firefox in an LTS release would really make our users
mad, even if it was working. And at this time, I'm not sure if midori
is even working fairly enough.
Xubuntu Marketing Lead
Web-designer, graphic artist
IRC: knome @ freenode
xubuntu-devel mailing list