Ubuntu Wiki

• Created: Date(2005-12-13T12:45:06Z) by JaniMonoses

• Priority: NeedsPriority

• People: JaniMonoses, NeedsSecond

• Contributors: JaniMonoses

• [https://launchpad.net/distros/ubuntu/+spec/xubuntu-roadmap-dapper]

Summary

Xubuntu Roadmap for Dapper

Scope

This page covers the xubuntu-dapper tasks to accomplish in order to meet our main goal - getting ubuntu-express/install CDs out for Dapper.

Implementation

Artwork

• DONE: we have nice usplash, proposed wallpapers and gdm theme
• TODO: somebody qualified volunteering to oversee artwork direction
• TODO: upload/integrate all current work
• TODO: get more works/feedback

Documentation

• DONE: Julio has written a nice about-xubuntu ff start page which is packaged in xubuntu-docs (on REVU only)
• TODO: somebody qualified volunteering to oversee xubuntu doc writing
• TODO: get doc people to use ubuntu-docs svn and go on freely with writing xubuntu docs.
• TODO: get xubuntu-docs voted for in REVU so it can be uploaded.

Sync with debian unstable (ongoing task)

Current status: we already have reduced the diffs between us and debian with the 4.2.3.2 sync, but there are still packages which diverge. Most of it is noise though, not real diffs. http://utnubu.alioth.debian.org/scottish/by_maint/pkg-xfce-devel@lists.alioth.debian.org/

• TODO: sync newer exo and thunar when they are uploaded to debian.
• TODO: push upstream our packages which they don't have (notably some xfce-goodies)

I am in the process of becoming a pkg-xfce member on alioth so I can help out in their svn.

Promoting packages to main (ongoing task)

Xubuntu packages need to be in main so the CD build infrastructure works with them and so they can be supported in ubuntu-updates and ubuntu-security.

• DONE: main inclusion reports for core xfce packages written and enqueued, pending review (by pitti probably)
• TODO: respond and correct the issues raised during review
• TODO: sync exo from debian (see above) as having it come from there makes the review process more straightforward and less risky.

App selection

The key apps we need to be usable are listed here. For more details go to XubuntuProposedPackages

Display manager

• gdm seems like the best choice as it is used by ubuntu, has lots of features, looks ok and is actively maintained upstream. Xdm/Wdm/Slim/qingy were considered but dropped for various reasons - ugliness, lack of features, immaturity, lack of ubuntu integration and not least nobody advocating them actively enough. This is a task still open for experimenters who can convincingly argue for one of the alternative DMs. The main point against gdm is it not being lightweight enough but that seems to be changing

• TODO: see below - building it w/o gnome deps

Volume manager

• most likely ivman. It is already in main as it was used by kubuntu until 5.10.

• DONE: wait for new upstream version which includes specifying alternate config dir so we do not mess up current

setting in /etc. Upstream promised this week.

• DONE: ivman config files uploaded as part of xubuntu-default-settings.
• TODO: wait for my main upload privileges to take effect and I can upload new ivman itself.

Default file manager

• thunar.
• TODO: wait for it to come in through debian

Miscellaneous hacks

There are no nice networking/printer config apps for Xfce.

• TODO: evaluate writing them using system-tools-backend. Anyone interested?

Integrate with laptop hotkey-setup Fix logout dialog to not require password Integrate hwdb-client, update-manager, update-notifier (these all have various gnome deps)

Only packages without gnome dependencies in xubuntu-desktop

A contentious subject, but lacking more scientfic data to be able to differentiate between heavy and light apps one guideline is 'no gnome dependencies'. Obviously also no kde/mono/java deps either. Actually I have used exmap, so there is some scientific data but on the machine I am developing on (1,73GHz, 512M) using an additional 4-5 Meg or loading 29 additional .so objects does not seem to make a visible difference. It probably does on lesser hardware.

• TODO: gdm w/o gnome.

gdm uses very few gnome functions calls. I built a gdm w/o gnome deps and sent part of the changes upstream a week ago. No response yet, I will follow up shortly with the rest of the changes. (Dec 13 update: upstream took the patches and released them as part of 2.13.0.3, and seems to go in the same de-gnomifying direction, so luckily we won't have to have a forked gdm in xubuntu)

• TODO: evince w/o gnome.

Built and runs fine, needs packaging, upload to REVU.

• TODO: gnumeric w/o gnome

both libgoffice and gnumeric have --without-gnome config options (they have a win32 port) upstream debian packager said he would accept clean patches to make gnumeric-gtk binary debs too.

Xfce version to use

Initial goal was using 4.4 since it was supposed to come out by April. However Xfce upstream is divided whether that release will be stable enough to base a distro on by that time. Still an ongoing research

• 4.2 pros
  • already packaged/used/tested since Hoary
  • stable
  • translations very complete
• 4.4 pros:
  • really nice new features in the wm and the panel especially

Colaboration with Xfce upstream

Need to investigate whether ubuntu can offer anything besides user feedback to Xfce. Maybe rosetta could be pitched to their i18n team but I don't know much about rosetta's maturity. They use bugzilla and svn so LP is probably not tempting either, especially not at this stage. Once bzr is default we'll import their svn anyway for testing purposes.

Other specifications relevant to Xubuntu

Espresso, specifically the gnome ui https://launchpad.net/distros/ubuntu/+spec/ue-gnome-ui

Desktop specification https://launchpad.net/distros/ubuntu/+spec/dapper-desktop-plan (particularly regarding what to have in the menus, logout dialog icons instead of radio buttons)

Edubuntu - for a lighter version

Other things

Contact ubuntulite and see if we can colaborate better than we do now. They have done some good research on lightweight apps. We could have icewm in xubuntu too as an option, as it is small, so a nice default install of ubuntu + icewm may cover at least part of ubuntulite goals. Must see how to incorporate their more intrusive changes without diverging from ubuntu core, but probably this will come with the price that the system won't be as lightweight as theirs. Maybe we could have a set of utilities which turn a xubuntu into an ubuntu lite (i.e boot directly into X w/o a DM, remove some uneeded services)