Reduce duplicated applications and libraries in Hardy

Registered by Martin Pitt on 2007-10-18

Hardy Heron will be an LTS release again, and thus we should spend some extra efforts to get rid of duplication to make maintenance as easy as possible and avoid CD space waste.

We identify lists of duplicated libraries (such as libdb4.[23456]) and outline the steps to eliminate them. We also present discussion results about applications which have a similar purpose (such as f-spot, gthumb, and eog) and thus both create unnecessary maintenance overhead and don't conform with our "ship one default application for one purpose" philosophy.

Blueprint information

Colin Watson
Martin Pitt
Needs approval
Martin Pitt
Pending Approval
Series goal:
Accepted for hardy
Milestone target:
Started by
Martin Pitt on 2008-01-04
Completed by
Martin Pitt on 2008-04-18


pitti, 2007-11-15: Set to review; Matthias, can you please have a look and bounce back to me or set to pendingapproval?

I actually wrote about something similar to this in the IdeaPool the day after release. If you're interested in hearing my perspective as a user, follow the link below:
-- JohnKarahalis

pitti: John, we did not get around discussing the "duplicated apps" bits of this (see the "Outstanding issues" section). Keeping it in the idea pool is fine for now, thanks for bringing it up.

pitti, 2007-11-22: Matthias confirmed that he is ok with it.

jwendell, 2008-01-26: With Vinagre in desktop suite, we can drop vncviewer. -- Pitti: that's already done

pitti, 2008-02-13: Tim is reportedly working on dropping the gcc-4.1 kernel build dep on hppa/powerpc; we did everything else that doesn't involve heavy upstream work

I think that in the case of gthumb vs f-spot, we should prefer gthumb because of its batch processing features that are missing to f-spot. Also gthumb is not mono dependent and, that's a personal non technical point of view, I think that's not a good idea to have mono installed by default with Ubuntu.

pitti, 2008-04-18: we managed to remove most of the library duplication mentioned in the spec. The few remaining issues are (1) depending on heavy upstream work and (2) are not too painful to maintain in Hardy, thus this can be considered sufficiently implemented for Hardy IMHO.


Work Items