seed changes

[Eero Tamminen]

Hi,

On Tuesday 05 February 2008, Jim Campbell wrote:
> What can we do?  Can we create test cases for possible app inclusion? 

Yes, there should be an acceptance criteria for an application.
Below would be my order of prioritization.

Basic things that need to work:
- Stability, does it crash or can it corrupt users data?
- Basic functionality and ease-of-use
  - Desktop integration, printing etc work
- How well maintained and regularly released in upstream
  - Do they respond to bug reports?
- If changes are needed, is there a willing and *commited* maintainer
  in Xubuntu for those (not somebody who resigns if there's a bit of heated
  discussion about distro goals)
- Localization coverage

Then secondary things:
- Resource usage
- Startup speed
- Functionality compared to other similar apps
- General responsiveness
- Battery usage (some panel applets, such as gnome-power-manager
  were really bad at this, they were waking the CPU several times
  a second)
- Network usage (for remote setups)

On low spec system resource usage is more important than
speed.  If it's just slow, it will come up eventually.  If it takes more
memory than you have it will make your whole machine to crawl
even if it would come up with enough swap.  I would declare low
spec machine as something having 128MB of RAM (with less RAM
you should be using just a window manager, not a full desktop).


See my "gnome apps" mail on 1st of February on how to measure[1]
performance and resource usage.

I think these can be measured even on Ubuntu, the packages should
be same.  Themeing can affect this quite a bit, but I think it's only
a problem with Kubuntu (it sets a Gtk theme that causes Gtk apps
to load *both* Gtk and Qt libraries!).


	- Eero

[1]  Performance & resource usage needs to measured, handwaving
about "gnome" and "feels slower" doesn't really cut it.  It will be
ridiculed by anybody who's seriously looked at performance.