Xubuntu meeting summary

Eero Tamminen oak at helsinkinet.fi
Thu Mar 27 09:17:49 GMT 2008


Hi,

On Thursday 27 March 2008, Jim Campbell wrote:
> I'd first like to start off this e-mail by announcing the Xubuntu
> community meeting was a *huge* success. We had roughly two dozen people
> take part (including old, current, and new faces) and a number of other
> individuals who sent in e-mails or left a quick IRC message to let us
> know that they were unable to attend but would be following up with much
> interest. After just under an hour of constructive discussion led by Jono
> Bacon and several free form votes, I'm happy to present following mission
> statement for Xubuntu:
>
>     "To produce an easy to use distribution, based on Ubuntu, using Xfce
> as the graphical desktop, with a focus on integration, usability and
> performance, with a particular focus on low memory footprint. The
> integration in Xubuntu is at a configuration level, a toolkit level, and
> matching the underlying technology beneath the desktop in Ubuntu. Xubuntu
> will be built and developed as part of the wider Ubuntu community, based
> around the ideals and values of Ubuntu."

Level of integration and especially usability can often be subjective
matters, but sometimes it's very clear which of the alternatives is better
in these respects.

However, performance and memory usage are something which can be
measured[1]. I think having them in the goals requires specifying what kind
of test-cases and tools&measurements should be used to evaluate them.
I.e. decisions related to them should be based on facts, not rubbish like
"I feel gnome libs are heavy...".

Maemo "Quality Awareness" document could be looked for examples:
http://maemo.org/development/documentation/how-tos/4-x/quality_awareness.html


[1] Some performance metrics:
- system & desktop startup time[2]
- system & desktop memory usage[2]
- application startup time
- application responsiveness
- application memory usage
- how application CPU & memory usage correlates to its data size i.e.
  scalability (e.g. archiver memory usage in relation to archive size)
- power usage (wakeups and polling can affect laptop battery usage
  dramatically, but this is getting important also for servers and desktops)
- performance over network (for LTSP setups)

There are tools to measure and analyze all of these, some can be a bit hard
to use though.  Anyway, it needs to be prioritized what kind of performance
Xubuntu cares about.  Is power usage important?  What about LTSP stuff? 
In what amount of memory (most) applications should work?

Also, if some non-gnome app initially takes less memory, but with larger
sets of data takes significantly more memory than the gnome-variant, I don't
feel it's the right one for Xubuntu.

[2] System part comes from Ubuntu and has unfortunately pretty large effect
     on these measurements.


[...]
> https://wiki.ubuntu.com/MeetingLogs/Xubuntu_2008-03-26.

"19:19   meborc  i believe the main fight was in either including or 
excluding some gnome applications... this should also be somehow said in 
the statement!"

Well, to me it seemed that the main issue of "the fight" was that
the changes were not:
- presented to community before hand so that they could comment on what
  issues the changes would/could have, what changes would be needed in
  documentation, support etc
- reasoned (until the reasoning for them had been asked for many times)
- backed up with facts about actual performance/memory usage improvements


"19:30   cody-somerville IMHO, I don't think we have the expertise to have a 
focus on performance."

I can help here.  I don't run Xubuntu myself currently, but I have a lot of
experience on this area (what tools to use, when and how to interpret
the results), please use it.


"19:36   j1mc    i get the feeling that, if we have a leader, that they will 
likely have the final say on some technical matters, and that everyone 
might not agree with their perspective, but that is part of having a 
leader."

I wouldn't go as far as to say that they have the final say, instead they
should have a veto on changes that:
- haven't been presented to community for commenting,
- don't contain enough/valid reasoning or
- aren't aligned with the Xubuntu goals or (after presentation) the majority
  of the community.

The leader facilitates this kind of decision making and makes sure that
the project goals and members are respected. If the community cannot produce
a decision with his guidance, he has the final decision, but I think this is
usually an indication that the matter should be postponed until there's more
information or better design (or in current case, more focused vision :-)).


	- Eero



More information about the xubuntu-devel mailing list