gnome apps

[Eero Tamminen]

Hi,

On Thursday 31 January 2008, Kaspar Kööp wrote:
> i'm sorry to come into between you two, but i have also felt the
> responsiveness i got out of my old lappy, slip between my fingers when
> running the xubuntu gutsy... it used to be so smooth and fast... now it
> hogs... as in fresh install hog... therefore i have moved my lappy to
> something lighter (fluxbuntu)
>
> i understand what Jerome is saying... and i also understand why Jani
> needs numbers to back it down... but how can i back down what i say with
> numbers, how could i explain what i feel using xubuntu every day? i love
> xubuntu... but i have to say the responsiveness has gone from it

Responsiveness where?  Booting your computer/desktop?  Starting
applications?  Applications UIs themselves? Some specific applications?

The base system and Gtk itself come from Ubuntu, Xubuntu is just
the desktop and set of applications.  Whether application is Gtk or
Gnome one shouldn't have any effect on responsiveness unless
your machine runs out of RAM.  Did it start swapping?

It's possible that Ubuntu base system has bloated instead of the Xubuntu
desktop itself, added more always running services etc.  Also, for example
Gtk starting to use Cairo for (antialiased) drawing affected performance,
but it had nothing to do with Xubuntu.   Similar transitions have been also
earlier, e.g. many years ago when font handling in Gtk moved to antialiased
fonts with Xft...


On systems where this is really visible, one might consider using just
a window manager like icewm instead of running a real desktop like XFCE.


> why not the developers (before changing the programs) run some benchmarks
> and get values that are really acceptable, so we (the users) could see
> that the change in programs will not effect the overall responsiveness.
>
> Jani, you need to defend your approach, i understand, but please,
> honestly, don't you miss the fast-xubuntu we used to have? or are your
> computers all too new and fast to see the difference?

Without some facts (what actually was tested, what was measured and how,
what were the results), the changes might actually be something that didn't
have effect on anything that matters, except removing useful functionality.

It's easy for developers and users to fool themselves.  It wouldn't be the
first time when users or developers think that something is faster because
they've made a change that "should" have improved matters, but the actual
numbers reveal that there wasn't any real change, it was just a 
psychological effect and bad testing (one developer: Hey, I optimized it!
second developers: ...yeah... it feels faster... (when it in reality
isn't)).


	- Eero

> ps. i'm not trying to push the argument... i just don't like the
> mentality "feels slow? prove it" ... if it feels slow, then it is time to
> move on to something faster... if it feels slow, but you are not able to
> prove it with figures - it is also time to move on... sorry, but thats
> the way it is
>
> On 31/01/2008, Jani Monoses <jani at ubuntu.com> wrote:
> > Jérôme Guelfucci wrote:
> > > Well, the changes I remind:
> > >
> > > - gnome-screensaver -> xscreensaver: doesn't make much difference
> >
> > If you think it does not make much difference why change it?
> >
> > But actually it makes a difference, g-s-s is better integrated with
> > power management, multiple users and other possibly technologies that
> > are developed in GNOME and which we can get for free if not so stubborn
> > on being different from gnome.
> >
> > > - file-roller -> squeeze: squeeze is fast, light, seems to have most
> > > features needed for an archive manager. It's stable, upstream answers
> > > quickly to bug reports and features requests. There is a crasher In
> > > Hardy (doesn't happen in Gutsy with the same version) but work is in
> > > progress to fix it.
> >
> > Fast and light as oppposed to what? Do you have numbers to back that
> > up? You do realize that for working with archives, most of the times
> > the factors determining resource usage are the size of the data being
> > worked on and not the app itself?
> >
> > > - gthumb/or whatever was before -> ristretto: ristretto is fast,
> > > light and stable, same maintainer as squeeze and same postive
> > > feedback.
> >
> > I have not used ristretto, if it is better thab gqview I am all for it.
> > If the only plus for it is that it comes from Xfce upstream then not.
> >
> > > - gnome-mount -> exo-mount: we just loose one feature, encrypted
> > > partitions, but it's not really a key feature. This change makes it
> > > relally easy to follow debian and upstream changes.
> >
> > So on the one hand you say you have maintainer resources to keep deltas
> > in goffice and gnumeric (see a recent upload of livecd-rootfs by Colin
> > Watson, other's time is indirectly spent because of goffice deltas)
> > which do not give us any real gain, but on the other hand for something
> > that is important as the features of gnome-mount (not only encryption
> > but integration with user permissions ) you drop them to be easy to
> > follow debian?
> >
> > I find that illogical.
> >
> > > - totem -> * : this hasn't been decided yet, maybe mplayer-gtk or
> > > keep
> >
> > gnome.
> >
> > it should be something that is patent free as Vincent said and
> > friendly. I am glad that xfmedia is not considered but I wonder why it
> > was put into the seed in the first place.
> >
> > > I've a P2 450MHz and I can tell you I really feel the difference on
> > > disk usage, memory and cpu usage ! The last annoying application is
> > > Firefox 2 but Firefox 3 should be released on time and should be
> > > really better.
> >
> > You feel the difference in which applications' case? Could you
> > elaborate on how do you feel the differences on the 3 resources you
> > mentioned?
> >
> > That would be really useful data if more precise and confirmed by other
> > users too.


	- Eero