[Usability] Re: [Desktop_architects] Good Job

[Lubos Lunak <l lunak suse cz>]

 Gee, won't this stop :( ?

Dne čtvrtek 15 prosinec 2005 06:27 Ryan Richards napsal(a):
> Mr. Waugh:
>
> Having recently read about Mr. Torvalds's comments on the KDE vs. Gnome
> matter, I found the thread and drilled down into it a bit. It seems to me
> that the features that Torvalds and others are complaining about are
> trivial at best, while the complaints themselves are bloated. I intend to
> continue using Gnome, and I'd like to tell you why: it uses 50MB less
> memory than KDE. Now, I'd only classify myself as an intermediate Linux
> user (not an expert, but not a beginner),

 And I'd classify myself as an experienced KDE developer who's spent quite a 
lot of time working on improving KDE performance (so I probably am one of the 
experts).

> but I can tell when Mepis and 
> Fedora with KDE use 140MB to show me a desktop, while Ubuntu and Fedora
> with Gnome use 90MB to show me my desktop. Yeah, I could use Ice WM or no
> GUI and save EVEN MORE memory, but Gnome gives me features that are worth
> the extra RAM. The features talked about in the thread are not worth--at
> least to me--50MB.

 This is either an outright lie from a troll intended to start another 
flamewar or it's a claim from somebody completely clueless and incompetent as 
far as anything performance related is concerned. Make your own pick.

 And since such nonsense spreads quickly and happens to live long for some 
reason, I think I should give just a few quickly measured numbers:

- my full KDE session, which is right now 8xKonsole, 1xKonqueror showing a 
webpage, 1xKWrite, 1xKMail (accessing my relatively large IMAP account) and a 
handful of systray apps, causes total memory usage of about 130M after 
startup. Right before I launched KDE the memory usage was about 47M (27M even 
without X running). That's less than those alleged 140M for just showing a 
desktop.

- A bare KDE session causes about 22M memory usage increase when launched in 
plain X while a roughly (I hope) equivalent GNOME session causes about 34M 
increase. BTW, it also starts faster, and I have pictures for this 
(http://www.kdedevelopers.org/blog/280).

Now, this is Garnome-built 2.12.2, which probably doesn't(?) have all of the 
recent GNOME performance work, I'm no GNOME expert, and my KDE3.5 has few 
small patches that haven't made it upstream yet, so this is not exactly fair; 
moreover I didn't bother more than just getting the -/+ buffers/cache numbers 
from free after a reboot, so the numbers aren't that good either. But it's 
certainly better than comparing apples and oranges using a thermometer, the 
numbers were measured the same way with the same conditions, and it's me 
giving these numbers (please see above for details).





 And now, please please pretty please, could all this flamewar material, 
crossposting from usability gnome org and other stuff that doesn't belong 
here please stop? Thank you. Some people would like to get some work done 
here, too bad I'm not sure anymore because of all this noise what this work 
was supposed to be.


PS: Just in case somebody feels like discussing the numbers above, there are 
mailing lists like kde-optimize kde org and/or performance-list gnome org for 
that, not this one.

-- 
Lubos Lunak
KDE developer
---------------------------------------------------------------------
SuSE CR, s.r.o.  e-mail: l lunak suse cz , l lunak kde org
Drahobejlova 27  tel: +420 2 9654 2373
190 00 Praha 9   fax: +420 2 9654 2374
Czech Republic   http://www.suse.cz/