bloated vs. lightweight WM's

Brian J. Tarricone bjt23 at cornell.edu
Tue May 25 19:54:53 CEST 2004


On Tue, 25 May 2004, Sami Samhuri wrote:

> You would have had to seen the entire thread [1] to understand what the
> original poster was trying to achieve. The poor guy just wanted to see
> the difference between memory use in different environments for his work
> with all his services and programs open to simulate normal use. He did
> reboot between each test and had the services running on purpose. After
> posting he was inundated with messages telling him why his test were not
> accurate and all sorts of other things similar to what's being said
> here. The difference between a window manager and a desktop environment
> was also pointed out to him. :)

but the problem is that you can't accurately measure the environments' 
memory usage with other apps running.  simulating 'normal use' is 
meaningless in this kind of test, and can only make the results less 
accurate.  as i said previously, adding other non-DE-related apps to the 
mix doesn't impact the resource usage of the apps that make up the DE 
(with the possible exception of the WM, but that's in a constant and 
predictable way).  adding 'real-world' applications to the mix doesn't 
in any way make the test more valid or less synthetic.  it just serves 
to make the results uninterpretable.

> He was not trying to benchmark the individual environments, rather he
> wanted to know how they stacked up against each other on his machine for
> his own purposes.

but... he _was_ trying to benchmark the individual environments.  the 
thread is titled "bloated vs. lightweight WM's".  that implies an 
environment test (WM/DE definitions aside), and the rest of his post 
bears out that implication.

personally, i think the guy had the right idea - it _is_ interesting to 
see how the different DEs stack up resource-usage-wise.  as i've been 
working on xfdesktop, i've been watching memory usage very closely as i 
add features (in an attempt to minimise mem usage, obviously).  but he 
went about his tests in entirely the wrong way.

speaking of the original post... he says he has a P4 2.3GHz with a gig 
of RAM.  and he's worried about resource usage... ^_~

	-brian




More information about the Xfce mailing list