Where are the goals of XFCE.

Benedikt Meurer benedikt.meurer at unix-ag.uni-siegen.de
Mon Jul 14 19:18:55 CEST 2008

Vincent wrote:
>     > All this "lightweight desktop" stuff is pretty stupid anyways. In case
>     > you didn't notice, it's 2008, and you will have a hard time buying
>     a new
>     > machine with less than 512MiB of RAM. Performance isn't an issue
>     either
>     > with todays CPU processing power. Instead, it'd be nice if people
>     would
>     > start to improve the user experience by - for example - reducing the
>     > latency of the desktop (i.e. make clever use of XCB instead of Xlib to
>     > avoid blocking on round trips to the Xserver) or making the Gnome
>     people
>     > actually accept Glib patches to reduce unnecessary overhead in the
>     > object and signal handling. Wrt. latency there's also a lot of stuff
>     > that could be improved in the kernel.
>     I hope I'm not intruding into the discussion, but talking lightweight is
>     still not stupid. We have some old computers from where I work (this
>     is a non-profit and it's very hard to get new computers even with all
>     the talk about replacing the old equipment) that we will no longer
>     upgrade because it is already too slow to run the old software, let
>     alone the new. Lightweight certainly still is useful and will still be
>     useful in the next few years.
> Of course talking light-weightness is not stupid. However, it is madness
> to assume that, because Xfce ran on old computers five years ago, Xfce
> should still run on those same computers today even though they've
> become five years older. "Old computers" of today can handle more than
> "old computers" of yesterday.

People have a somewhat wrong understanding of what "efficiency" really
means in terms of software. It's not about overall memory usage + CPU
cycles consumed. For example, I can write extremely efficient programs
that require 2GiB of memory and take 99% of the CPU time. So what does
that tell us?

For desktop software people tend to assume that "lightweight" implies
"efficient" or vice versa, but _most_ people don't know how to interpret
(or measure) either of these terms. If you ask them, you don't get
useful answers (as can be seen here); it often comes down to
propositions like "XYZ feels sluggish" or "XYZ takes too long to start"
or "XYZ redraws too slow". Please feel free to prove me wrong and make a
valid, useful statement about how CPU time and memory usage actually
affects "efficiency" and "lightweightness" of the desktop.

If you'd actually read what I wrote instead of just posting the same
crap again and again, which is indeed stupid, then you'd notice that the
real problem cannot be measured in terms of memory usage or CPU cylces.
This kind of non-sense in the OSS desktop community is what keeps me
(and probably not only me) from spending my spare time with Xfce (and/or
other OSS desktop software). This might be benefical for some projects
now (i.e. Gnome devs posting nice graphs about reduced memory usage and
increased performance, even tho the overall desktop feels even more
sluggish than the previous release; and that's by no means a Gnome-only
problem, just serves as a good example), but in the long run it's just a
waste of valuable time.

So be sure to actually read what I wrote prior to starting your reply.
And be sure to read the quoted part of my original mail.


More information about the Xfce mailing list