sent this to Jeremey directly by mistake (*sigh*). sometimes I don't
know how I get by at all....

Jeremy Bertenshaw writes:
 > A lot of these problems have been echoed to me in the past 
 > from people trying out linux for the first time, especially simple
 > things like changing your screen res in X...
 > 
 > http://people.trustcommerce.com/~adam/top10/wrong.html
 > 
 > I think it's about time to biff X if Linux wants to get serious
 > in the desktop market, anyone know of any projects in place
 > which are looking into this?  I know the mac on linux guys are
 > not far away from getting a port of OSX onto linuxppc, but 
 > whattabout for us wintel guys?
 > 
 > jeremyb.
 > 
 > 

I suggest you look at http://www.cbbrowne.com/info/xbloat.html .
there was the Berlin project for a next-generation windowing systemm
but it has turned into Fresco it seems, http://www2.fresco.org .
They're still calling it Berlin in places though, don't know what's
going on there. 

A breif rundown of points from the article: 

*) X isn't bloated in terms of memory. Tops lies about the memory
usage. The memory usage is of the order of 10MB, and most of that is
screen data (and thus, can't be reduced any further by any X
contender). Moreover, with the ever-increasing memory available these
days, it doesn't seem like much of an issue. And it's hard to believe
that a contender will make a serious effort to run well on your aging
486. 

*) X isn't really all that slow. The only real concern is games and
animation, and this is getting better with more recent released versions.

*) there is no replacement contender. 

*) if there was one, it would only replace X if it offered a real
advantage, did everything important that X did, interoperated with X
somehow, and was portable to all (or almost all) the platforms that X
supports.

*) X is getting better by the day. 

He also notes many of X's flaws, like the 'hacks upon hacks' and
things, but it seems that the flaws are getting fixed at least to a
certain extent. 

He doesn't address the configuration issue, though, but I doubt that
this is somehow necessarily inherent in X. In point of fact, you *can*
change the screen resolution, by pressing ctr-alt-+ or crt-alt-- to go
up or down a resolution respectively. I think the author of the "10
things" wants it to be in a menu or something.  Doesn't windows
require a reboot to do this? 

Some other problems that could end up getting blamed on X aren't
really anything to do with X, like: 

*) GUI proliferation, ugly apps. X isn't a GUI, it's a display driver
and communications protocol. 

*) fonts. X does its bit fine, as far as I can see. It now has
antialiasing and support for TrueType. The difficulty with fonts is
that their installation isn't straightfoward and printing isn't
necessarily coupled to the screen display --- it's a totally different
kettle of fish. However, this is a job for Gnome and KDE, not for X.

Other problems, like making good use of the capability of the graphics
card to provide better rendering of windows etc. are a problem, but
are likely to be just as much a problem if not worse for any
contender. 

In summary, any possible X contender is a looong way off, and X's
flaws either aren't flaws, or can be worked around, or would be a
problem for a contender. 

Sure, it's entirely possible to either redesign X from scratch or
design something else that's better, but if you want Linux to take
over the desktop it can more easily and quickly be accomplished simply
by making X better. 

Andrew. 

Reply via email to