Tijs van Bakel wrote:
> Personally I wouldn't like a graphical-only install. I like to run
> Linux on a 'slow' machine like a 486. A graphical installer would be
> nice, but fallback support along with it would be even nicer. I think
> being able to run Linux on these cheap machines shows off the real
> power of Linux even now.
Yes. One should be able to fallback to text, or even to a serial console.
But I do like the new Caldera installer, because it really uses multitasking
during install. One can start configuring whilst packages are being
installed. This can save time for admins.
> In order to get libggi included with distributions, I think plain
> eyecandy would suffice. Especially 'demos' are nice. Take a look at
> http://www.scene.org and note that there are zillions of DOS/Windows
> demos and about 5 to 10 (mostly pretty bad, or DOS ported) Linux
> ones.
Demos are not applications, and (at least that kind of demo) never gets included
in a Linux distribution. But yes, Linux needs more demos.
As of yet, there is not a single application that uses libGGI that has made it
to any distribution. (correct me if I'm wrong)
As of yet, the only people using GGI are the developers and some true fans.
All of the other libraries that are succesfull (gtk, qt, libc etc.)
are succesful because there are actually applications using them (gimp, gnome,
kde, linux). This stimulates use of these libraries in other projects, by other
programmers. I don't think a library without a sample application that gives
people somthing they didn't have yet but want to have, like the gimp for
instance,
is ever going to be succesful.
> Wouldn't it be
> enough for the library to prove itself, as it is doing already?
In principal yes, in practice no. There are so many libraries around, all
claiming
to be good, but unless it has proven itself in a real-life situation, like kde
for example,
only very few are going to use it.
> One could make a screensaver in native X11 that runs 0.000001% faster than
> the libggi version in X11, for obvious reasons, so the speed gain
> proposed by the original author wouldn't be due to libggi.
I'm not sure how libGGI works when used in X window. Does it make calls to Xlib
?
It might still be faster if it is easier and more efficient to program in
libGGI.
(like programming in ncurses might be faster then trying to write out all escape
codes yourself)
But ok, say that programming in libGGI with target X is not faster than
programming
in Xlib, it is still easier, supposedly. And most of the 'hacks' in xscreensaver
are
really slow, easy tricks.
Im sure people on this list can make much nicer 'hacks', faster as well as more
imaginitive.
> By promoting libggi this way, people think they are being fooled, which
> would make them angry in the end.
I don't think so. Mesa is being used in xscreensaver, it looks slow at least on
my
machine, if you don't have a 3dfx card that is supported, but still I don't see
people get mad at it.
----------------------------------------------------------------------
UNIX isn't dead, it just smells funny...
Run Linux! Keep The Net Free!
----------------------------------------------------------------------
