On Sat, Oct 10, 2009 at 8:01 AM, Michael Thaler <address@hidden>
To see what? A user interface that most people consider looking really dated?
> 1. Marketing to get people to give us a look.
Here are some numbers from the 2006 Linux Deskop Survey:
BlackBox 1.6 %
GNOME 35.1 %
Enlightenment 3.8 %
Fluxbox 3.9 %
IceWM 3.2 %
KDE 37.7 %
WindowMaker 2.2 %
Xfce 9.8 %
Other (please email us)
I could not find any results for 2008 or 2009 but I doubt that the market share
of WindowMaker increased. Don't you think that a huge majority of Linux users
prefer a more modern looking desktop environment with some eye-candy and will
be just dissapointed if the see gnustep in its current state?
I don't really like too much eye-candy personally. The first thing I did at
work was to change Windows Vista from Aero to Classic mode because I prefer
Windows Classic (Windows 2000?) look compared to Aero. On the other hand, I
think Snow Leopard looks quite good and I also think KDE4 and Gnome look sort
But the NEXTSTEP look is too old-fashened even for me (I don't care if it is a
masterpiece. I don't want to put a picture of it in a frame on the wall, I
want to use it as a desktop environment). I really like ObjC and the
openstep/gnustep/Cocoa APIs. But everytime I sit down to develop something
using gnustep, the old-fashened Look & Feel kills my motivation because I
think nobody will use it anyway and I decide to use Qt/KDE instead (I am
actually a former KDE developer).
I just completely disagree with your arguments here. So what if you like "eye-candy"? Riccardo and Richard like the grey NeXT look, and using the mailing list as the sample space I would say it's divided roughly 60/40 for the NeXT look over the so called "eye-candy".
Have anyone here using GTK or Qt applications ever actually built these from scratch? I would assume no, because the idea of "an easy install" always comes up. I've personally next built Qt, but have done GTK. Simple put, it's hell! You have 15 dependencies you need to satisfy before GTK even configures without an error, and another 10 dependencies to get decent support for everything you want (http://www.linuxfromscratch.org/blfs/view/svn/x/gtk2.html
everything in "Required" + their dependencies). Then, after you're all done with that you still end up with a "dull and grey" look... so you go out and install the clearlooks theme engine. How is that any easier than building GNUstep? I can truthly say, it's not. I still say we need distribution support, which the little that we do have we seem to be loosing. How do we get their support? Marketing will become much easier if all we need to say is "do apt-get install gnustep-core gworkspace" instead of "grab the sources from svn and compile".
To be honest, I don't like WindowMaker. don't like using and think those icons are a waste of my precious screen space. What I'll generally try to do is use nothing but GNUstep applications with no window manager (since GNUstep supports it, even though it has issues).
On top of all that, GNUstep has a serious identity crisis. It's such a far departure from the usual Gnome/KDE/Windows desktop metaphore. So you end up with the problem that most people expect you provide at least a half working desktop in order to feel comfortable, but that's not GNUstep's goal, it's just a development environment. You can see that littered all over Michael's post, he's trying to compare GNUstep with KDE and Gnome instead of with Qt and GTK (+ GLib and GDK). Etoile is definitly working to bridge that gap, but even so it's not easy to get it. I personally do not build all of Etoile because it's just simply too much work. I would not use Gnome if I had to build it everytime either.
> 2. Eye-candy to draw people in and get them to try things outFor me, the fundametal question is what direction gnustep wants to take. Does
> (changing the default theme won't do that ... we need to have a group
> of three or four good themes to appeal to different people)
gnustep want to appeal to former NeXTSTEP/Openstep users? Or does gnustep want
to be a MacOS X for Linux and other OSS operating systems? In the former case
I am not really interested in gnustep. Openstep/gnustep might provide a nice
API, maybe it is even a bit nicer then Qt, but I don't really see gnustep
being adopted widely if it just tries to provide an Openstep-like API with a
Nextstep-like inteface. If gnustep aims to provide APIs and a desktop
environement similar to MacOS X I would be very interested. But I don't think
gnustep can do both. Either it will continue to try something similar to
Openstep or it will change direction and try to be something similar to MacOS
X. A simple theme will not be enough to make a gnustep desktop really look
cute and appealing.
Again with the "eye-candy". GNUstep doesn't need to be providing this by default, and no project does. GSTheme is a great step forward, but it's still unfiinished, and even when it is finished it shouldn't be
For this to happen, gnustep needs more developers. In my opinion the only way
> 3. Enough good quality stuff so that people don't try once and then
> give up.
that this can happen is if more people start to use gnustep. Apparently
gnustep did not attract a lot of users / developers for the last 15 years. So
maybe it is time to change direction?
Here I agree with one of the messages that was posted before on this thread. GNUstep needs to stop chasing butterflies. GNUstep barely has full 10.3 compatibility, yet there already are 10.5 features in. In my opinion, and that's all it is since there's not much I can do to help in this aspect, GNUstep needs to focus on finishing full compability with one version of OS X before moving to the next. Pick one, and stick with it until you're at least 90% finished before moving to the next. I'm going out on a limb here, but I'd say it will take a few years for a good chunk of applications to move to 10.5 features. New application might start using these features, but application being ported from one version to the next will probably not use them.
Anyway, that's enough for me. I've already spent too much time typing.