-------
The GNOME project was started in August 1997 by Miguel de Icaza and Federico Mena to provide an alternative to KDE.
KDE is a free software desktop environment that relies on the Qt toolkit — a piece of software written by Trolltech that did not use a free software license. Members of the GNU project became concerned about the use of such a toolkit for building a free software desktop and applications and launched two projects: "Harmony", to create a replacement for the Qt libraries, and the GNOME project to create a new desktop without Qt and built entirely on top of free software.[2]
In November 1998, the QT toolkit was licensed under the open source Q Public License (QPL), but debate continued about compatibility with the GNU General Public License (GPL). In September 2000, Trolltech made the GNU/Linux version of the Qt libraries available under the GPL, in addition to the QPL, thereby removing most of the objections that had fuelled years of licensing debates.[3] The licensing of Qt is still controversial for some people because the use of the GPL for a library imposes restrictions on the licensing of code linking to it, including the KDE framework and any applications written for it. In particular, in order to develop proprietary software with KDE and Qt, it is necessary to purchase a commercial license from Trolltech.
In place of the Qt toolkit, the GIMP Toolkit (GTK+) was chosen as the base of the GNOME desktop. GTK+ uses the GNU Lesser Public License (LGPL), a free software license that allows software linking to it, such as applications written for GNOME, to use almost any license.[4] The GNOME desktop itself is licensed under the LGPL for its libraries, and the GPL for applications that are part of the GNOME project itself.
The GNOME desktop is written in the C programming language. A number of language bindings are available, allowing GNOME applications to be written in a variety of languages, such as C++, Java, Ruby, C#, Python, Perl and many others.
----
So I think that is a pretty good summary of Gnome's history, but my
question to the community, how objectionable is the Trolltech license
today? Regardless of the state of each desktop environment, is it
better to support Gnome since it is truly free and, in part because of
this, will provide a better platform to build on for the future?
I've heard GTK+ is improving rapidly (since this was a weakpoint of
Gnome compared to KDE for development) and that GTK+ 2 is around the
corner with various improvements (although no backward
compatibility).
I'm using Gnome now and will support it and encourage others to use it
almost entirely on the basis that the licensing is better than KDE,
although the fact it is community developed and improving rapidly are
also key factors. Do others agree with this, or do they believe
the licensing of the Qt toolkit is not of that much importance?
Ancient Fujitsu LifeBook 530T, circa 1996. Nothing other than the default apps installed except for a demo of some silly game. Norton AV physically removed but not uninstalled. Win95 in all its glory. CD Rom drive , no floppy drive, no USB drive (1996! you gotta be kidding!), internal modem, no network card. 1.3 Gb hard drive. Not much RAM (I'm not currently remembering how much). I should mention that such experiments are not for the faint of heart. Installing onto sufficiently old hardware, especially laptops, can be a frustrating experience. It might work, it might turn the computer into a brick. Fortunately this computer was free (for me, anyway... I think that in 1996, this may have been state of the art).
The original owners were friends of my grandmother. They wanted to throw it away. She wanted to see if one of us kids could use it.
First thing I want to try is a Live CD. Since I'm a musician, I like to have a dyne:bolic CD around. Pop it in, reboot. Boots to Windows. Check BIOS settings. Asks for password. Damn thing is locked. Boot Windows again. Look through the contents of the dyne:bolic CD to see if there is a way to load linux from DOS (I'd remembered reading something like this on their website once). Found Loadlin.
After several reboots and the problems that led to them, I successfully run loadlin. Kernel panic. Swearing ensues. Decide to examine the guts of the computer to see where the battery is so that I can reset the BIOS. I take it apart once, tentatively. I've never done something like this before. Put it back together after finding no battery. Try loadlin again, but with the Debian disk. Debian installer needs too much memory?! Damn ancient laptop! Not enough memory to run the Debian installer from DOS!
Try taking the thing apart again. Find new screws to unscrew. Take it more apart than I need to, fully removing the keyboard, afraid that I may not be able to get the tabs back together properly. Damned battery is soldered. Since I don't solder, I leave it, consider other options.
Boot back to Windows, since this is the only way I can figure out how to get the damned computer to recognize the CD Rom. There's got to be a way to not have to boot Windows. I hate having to look at it. Call the brother, leave an urgent -- but not desperate -- sounding message. Call another friend who is able to offer sympathy, but can't understand why on earth I'm wasting my time like this, who advises me to edit the autoexec.bat and config.sys files. Shit. Google offers little help, and my 4am brain isn't making matters any better.
I think at this point, I recognized that there were things I would want to install that come on Debian disk 2, and began downloading it.
A day or two passes.
Eventually, the brother comes over. Probably to drop me off after climbing at the gym. Looks at it, wonders why the autoexec.bat comments out the cdrom drive. Now, at least, no more having to load Windows just to access the CD Rom.
Memory comes to the rescue. I have a Slackware disk lying around somewhere. I root around my room and find it.
Using my recently-acquired Loadlin know-how, I easily load the Slackware installer. Not looking forward to trying to configure x11. There is an installer kernel for low memory situations. This is a low memory situation.
I install a base system with far more than I need. Every few packages fails. I try installing again. Same problem, but different packages fail. Try again, with minimal choices, but keeping X11. Fewer packages fail, none important, but including an X11 app or two... I can reinstall those. Installation succeeds, reboot. Wait. If it's fucked, this has been a waste of hours and hours of my life. At least the damned thing was free.
Boots. There are very few things more satisfying than the first login prompt of a recently installed OS. Ok, that's a lie. There are a lot of much more satisfying things. But this time was more satisfying than any other. I bask in my own glory for a few moments.
Run startx. As lucky as I expected to be. It crashes. Complains about pointer devices, displays, and many other issues. Since this is the exact thing I didn't feel like doing from the beginning (lazy me), I modify the Lilo config file to include the Debian CD. Forget to run /sbin/lilo. Reboot. Slackware loads. Dumbshit. Run /sbin/lilo. Reboot. Debian installer runs! Not too little memory now!
Debian installer goes without problems. I obsess a little too long over how to partition the drive. Base system installs without a hitch. First reboot, no problem. This is a little eerie. Ok, one problem. Debian, for some silly reason, changes the console font. Now it's harder to read. But at least I have a Debian laptop. Install X11 and windowmaker (what can I say, I like windowmaker), answer debconf's questions about the hardware. Run startx. I have a working gui!
At this point, I have three disks of the Debian repository (out of, what, 14?). A few packages I'd like to install, but am not interested in wasting another platter for the total of six or twelve packages that would end up being installed. I'll buy a cheap ethernet or wireless adaptor that's known to work with linux, install them then. I have some minor, but odd, questions for the Debian community. Namely:
Why is the ASClock themes package on disk 3, but ASClock is on disk 4. ASClock themes are useless without ASClock.
Similarly, Why can I install freeciv-data (on disk 3), but not freeciv (on disk 4)? Same problem.
I know the actual reasons for this (the popularity-contest package that lets the debian developers know what's most popular, so what should be on the earlier disks), but still. Useless is still useless.
But. The damn thing works. As soon as I can get it hooked up to the internet, I'll install LilyPond and ASClock. Freeciv if I have space available. (Hell, if I have time available!)
http://distrowatch.com/
As you can see on the right side, Ubuntu has rapidly vaulted to the #1 place for visits to the distrowatch website, wich is probably a decent measure of "home user" installations of Ubuntu. I am now running Ubuntu and highly recommend it to all. The Unofficial Ubuntu Guide http://ubuntuguide.org/ also rocks and the community behind Ubuntu is awesome. Just my 2 cents on my favorite distro.
Jet
After I clicked join this group I was just waiting for corporate security to come get me....
Now, if I had to choose between KDE or Gnome, it's hard to say. The philosophy behind KDE is one of unification of interface, and having a powerful core that the KDE apps can share. This is a big deal for users because KDE wants us to be able to do anything we want in a KDE environment (as in, no non-KDE apps are needed)... or, excuse me, a KDE/GPL environment. If all of the applications are GPL'd, it doesn't matter if the libraries are GPL or LGPL. LGPL is useful when you want applications with GPL-incompatible licenses to link with Free Software modules. KDE people don't care about this. They want a Free Software environment, with no incentive for proprietary KDE applications. Proprietary Qt applications are different, and there is no need for them to like to the GPL'd modules.
The Gnome camp has a totally different philosophy. They want to be able to take the best GTK applications and incorporate them easily into the Gnome constellation. AbiWord is a fine example, so is Firefox. They give incentives to write GTK apps by allowing more flexibility with licensing, but as a result, there are more options -- do you want to use Galleon, Epiphany, or Firefox as a browser? Do you want to use Evolution or Balsa for your email? This is intended to be a philosophy more conducive to the "Unix Way", of smaller applications with a smaller scope, that can be used with other (small) applications.
When I'm setting a computer up for someone who is more accustomed to Windows, I usually put them in KDE. Not just because it looks more similar -- both Gnome and KDE can be configured to look a lot like Windows, very easily -- but because the philosophy is more similar. One unified set of applications, that all use a tightly woven set of core packages and libraries. Everything else is third party.
All that said, I prefer WindowMaker. I like the dockapp thing, and enough to sacrifice the lost screen space. Mostly, though, it is fast, lightweight, looks pretty nice, and is almost exactly what I want. And there are no licensing issues.
I use KDE at work on my desktop and on my relatively old P3-700Mhz-256MB-RAM laptop for the simple reason that it renders much better than any version of Gnome I have seen to date (and it doesn't run THAT slow). The KIO slaves, among other relatively cool features, are nice; as is the centralized management. And I can't really fault Trolltech for wanting to make some money off of the deal. Us software developers have to eat after all. I feel like while the license is more restrictive, they have made the "correct" accomodations to the Open Source community, allowing free innovation to continue based off their work.
However, what really scares me about KDE and Qt is the way things could go wrong if it became the preferred platform and Gnome went by the wayside. Imagine a situation where all your favorite apps are tightly coupled the Qt API, then Trolltech starts charging for their security updates, performance improvements, and bug fixes. There are essentially no guarrantees that Trolltech will play nice the future; one can only hope.
So while I appreciate and use KDE/Qt I do hope that the Gnome effort continues. Also, on an aesthetic level, I hope that my favorite apps are built in a relatively GUI agnostic fashion.