Linux Today: Linux News On Internet Time.

Community: It's Not About Linux, It's the Desktops, Stupid!

Sep 15, 2005, 00:15 (27 Talkback[s])
(Other stories by Zeek Greko)

[ Thanks to Zeek Greko for this article. ]

By Zeek Greko

Former US President Bill Clinton's election campaign office (dubbed "the war room") had a sign on the wall that simply stated "It's The Economy! STUPID!!" Put there to remind workers that despite the myriad of other social and political issues needing their attention, they should keep their primary focus on the issue they had determined was of most concern to the voting public. The Economy.

The Linux pundits' ever-present "Is Linux Ready For The Desktop" articles, and their recent rash of "The Honeymoon Is Over For Linux Adoption" articles, refer to a waning interest among the general computer-using public toward moving to Linux kernel-based systems. Some valid peripheral issues are raised in these articles but they are in fact peripheral. They miss the point. The underlying core issue is the desktop.

The Desktop Is The Operating System, Isn't It?

To most non-technical users, the desktop they see and use is, in totality, what they understand the Operating System to be.

If they are using KDE, KDE is Linux to them. If it's GNOME they're using, then GNOME is what Linux is to them. They don't care, nor is it absolutely essential that they know more than that. That nearly all endeavors that make use of the Linux OS kernel are referred to as "Linux" is an aberration. The sheer broadness of the term "Linux" is also a source of end-user confusion. Normally, to an end user, the underlying OS kernel is never mentioned when making reference to or describing their system, if it is even known to them at all.

In my experience, most Windows 2000/XP users don't know that the core OS kernel they are using is NT. Neither do Windows 9x/Me users know they are using DOS. No, to them they are using a "Dell Model 001 with Microsoft Windows XP" or some other such comparably defined description (Make:Model:Microsoft Windows Desktop Version). Applying this end-user friendly, easy to understand and remember description to Linux kernel-based systems would logically put the KDE and GNOME desktops into a place of prominence within that description (Make:Model:Distribution: KDE/GNOME Desktop Version) relegating Linux to lesser or no mention at all.

This would be in line with current OS norms and much less of a source of confusion. What's your computer? "My Computer runs Windows," "My Computer runs KDE," "My Computer Runs OS X," "My Computer runs GNOME." If they have ever seen these desktops before, these responses convey a basic understanding of your user experience to someone's "minds eye" as quickly and effectively as "Big Mac" conveys what you had for lunch and where you had it. "My Computer Runs Linux" is so totally nebulous as to tell them next to nothing more than that the answer is going to take a while and require further dialog. Maybe even a notepad.

Linux Desktop Polarization

Nothing starts a prolonged and downright ugly flame war in the forums faster than a KDE vs GNOME discussion. This in itself says a lot about how even technically savvy people within the Linux community personally equate Linux to either KDE or GNOME. People are passionate about loving or loathing one or the other. This polarization is a major forking point within the Linux community.

Who among us, who were here at that time, could forget the reaction to Novell, a GNOME-centric company, buying a major KDE Distro, SUSE? And the sheer panic that ensued among SUSE users that Novell would dump KDE and replace it with GNOME? Sage-like predictions of the imminent demise of KDE abounded because of it.

On the now defunct The Linux Show, Eric S. Raymond's prophetic utterances were to the effect that KDE had six months to live. Well, that obviously didn't happen. If Novell had in fact switched to GNOME, there simply would have been a mass exodus from SUSE to some other acceptable KDE distro, with Novell losing most of the "good will" they paid good money for in the process.

The die has been cast. Both KDE and GNOME are hear to stay for the foreseeable future, and are as much a part of the Linux community as Linux itself. They are even more a part of the F/OSS community than Linux. When the SCO debacle happened, the speculation (if they won) was not whether F/OSS would survive, but what other kernel would need to be thrown under our KDE and GNOME desktops. KDE and GNOME separately enjoy "Brand Loyalty" that most corporate CEO's can only dream of, and we would all do well to keep that in mind and build on it.

The Butcher Shop Branding Pandemic

Common among all major Linux distributions, and most minor (spin off) Linux distributions, is that they use KDE, GNOME, both, or a hybrid of both (like Red Hat's Blue Curve). Also, all major Linux distributions, and most minor Linux distributions, create their own unique and customized configuration tools (DRAK, YAST, etc.). As a consequence, and without exception, they break, disable, remove outright, or just leave them to crash when you try to use them, whatever configuration tools that are natively provided by the desktops. KDE and GNOME are treated simply as packages to them, albeit large ones. They also summarily remove components and hack and butcher away as they please.

Kubuntu, as a recent example, removed KDE's login, replacing it with GNOME's (see the "Polarization" section above for why that's a stupid move). Further, in their attempts at distinguishing their product (branding), the various Distros so modify the desktops that they are nearly unrecognizable as their core desktop. An example of this is Linspire, which so reworks everything it touches, including the desktop, that an actual KDE user would need a lot of time just to acclimate themselves to it. Because of this, many of you have never actually seen KDE or GNOME in their pure states. What you see is your particular Distro's modified (corrupted) versions of them.

Any MS Windows user can sit down in front any other MS Windows computer of the same version and use it, or even configure it without skipping a beat. Not so with Linux. The same distribution, yes, but Linux distributions in general, no way. Not even with different Linux Distributions using the same desktop and version.

Distro Du Jour Burnout, Linux Becoming Unsupportable

I think the pundits' perceived waning of interest toward Linux has some truth to it, but it might be more accurately defined as "Distro Du Jour Burnout."

Linux kernel-based Operating Systems have now become hopelessly mired in convolution, rife with incompatible redundancy through the over-proliferation of the various distributions' multiple package management systems, techniques, and repositories; their unique configuration tools; and other unique incompatible branding modifications. This situation is definitely not conducive to hardware vendors, OEMs, or third-party commercial software vendors wanting to come on board Linux.

Also, competent general-purpose Linux system support has become virtually impossible by all but the the most fanatical devotees of All That is Linux. The volume of specific, esoteric knowledge that must be maintained is huge and continues to grow exponentially. This is an absolutely critical situation that cannot be overstated.

We Need More Distros--Exactly Two More

If KDE and GNOME are to become (and rightly so) the instantly recognizable "Big Mac and "Filet O' Fish" of Linux, they will need become separate core entities and build their own core OEM Linux distributions. They may have to just to retain their identities and continue meaningful desktop development that is free from the distros butchering their work.

These KDE and GNOME OEM cores would ideally become two separate community projects. This would leave us with one primarily Qt-based core distro, and one primarily GTK-based core distro. This two separate-core concept is strategic to help maintain healthy competition among the OEM cores and to prevent the creation of another dangerous Microsoft-style "OS Mono-culture."

This would also reduce overburdening (as they are now) device and applications vendors who might be persuaded to support Linux kernel-based systems if supporting them was simplified. GPL'd projects too. for that matter. Since only two basic core binary packages of a product, project, or driver, would be all any of them would need to provide, centralized package repositories for KDE- and GNOME-compatible software could then become obsolete. Plus, two separate cores also sidesteps the problem of two fierce competitors (the KDE and GNOME teams and their supporters) having to come together and agree on anything, while simultaneously eliminating wasteful redundancy and enabling tight integration of each of their installers, package managers (if needed), and configuration tools.

Lesser known, or yet to be built, desktops could easily wed themselves to the OEM core that uses their primary base (XFCE = GNOME/GTK base), and could reuse all/most of the same software packages and configuration tools.

The various distributions would jointly support these OEM desktop cores, then build their offerings on top of them by selecting and bundling best-of-breed GPL/Proprietary tools and applications to provide complete usable end user and/or special purpose systems with their related support services. Healthy competition between the distros would be changed somewhat in their primary focus, but would continue more or less unabated. They would move beyond just successful basic installation and setup of the desktop to being actual value-added solution providers. No less importantly though, Linux-based systems would in short order become competently supportable by many more people of average technical ability, thereby drastically reducing TCO. And, as a result, they would become more widely accepted as truly viable alternatives to the MS Windows monopoly products.


End users, support people, and independent software providers that are interested in participating in Linux kernel-based alternatives to Microsoft are starving for the kind of clarity and simplicity such as is outlined here, or some similarly acceptable solution. We should discuss, debate, and/or offer alternatives to it. Whatever ends up making the most sense, Then we should advocate it. Whatever we as a community decide to do, we can't stay here.

Related Stories: