GNOME on the range
Oddly enough, the quality that struck me the most about the Expo was its normalcy: Wandering through the San Jose, Calif., convention center, it seemed like Just Another Trade Show. I was pleased to find representatives from the BSD camp, as well as numerous big names (Compaq, IBM, HP, Sun, as well as SGI and its cool-looking big iron equipped with lots of blue blinky lights), in attendance.
But the LinuxWorld event that seems to have stuck in most people's minds (or craws, depending on which faction one is aligned with) is the announcement of the GNOME Foundation.
Now I'm generally a good little capitalist and consider competition in markets a good thing in most cases. (That said, the cellular standards situation in the United States -- viz. the utter lack thereof -- is a perfect example of where Letting The Market Decide turned out to be a Really Stupid Idea and Those Responsible in the FCC Should Be Fired for It, If They Haven't Already Been.)
Wireless vitriol aside, I think that all this buzz about GNOME vs. KDE is completely ignoring the real issue. I mean, really, who cares whether the underpinnings are CORBA-compliant? Before any developers out there begin squawking that they care quite a lot, thankyouverymuch, it bears pointing out that GNOME and KDE aren't really for them. They're for those hypothetical Linux Desktop Users Of The Future.
And what is most important to users about their graphical computing environments? Two words: user experience.
I think it's very nice that the founding members of the GNOME Foundation sat down, had a press conference and decided to consort with one another. I think it's mildly unfortunate that the KDE camp considered itself so cornered that it felt compelled to issue miffed- and wounded-sounding position statements that it will continue to fight the good fight.
But none of this politico-marketing posturing folderol really matters. You see, the business about history being doomed to repeat has a rather large salt lick of truth to it. And the important lesson learnt waaay back in the 1980s is that human-interface consistency matters.
In my spelunking around the GNOME site today, I found loads of documentation, white papers and other forms of text describing in intricate technical detail how the guts of GNOME work. But nowhere could I find the GNOME equivalent of the now-legendary Macintosh Human Interface Guidelines.
Author Neal Stephenson is well-loved by the Linux community for many reasons. In the acknowledgments to his novel "Snow Crash," Stephenson explicitly credits Apple's Human Interface Guidelines. Those reading "Snow Crash" with a slightly analytical eye will recognize that the metaverse has a great deal of consistency, continuity and interface discipline to it. This is by design.
In the eighth paragraph of "In The Beginning" -- a rant about operating systems -- Stephenson is quite clear about his feeling that "interface is very important."
Yet it's precisely this perceived importance of interface consistency, continuity and discipline that is AWOL in any of the current Linux desktop efforts. Without it, Linux will not appear on the desktop in any meaningful numbers. It seems that the fervent engineers have neglected to internalize that users actually care about usability.
Eazel is developing Nautilus, GNOME's next-generation file manager. Nautilus is to the GNOME environment what the Finder is to the Mac and IE has become to Windows. It's good that Eazel, of all companies, is working on this, since it's staffed by many of the people who are responsible for the Mac's original human interface.
Eazel's current job posting for a Human Factors Expert, which also includes a reference to in-house interaction designers, further underscores the fact that Eazel is doing more than just going through the motions.
But Eazel is not the GNOME Foundation (despite the fact that www.gnomefoundation.org does bring you to Eazel's site, albeit with some registrar's nav bar down below). And Nautilus isn't the sum total of the GNOME user experience.
While Eazel can lead by example, it's doubtful that this will be enough.
Open-source developers don't like being told what to do. That's one of the reasons they are open-source developers. And other than Linus Torvalds (who maintains total control over the Linux kernel but wants nothing to do with the interface), there is no one in the open-source community who commands the respect necessary to make a usable human interface a reality.
I will never forget a conversation I had some years ago with Bruce Tognazzini while he was at Apple. It was around the time when the company's Hypercard software was considered a cutting-edge RAD/end-user programming tool. However, unlike development systems that provide a collection of standard interface elements from which to construct an interface at least recognizable to the ordinary user, Hypercard had a total of two interface tools: bit-mapped graphics and clickable regions.
As many will remember, this flexibility resulted in some pretty gnarly interfaces due to their completely inconsistent appearances and behavior. So I asked Tog (arguably the foremost authority on the Mac interface at the time) to give me his opinion on what Hypercard had done for the discipline of interface design. His response: "I cry a lot."
I've worked with highly trained, black-belt interaction designers and have absorbed a fair bit of knowledge from them through osmosis. Yet despite all their learning and experience, they continue to think this stuff is hard. If open-source OSes don't develop their own human interface guidelines and find a way to make them stick, Linux on the desktop will become a forlorn dream.
Stephan Somogyi doesn't really like Aqua all that much, either. Especially that dock. What a senseless waste of valuable vertical screen space.