GNOME hacker: Culture isn't holding desktop Linux back
'There is no single simple issue to fix'
Open source bigwigs like Alan Cox, Miguel De Icaza, and Linus Torvalds can bicker all they want, says GNOME hacker Michael Meeks, but changing developer culture won't do a thing to attract more consumers to the Linux desktop.
"I happen to mostly agree with [De Icaza's] conclusion – that we're still facing a huge uphill struggle there," Meeks writes in a blog post. "While I have huge respect for his experience and insight, I think the causes are larger. My punch line is that the Linux Desktop faces a huge and multi-factored ecosystem challenge, there is no single simple issue to fix."
Meeks' comments come in response to a recent online dust-up that saw several high-profile open source developers pointing the finger over Linux's lack of success as a desktop OS.
In August, De Icaza accused the open source GNOME desktop environment project – which he founded – of "killing the Linux desktop" by placing too much emphasis on "engineering excellence" over stability, in a blog post that heaped much of the blame on Linux founder Linus Torvalds.
It was Torvalds' often cantankerous personality that set the tone for GNOME and much of the open source developer culture, De Icaza claimed, by encouraging other developers to be "mean and harsh" in response to criticism and to run roughshod over others' needs and expectations.
The result, he said, is that core open source code changes too often and in too radical ways, resulting in systems that are too difficult for admins to maintain across versions and that break third party applications too often.
Torvalds quickly shot back on Google+, saying the idea that his attitude is what causes the GNOME project problems was "laughable" and that GNOME developers were "in total denial about what their problem really is."
Fellow Linux kernel developer Alan Cox chimed in on Torvalds's side, saying the GNOME team's track record for delivering stable, backward-compatible code was so poor that the GNOME desktop should really only be considered "a research project."
But according to Meeks, even if Linux and GNOME are turning away some independent developers, that's only one of the reasons why Linux hasn't enjoyed more success on the consumer desktop.
For one thing, Meeks says, it's not as if Windows, iOS, Android, and other, more successful client operating systems are necessarily that much more attractive to developers than Linux. They each have their own quirks, bugs, and performance problems – but what they also have are large addressable markets.
In other words, Meeks says, it's the same old Catch-22: Consumers won't come to Linux until they're sure they'll have access to a wide range of applications, and developers won't write applications for Linux until they're sure there will be a large market to buy and use them.
Even more important than encouraging developers to build apps for the platform, Meeks writes, the Linux community needs to figure out ways to encourage OEMs to ship hardware that comes with the Linux desktop preinstalled. Right now, he says, that's not happening – and the issue is all about money.
"One of the major business problems of hardware enablement is that it takes a constant investment of real cash to pay excellent engineers to make (brand new) hardware work reliably," Meeks writes.
As long as OEMs and resellers treat Linux as a "free" OS, Meeks says, there simply isn't enough money available in the ecosystem to support those developers. Even if downloading and installing Linux costs nothing, OEMs incur additional costs to get the OS working well on each specific hardware configuration.
By comparison, OEMs who build Windows systems can earn revenue from co-marketing deals, bundling antivirus and trial versions of software, advertising, and other opportunities that typically aren't available on Linux.
As a result, bundling Windows with hardware essentially costs OEMs nothing. In fact, it probably earns them money, which in turn makes it easier to pay developers and component makers to ensure that specific hardware configurations are well supported by the OS – all of which makes Windows more attractive.
Furthermore, Meeks writes, even when PCs that come bundled with Linux are available, attracting customers to them is another problem. Often, Linux models are difficult to find in stores, and distribution may be limited to small volumes or certain geographic regions. Advertising is limited – money again – so even customers who want Linux PCs seldom know what specific models to ask for.
Meeks has considerable experience with desktop Linux himself. In 2000, De Icaza hired him to work on GNOME at Ximian, and he stayed on at Novell when it acquired Ximian in 2003. Meeks points to Novell's commercial Suse Linux Enterprise Desktop product as an example of how a Linux desktop system can be done right.
"In my view the most hopeful strategy for the Linux Desktop is to make it ideal for an enterprise, while not crippling it for consumers and very early adopters," he writes.
The fact that enterprises don't like change much, tend to purchase new PCs in batches, are typically concerned with only a limited number of features, and have technically literate IT staff on hand all contribute to making the economics of Linux desktops more attractive in those environments, Meeks says.
To that end, he wishes more hackers would spend less time worrying about things like GUI controls, flashy desktop effects, and how their desktops compare to Windows and Mac OS X. Instead, they should work to build systems that are secure, easy to manage, integrate well with existing network services, upgrade smoothly, and require little retraining – all qualities that could help desktop Linux gain better footing in enterprise environments.
"It is perhaps not glorious – enterprises care little for the bling that doesn't have a direct, unarguable business benefit – and bling is fun to hack on," Meeks writes, "but surely creating real value, that allows people to work more efficiently, reliably, and speedily has to be a satisfying thing to do as well." ®
Sponsored: IBM FlashSystem V9000 product guide