Hardware manufacturers should do more to make their devices compatible with Linux, experts agreed on Tuesday.
Making drivers available promptly and automatically would help open source users, according to attendees at a panel on the Linux desktop at the LinuxWorld conference in Boston.
"Things like wireless, customers expect to just work. We need power management to just work, to make sure our laptop batteries work for longer than 15 minutes," said John Cherry, the manager of OSDL's Desktop Linux initiative.
Guy Lunardi, a desktop architect at Novell, agreed that the automatic support of drivers is vital, and must be provided as soon as the device becomes available. "The key point is that it must be timely. The drivers will eventually get there, but the driver must be there when the device is supplied [to customers]," Lunardi said.
The OSDL is trying to help the open source community to get access to hardware specifications before a product is publicly available, by organising nondisclosure agreements between developers and hardware vendors.
It is not just down to the open source community to write drivers — hardware companies are increasingly providing drivers for Linux as well as more mainstream desktop operating systems, according to Waldo Bastian, a desktop Linux architect who works for Intel.
But, although Linux drivers are increasingly being developed in a timely manner, many drivers need to be installed separately from the main installation.
"How can we make it that drivers are included in the default install, so you don't have to do extra installs to get wireless working?" a member of the audience asked the panel.
Some drivers cannot be included in the kernel as they are proprietary, but Novell is working on a process that could automatically install drivers after the main installation has finished, according to Lunardi.






Talkback
The experts are wrong. The loudest voices within any market are the ones you don't hear. Sure, we hear loud voices yell out for "instant driver support", "ease of use", "better user experience", "educational programs", "nationalized support" and what not. And you know what? They are all wrong. OK, it's important but it's not a top priority since the focus needs to be on largest group and that's the group you don't hear.
Let's focus on the people who can't keep up with the latest and greatest. Because that's the largest group. Let's focus on the people that can spend less then E100 each three months or so on whatever hardware they currently have. Service the needs of that group and you'll have found your "killer app". The "killer app" isn't that what makes people buy the latest and greatest. The "killer app" is allowing most people to make most with what they have. And the fact is, that ain't much. So focus on providing, focus on enabling, focus on making due with what one has. Many small ones still make a large one. It's a perfectly viable business plan. Just think big and small at the same time. Think total.
For Linux the focus should be on mainstream games support because that's what I've been hearing for years now: "if it wasn't for the games I would have dropped Windows years ago".
Start simple. Start with a CD that can also be used to boot a PC into a Linux environment in which your game starts. Add a loader so that those that want to play the game under Windows actually load a little, stripped down, virtual Linux OS in which the game starts.
Most people that currently buy mainstream games wouldn't care less how it starts just as long as it starts. So simply by doing that under (virtual) Linux you would reach several markets without splitting your developer R&D. Don't let bad strategic decisions in the past get in the way of that because someone, someday, will be the first to go mainstream with a more modern approach and they'll hit it big if they go about it in the right way.
Don't listen to the markets. Read their signs in order to understand the larger, underlying, picture. The market are your neighbours down the street and a nation or two away. Expert opinions are an echo of statistical research you read in a published news flash somewhere. There's a difference between the two.
Back in the day, you booted the game disk. The OS and the game were tightly bound and there was nothing else running on the machine when the game had sway. Back then it was because there was no universal graphical API or indeed OS. A good friend of mine calls Windows "The Games Loader". I think it would be a sound business move to switch this logic around and make Linux the "Games Loader".