This is all fairly top-of-the-mind, and I am certain I am leaving things out, but here are the challenges I see:
- Too many Linuxes. The desktop customer doesn't want to worry about different--sometimes very different--versions of the same OS. This only highlights compatibility and complexity concerns. And business customers don't want an operating system called Mandrake or an interface called KDE.
- The Linux community needs to learn to speak the language of the common user. And I mean the very common user. Education is key--for everyone involved.
- Getting the Linux community to rally around a single "preferred" Linux distribution and desktop would vastly improve the marketability of the operating system to desktop customers.
- The "free" software model turns customers off. They don't understand it. It sounds like communism. How do we deal with this?
- We'd need a major (by that I mean top 10), Intel-based, desktop hardware vendor to ship this preferred desktop Linux as a second operating system on all its machines, offering customers a very real choice. Those who preferred Linux could order the machines without any Microsoft content and save some money.
- Some people say the lack of Microsoft Office for Linux is a major limiting factor, and it is. Having Microsoft Office has certainly saved Apple from oblivion, but it hasn't really helped Apple gain market share. Microsoft is highly unlikely to develop a Linux version of Office. I suspect they are also having considerable problems with the idea of building a Mac OS X-native Office because it would essentially be a UNIX port. That could start Microsoft down a slippery slope--especially if some future court ordered the Mac OS X Office ported to Linux.
- Running Windows atop Linux would be a good thing, especially if it can be made to work as well as the current crop of tools for running Windows atop Mac OS 9.1. This is more a propaganda victory than anything, but it would provide support for apps that will never be ported over but users feel they need.
- Customers need to see important software developers announcing Linux versions of their products. It doesn't really matter how much free software exists; we need popular commercial apps.
- The more lines-of-business apps--the sorts of things that run vet clinics, law offices, and small retail shops--we can get ported to this preferred desktop Linux, the better its chances for success.
- Improving Linux's support for multimedia is critical. Linux must do all the important things Windows does.
- Customers, especially corporate customers, are sheep. They don't want a second operating system, even if it is more stable, easier to use, and cheaper to maintain. They have had this opportunity before, and they have rejected it in favor of Microsoft.
- Getting people to change what they are doing requires an order of magnitude of perceived improvement just to get people interested. How can desktop customers be convinced Linux offers this?
- WordPerfect for Linux seems doomed. StarOffice is likely to continue, but I am not sure I would want to convert a bunch of desktops over to Linux based solely on Sun's promise to keep StarOffice both current and available for Linux.
- Even if Microsoft is broken up, I can't imagine any of the new companies supporting Linux. Why? They will still be staffed by Microsofties, and any non-Windows OS will trigger an autoimmune response for a very long time.
So what's the best hope for desktop Linux?
The only real way I see Linux becoming anything like a common desktop operating system would be for Microsoft to endorse it. Figure out the likelihood of this and you'll have a fairly precise measure of the chances Linux has of becoming a real desktop player.
That doesn't take away any of the good things about Linux--it may indeed be a very fine desktop OS--but it will never be a common one.