But that is a big reason why desktop linux won’t take off. Now if there was guaranteed libraries on a system and there wasn’t so many distros that had all their own little quirks, there would be something that game makers could consider, but alas that didn’t happen.

Personally, I pretty much laugh at all this rabid “how do we get them to switch” business. I usually just run Linux and Windows at the same time.

* Rick
I sort’ve agree that it’s a hybrid OEM/distro I’m describing, but it’s also a lot more. It incorporates things that no other OEM or distribution do in a unified or complete way.

I am convinced that the whole model for what is being offered should change. Take the offering into a new direction, playing by new rules that can’t be accommodated under the existing/legacy OEM/OS sales models.

* EnviroTO

The way you pulled in ISP support is great… It really is an issue. in my opinion, in should be a part of the solution package offered by the “Fictux” sort of company.

* AMDUSER

The quote, “Desktop Linux will take years to achieve acceptance if it just copies what MS and Apple do, it needs to be more compelling…” is great. Virtualization could loosen a user’s legacy dependence in some ways. I’m not sure I can see it as being immediately compelling to a majority of users though. Ideally, a Fictux sort of company shouldn’t rely on that as it puts a kink into the whole solution flow I proposed. Perhaps in the short term it could be an acceptable workaround for a few rare tricky spots. I can’t help but worry that it would introduce more complexity than simplicity for the average user though.

* Ron

You asked “…what basis does a person who doesn’t give a hoot about what is under the hood have for moving into Linux as an OS?” I think, barring a lengthy discussion on the greater philosophy behind it, the easiest, most direct way to convince people, would be to stop providing a straight-up comparison to the systems you can get with Windows or Macs. Without arguing specifics, they’re, in the most general sense, at par. The weakest common spot for the general user is that Windows and Mac systems require the user to spend countless hours hunting for software and services, configuring them, spending money on them, figuring out what new hardware is needed for new applications, not seemlessly working with net services and data backup, etc. for the life of the system and beyond. Change that whole way of doing things to make it easy, unified, and not a waste of the cusotmer’s time. Then you’ll have something compelling in a whole new way that Windows and Macs can’t touch.

See the problem is that you have duplication of effort, no leadership, and no courage to really go out on a limb and take the risks to make desktop linux more compelling.

But virtualization does bring up an interesting mix into the situation. Nextgen hardware will have virtualization built into it, so we might get to the point where you can pop in a CD or download an exe and have a linux system running on near bare metal.

But the bottom line is that cloning Apple and Microsoft’s offerings won’t really move desktop linux into any kind of significant mainstream presence.

In the Enterprise, a user’s desktop is filled with office based applications (MS Office…) and the odd client application. Some of these are entrenched into Windows and prevent a company from moving to a Linux desktop (OpenOffice 2 is still not 100% compatible with MS Office – Impress and Powerpoint for example do have the odd translation issue, especially around images).

However, what is needed is virtualization. Once the user’s desktop has a VM of Windows for their legacy environment then the door is wide open for Linux Desktop to appear on everyone’s desk. Due to being opensource it is of course possible to lock down and minimize the size of the Linux environment and create multiple desktop VM silo environments for say Corporate Emails and for confidential internal applications and even possibly a sort of DMZ VM for the user to browse the internet etc. Much more secure, much more portable (a VM would no longer be locked to a workstation and it’s state coud eventually even be saved on a memorycard for immediate access elsewhere in the corporation).

Desktop Linux will take years to achieve acceptance if it just copies what MS and Apple do, it needs to be more compelling and this is why I believe Desktop virtualization is the way forward as it brings many new possibilities to the desktop which really makes Linux the ideal O/S to sit alongside the user’s legacy O/S and eventually be the dominant system of choice due to having the smoothest migration in the history of operating systems.

Your thoughts?

The main obstacles in my opinion are:
1. awareness,
2. software availability,
3. standardized hardware certification, and
4. ISP support.

Of these key issues the most important is the lack of awareness of what Linux already can do, how affordably it can do it, and how easy it can be to use. I have installed Linux on more than one person’s computer knowing that they did not use their computer in a way that would run into obstacles 2 though 4 and after installing it the response has been something along the lines of “if Linux is so cheap and I am able to do everything I am doing then why are people paying more?” The issue is that there is a lack of awareness amongst people who only use a computer to do internet activities and transfer digital photos or music that Linux does this with little effort. The more tech savvy are the ones that know what Linux can do but need it to do more while the less tech savvy don’t know about Linux and are constantly getting viruses, filling their computers with freeware or spyware, and need support from family, friends, and tech support. The software delivered on the computers of these less tech savvy and more casual users of the computer plays a part in this because they stick with what they have and they really don’t feel comfortable with switching.

The second issue is software availability. For people using a computer to get work done there are some missing pieces. Linux now has a good office suite and a great web browser but personal and small business financial software, graphic software, and web development tools aren’t yet to a level of polish expected by many users. There are many pieces of software in the personal finances area making good progress such as Moneydance, KMyMoney, and GnuCash but these are key features missing such as tie-ins to tax preparation software, the tax preparation software itself, and the ability to import reliably from Quicken or Money where people who track their finances currently have their data stored. Most importantly for small business what is lacking is an accounting and payroll solution (with tax table updates) which is as affordable as the equivalents on Windows. In the graphics area it seems like the hole in the offering will be filled very soon with projects currently underway such as Xara and Gimp although some focus probably needs to be given to usability for users coming from Adobe products. In the area of web development Bluefish, n|vu, Quanta Plus, and MonoDevelop are making good progress but in web design Macromedia components are common and there isn’t really much offered yet for web developers who have been taught web design using Flash and ColdFusion

Standardized hardware certification is important because one of the things that really detracts from the Linux experience is when hardware doesn’t work and troubleshooting and research is required. People need to be able to go into a computer store and buy computers or components and know if it will work by seeing a Linux Certified logo on the box. This is one of the harder problems to solve when there are so many distributions out there all using different kernel versions, different hardware detection mechanisms, and in some cases custom kernels. Obviously a hardware vendor shouldn’t need to spell out the hundreds of distributions its hardware may or may not work on. Some hardware compatibility versioning is required to deal with backporting features in the kernel, hardware detection methods, etc so that someone can look at a logo on a box and a coresponding version number and know that the hardware will work without effort on their computer running Linux.

Lastly, ISP support is an issue that causes problems with less tech savvy users. Not many ISP support Linux and that is a deterrent to those who need help setting up the internet. In addition, some ISPs have special connection software which makes it an unpleasant welcome to Linux. BigPond and AOL clients need to be provided with Linux to convince users of these ISPs to make the move to Linux.

But I don’t see anything beyond that. I didn’t see you mention any new software development. What I”m saying (and I’ve been using Linux at work for 9 years now, but I’m not emotionally attached to any kind of open source “movement”), is that the current momentum of desktop linux will never be great enough to make the vast majority of people to switch. That until people stop looking at KDE and Gnome as the answers to desktop Linux then there’s really no hope. Those desktops will never give anybody any reason to switch.

What is needed is what I’ll call a “winner”. And what I mean by that is what is needed is a distro/desktop/(maybe OEM too) that will come out on top – where ISVs can say ok, here is our standard desktop linux environment. But that’s just not happening and part of the reason is that everything is open source and various bits can be cobbled together by anybody, which in itself causes fragmentation because it’s hard to target and the installers are all different, and the directory structure layout is different enough on these distros, and there’s no universal installer, etc…

Tom is basically right. OSX is what desktop Linux aspires to be, but OSX is special because even if Apple decided to shrinkwrap OSX for vanilla PCs, you couldn’t just recombine various bits and make your own “Josh’s OSX distro”, causing fragmetnation.

There’s no incentive to put the resources into doing “the right thing” if everybody can just steal your code. A hybrid approach might work, where some parts are open and other parts are closed, but I’m doubtful that a pure open source desktop will ever be able to leap out above the others and become a standard. And a standard is what Linux desperately needs.

Put it this way, if Google decided to hunker down, lay out the resources, take the kernel and X, and build a desktop system then it would probably be the standard. But they would probably keep parts of it closed too. Free as in beer is barely an incentive and free as in source code is irrelevant to almost everybody that uses computers.

If a company targeted toward the home user would offer what Linspire does but stay up to date on software and provide telephone support, then I could tell people to use Linux. Also there needs to be some serious redesigning of Gnome and KDE to make them easier to use and set up so you *never* need to use a command line no matter what goes wrong. One of the things that would make it much easier for the average user to install applications would be to get Klik running reliably on at least one major distro.

Linspire is the closest that you can get to this as I know they do a number of your suggestions like: have a music service integrated, bundle ISP dialers, they sell computers with Linspire pre-installed, you can install hundreds of user selected applications in one click, and probably do more things that I am not aware of.

The big problem is getting someone with enough money and resources to pull this off. I am not sure who could do this. Linspire has the sense – they just need a sponsor with big pockets.

The year of the “linux desktop” will come when someone starts from the bottom (kernel/drivers), designs a real desktop operating system on top of the kernel, and doesn’t rely on upstream providers to control their destiny.

That takes resources and until someone with the resources and commitment comes along you can forget about it.

Apple is interesting because it seems to be striving in a few ways to actually do some of these things. I know it offers a number of useful net-based services, but it doesn’t provide all the steps in an integrated flow. When it comes down to it, no matter how nice its products are, Apple can’t break out of its proprietary model, so it is trapped competing with companies like Dell. It won’t get past that small user percentage because it’s participating in a losing game from which it can’t free itself. In fact, having owned a number of Apple products, I think the company sometimes even takes a stance in what they provide, which may be considered hostile toward the user. Some of the current DRM/iPod issues may provide a good example.

(1) I think if you\’re talking about things like command lines with people, you\’re probably talking to a smaller group of people who already have at least an intermediate background/history with computer usage. Yet, I\’d venture to say that the majority of people using computers have, at best, only a vague notion that there is something called a \”command line.\” So I don\’t really see that as much of an issue, especially since it is trivial to dispel with one or two screenshots.

(2) Publicity is a great point. People always say the best publicity is word-of-mouth. If my hypothetical vendor, Fictux, delivered its solution, I think customers would quickly and enthusiastically spread the word-of-mouth notice of how much happier they are with what they got from Fictux.

(3 & 4) I think this issue may actually be the most challenging. Fortunately for business uses, in most cases, there are FOSS-native equivalent applications. Where the issue is really highlighted though, I think is in the gaming side. There really are not game equivalents in the same way as other applications. Perhaps the best option is through something like Cedega. Unfortunately, even with Cedega, the games issue is still one that prevents Fictux from providing every single detail I recommended in the solution plan. I\’m afraid I\’m at a loss to offer a workaround for this. However, I don\’t see this one point as being so problematic as to foul up the whole thing. Rather I think that Fictux providing everything else would be compelling enough to capture such an audience of customers/users, that it would be the necessary catalyst to Linux growth and adoption, for game developers to finally want to natively support Linux. In other words, presenting the solution could, in essence, bring the change required to ease its remaining problem.

(5) The Fictux sort of solution, done correctly, would not have to offer basic teaching for things like Bastille Linux. It should solve that in advance to requiring the user to waste any time learning about it. A user should not have to spend time on things like that, it will not benefit him/her. It should just work from the start. I do believe Red Hat made some huge progress in this regard with its SELinux integration.

(6) I think this misses one of the more important aspects of the solution. You should not just go to any vendor\’s site and get the option of Linux on its hardware. The current vendor game rules are for a proprietary vendor model (Dell or Apple). More companies offering the option of a preinstalled Linux system won\’t cut it. I think the rules have to be redefined to support the continuous flow in the product/service chain I tried to outline in the article and as only a FOSS-based vendor can really provide. The vendor must ask the options first and make the distribution, the user’s distribution (point 2 in the article).

(7) I agree that wouldn\’t hurt. I think however, that those are only two specific exceptions. Providing the Fictux solution as outlined would eliminate that issue from the start.

I bought a Mac when they went *NIX underneath because I wanted a good unix desktop that didn’t encourage me to spend too much time tinkering with my work environment when I should have been working. Since then, when friends ask for help fixing their windows machines or for advice on a new computer, I always tell them the same thing, “Get a Mac”. Mostly the responses I get are that the user has too much invested in Windows, or less often that Macs are too expensive, even when the product the user has in mind has fewer features and is more expensive than the comparable Mac.

Once people do switch, they’re happy they did and become ardent supporters, but getting them there is a very hard sell. I imagine that that is the biggest challenge for the Linux desktop too.

I would argue that what needs to happen for the Year of the Linux Desktop to arrive would be 1) coporate acceptance and usage (and with IBM having declared some two years ago that they would be fully Linux on the desktop by the end of 2006), and Microsoft coming out with an upgrade unacceptable to many. In which case, the next twelve months may indeed be the year of the Linux Desktop.

(1) There’s still a general view that Linux is about command line work, even though the desktop distros show a different story.

(2) Not many people know about Linux or have misconceptions about it. Desktop distros need more publicity to reach a wider audience.

(3) Windows games. Most enthusiasts still refuse, because they can’t place their games! There needs to be a good way to allow this, but not compromise on Linux security.

(4) Windows apps compatibility. Business users want to jump off, but they need their apps. This point and the previous one (3), need to be addressed.

(5) We need to offer some basic info or teach them how to use Bastille Linux, because we do NOT want to repeat history! As well, we need some sort of video tutorial solution, to teach the basics.

(6) Need more companies to offer Linux OSs as a choice, when buying a pre-built system.

(7) Hardware support. It needs to be wider. We need to write open-source versions of current ATI and Nvidia video drivers.