In this DesktopLinux.com guest column, Gregory S. Hopper takes Linux beyond the desktop. Declaring the ‘PC is the Model T’ of today, Hopper makes a convincing argument that disruptive technologies and changing business models will dictate the PC market of tomorrow. By adopting a one-size fits all approach, the market has not yet fully realized the opportunity that open source, Linux, and the GPL offer to a dynamic user base.
Linux Beyond the Desktop
About The Author
Eugenia Loli
Ex-programmer, ex-editor in chief at OSNews.com, now a visual artist/filmmaker.
Follow me on Twitter @EugeniaLoli
Thanks to that quote I now understand the T in tungsten|T.
Here in Europe we need to get acquinted with things like the Ford T.
—
http://homepage.mac.com/softkid
The man is right, I have very basic nead. open office, web browser, e-mail, download music
no wonder linux desktops look weird in terms of color scheme
for the time being, linux is everything but a platform with a decent browser that doesn’t segfault – about win98 + IE6
people have tasted the freedom of running whatever they want on their PC’s and i dont think working in an environment where the IT department tells you what to run and not run will create a nice work-atmosphere.
That said though, at my place we run sparcstations (Ultra10) that are jumpstarted by the IT dep (i have no root access). the IT dep provides us with the stuff they think we need. But still i can compile and install things in my homedir. To me this is kinda the best of two worlds. they keep track of the enviroment they provide and i get to install my toys in my homedir
“Here in Europe we need to get acquinted with things like the Ford T.
”
I strongly recommend reading “My Life and Work” by Henry Ford. The
parallels with the computer industry are highly interesting,
especially the steady reduction in price.
An important point made by the author is that the hardware and software industry is dominated by a few. It is not the features that cost, but excessive profit margins. To sell many copies of the same product makes more sense economically. And if you are big you might just get away with it. This is the advantage of the one-size approach. But it only works if you dominate the industry.
Will industrial structure change? Maybe, but not in the nearest future. Until it changes, there will be little incentive for Dominators to offer more custom made solutions.
Of course there will always exist options, but they will be more expensive than the standard option. You can get your sunglasses at a gas station or get a pair of brand new RayBans: both provides basic functionality, to protect your eyes from sunlight, but at different prices. Go for cheap standard or expensive custom made.
I admit that Linux is somewhat special since it is opensource and that you can get it for almost nothing. But this only applies to standard Linux. It is easy to make a default installation of Redhat or Mandrake, but if you want a more custom made solution you have to hire a skilled admin. So you choose between going for the cheap standard option or paying extra for the custom. Pay more to get less – not a good combination.
So the challenge for all you custom system-builders will be to change that equation – make your customers pay less for less features. Then you will succeed. Unfortunately, today there is a tendency for the opposite. Those new easy-Linux distros (Lindows etc) are rather expensive and offer less than cheaper Linux or Windows systems.
I’m not really sure how what the author proposes (to the extent that he proposes anything) is that much different than what we have now. He wants a general purpose machine that can be loaded with a variety of softwares.
He juxtaposes “Total Cost of Ownership” with the operating systems and software, but fails to explain that most of the TCO is support costs, and these support costs are best reduced through standardization. It may sound like a wonderful idea to have customized machines, but that requires more specilization on the support end, which is what drives TCO up.
And this is one place where the similarity between PC and the model T starts to break down. For example, if you have a user who never needs to print, removing print services rom his/her machine doesn’t make it demonstrably better. An OS that can do everything for all the users, even if it means that, say, 20% don’t need all the features is often the best solution from a support standpoint than customizing each machine.
Another thing we have to keep in mind is that computers are a tool, not a commodity product, so business will willingly pay more for a machine if it gives greater utility for the money.
In many ways, a Linux loaded PC resembles the model T just as much or more than a Windows one. Both are kludgy in their own ways, both rely on the WIMP interface (not necessarily a bad thing), both have difficulties in their use that need to be overcome with some sort of specialized support.
But what are supposed to do, run “thin network clients”? Bwahahahhaha.
Maybe what he’s talking about is more like this?
http://byzgl.sourceforge.net/
He juxtaposes “Total Cost of Ownership” with the operating systems and software, but fails to explain that most of the TCO is support costs, and these support costs are best reduced through standardization. It may sound like a wonderful idea to have customized machines, but that requires more specilization on the support end, which is what drives TCO up.
And this is one place where the similarity between PC and the model T starts to break down. For example, if you have a user who never needs to print, removing print services rom his/her machine doesn’t make it demonstrably better.
You’ve obviously never administered *nix on a corporate desktop. You intall the bear minumim and then you can chuck people packages down the network willy nilly, they install in their home directory and they carry their apps around the network with them. A new package or an update comes out, people install it when they use it – you don’t need to send 20 support staff to each computer with a CD to do it manually! Its so easy it makes you scream when you go back to administering Windows and having to install Outlook 2002 on 400 machines!
You’re too stuck in the world of Windows – when you leave it you realise anything is possible!
The kind of Unix environment you describe (like the one I work at, incidentally) depends on all the machines running the same OS to work. In the kind of hetereogenous environment the author describes in his article (the very opposite of the typical corporate UNIX network) you couldn’t have the apps follow the users around because there would be no consistency from machine to machine.
And it seems you’ve never worked in a well-run Windows environment. With tools like Castanet Tuner, SMS, and others, many software installations can be done remotely. And why on earth would you send “20 support staff to each computer”? That would make for some crowded offices!
How much do those tools cost for Windows? They are included with most (Li/U)nix flavors.
In the article he equates todays PCs with the Model T cars of the past.
I don’t think the PC is the car … the PC is the road.
It is the invention that is letting you run the things you really want to drive – the applications. So when he is saying that the PCs are trying to be one-size-fits-all, and that the Model T shows that this approach doesn’t work – he’s missing it. The PC (OS) are trying to create the standards so that whatever car (application) you want to run can do it.
The PC and OS needs to build an environment so that other things can operate. That’s why our roads are built similarly. Lanes are sized so that vehicles of different sizes can work, overpasses and bridges aren’t built so that only a car can get underneath. They try to find the environment that suits the best needs of most.
Yes, there are functions in every OS that aren’t needed by everybody. But they’re there so that they are accessible by anybody. The road is there so that any auto manufacturer (software developer) knows that certain things can be counted on.
Some people only use web browsing and email. I use them, plus do graphic design, video, and audio editing. The customization is done by how many cars (applications) we each have in the garage. We both benefit by having roads pre-built.
Or those free as in “beer” or “speech” or “I wouldn’t pay for that in a million years.”
Seriously, Windows and Unix are so different, that nothing like there’s no correspondence between the tools. Windows will never be as configurable remotely as is Unix.
At the same time, you can’t really group together Linux and Unix in this categories, because Linux is just a kernel. You could have a “Linux shop” where each machine was so different that the kind of remote configuration UNIX admins take for granted wouldn’t be possible.
And then, who knows….you might have to write a check.
Perhaps somebody would, but *I* wouldn’t. 😉