| Linux Beyond the Desktop |
a guest column by Gregory S. Hopper
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
by Gregory S. Hopper
Imagine a machine that dominated its industry, was produced essentially unchanged for almost 20 years, generated billions in profit, and completely transformed life and business. The PC? No, the Model T. And it went from “wired” to “tired” overnight.
As we explore Linux on the desktop, we can learn from the lesson of the Model T. After nearly 20 years in production, customers started to want different things from their cars. Some wanted performance, some wanted luxury, others wanted safety, cargo capacity, or any of a number of other attributes.
Ford wouldn't give it to them. The company steadfastly stuck to their “any color, as long as it's black”, one-size-fits-all mentality. (Even if Ford had wanted to, though, their vaunted assembly line actually inhibited their ability to offer new products. It was optimized to only make Model T's.) Today, we'd say their business model wasn't flexible enough.
Ford's competitors – mainly the multiple divisions of General Motors – rose to the challenge. They created exciting new products for all types of customers. Cars were faster, lower, and quieter. They came in a variety of models and a spectrum of colors. Rather than one product for all customers, GM's business model was to provide different products for different types of customers.
The Model “PC”
A similar situation exists today with the PC. It is a tremendous machine that has transformed our lives, and will continue to meet the needs of many people. But...while improvements have been made (most notably graphical user interfaces, multimedia, and networking), the fundamental design – like that of the Model T some 75 years ago – has not changed.
More importantly, neither has the basic operating system design and license. It is highly integrated, comprehensive in scope, and resource intensive. (One result of this integration is that PC's do not, in the words of human factors engineers, “fail gracefully”. A problem in one program can cause the entire machine to crash.) You have no rights to modify the operating system to meet your needs. You get the entire operating system and all its extended function, or you get nothing. Every end user has to pay for everything, regardless of whether they need or use it.
The capability of today's PC far exceeds the needs of most people, and the gap grows wider with each new incremental “improvement” in chip speed. The fundamental design is 20 years old – an amazing achievement in some respects. Processor and bus speeds increase, power supplies are beefed up, fans are added to cool the machine at the cost of additional noise, hard drive capacity dwarfs what needs to be stored – and what benefits do people get for this progress? Do they surf faster? Write email faster? Compose documents faster? Generally, the answer is no. One might conclude that the PC is the Model T of today.
Today's environment
CIO's like simplicity. For ease of administration, they create and deploy a “single software image” across all desktops in the organization. The “single” image doesn't stay single for long. People, being who they are, bring programs from home, install screensavers, and download programs from the Internet. This can (will) create a support nightmare.
Productivity software is far more complex than most people need. User interfaces become more difficult to navigate as function is added, creating training costs. As products get “upgraded” with more function, the user interface changes accordingly, requiring “upgrade” training.
Then there is the somewhat bizarre “Total Cost of Ownership” story. The cost of owning a business PC through its lifecycle has been calculated to be around $10,000. PC vendors actually compete on the basis of comparatively lower TCO. The basic message: “Buy from us – we give you less of a bad thing.” (Imagine this scenario: a small business budgets $50,000 for 50 PC's, and then learns that each PC will actually cost $10,000. How many of those 50 PC's won't be bought?)
Open Source: Linux and the General Public License (GPL)
CIO's are beginning to rethink their desktop strategy. IT departments are looking at replacing aging PC's bought in advance of Y2K. Microsoft has made recent licensing changes and activation requirements that are more restrictive, more expensive, and widely unpopular. In parallel, the press has been filled with increasing coverage of open source software (OSS), and Linux in particular. This coverage serves as evidence of the viability and maturity of both the concept of OSS and the various Linux offerings.
End user applications – good ones – are available for Linux that do most of what users need. The architecture of Linux is attractive. So are the terms of the GPL. Of some concern to CIO's is the dependency on the open source community for support, but the presence of vendors to stand behind their distributions and the availability of Linux talent in-house are easing these concerns.
But the world doesn't need another OS “holy war”. Simply replacing the current operating system and applications with new ones would be like taking out the engine of the Model T and putting in a different one. Many of the fundamental problems remain.
“Users” are actually “people”
It's been said that the PC business is one of only two industries that call their customers “users”. That isn't very nice. They deserve better. So let's look at things from their perspective.
People value different things about their computer. Some use them as digital music studios. Others value the utmost processing power for virtual environments, like games. Some do complex modeling or design work. Some people write stories, or create presentations. Some just like to do e-mail, surf the web, and instant message. Some track packages, others follow stock prices.
These people don't really care what OS they use, or even what application they use. They don't care if they use KDE or GNOME on the desktop, or CUPS or lprng for printing. All the “choice” of programs in the Linux world is wonderful, but not really relevant to the people who use these machines. They just want to get something done.
People want machines that are easy to use. Ones that start right up and don't need esoteric “shut down” procedures. They want reliable machines that look good and are quiet. This is true in business and at home.
The Curse of the Product Cycle
The PC business is brutal. The vendors all assemble the same basic product out of the same standard parts. Product cycles are insanely short. The chips change every six months, the hard drives every other month, the software every year or two. The number of stock keeping units (SKU's) balloons with each change. Support costs rise with the number of SKU's. These factors contribute to create a commodity business with little differentiation, low prices, high costs, and razor thin profit margins for the assemblers. (The dominant chip and software vendors do much better.)
The short product cycles impact buyers, too. CIO's can't depend on a machine remaining the same during a rollout. Asset tracking is a nightmare. Software compatibility needs to be re-assured with each change. Staff retraining when software changes is another expense. All of this “churn” adds cost but no benefit.
Opportunity: Rethink the desktop
What if the PC industry changed how it thinks of itself? Instead of “PC makers”, they would be “network client solution providers”.
As I've noted, people who use PC's don't all have the same requirements. Therefore, they shouldn't all have to use the same machine. If an executive only needs to do email and access business metrics in a browser, why not simply give him an executive-looking machine with Ximian Evolution and a fully enabled Mozilla browser? (Maybe hook up a Palm Pilot to it, too, for convenience.)
A machine on the shipping dock might be able to get away with just running Mozilla to track packages and schedule shipments. (The Messenger component might be sufficient for the email needs of these people.) All these machines should run 24/7, without “rebooting”. Ever.
I'm not talking about thin clients, where programs run on a server and display locally. I'm not necessarily talking about embedded devices, either, but a lot of principles from the embedded space make sense for these machines. I'm talking about complete machines that are functional even when not connected to the network. I envision a hybrid of thin clients, PC's, and embedded systems, where the function is well defined, and the hardware and software are designed together to maximize the productivity of the people who use it.
By confining the scope of possible tasks done on a given machine, that machine can be configured and tested to be perfect. It might not ever need to be changed. If our executive's machine is good for e-mail and browsing today, it should be good for e-mail and browsing for a long time. If something comes along where the machine isn't sufficient (say, new plug-ins), then IT (or a vendor) builds and tests a new system image, and the client system is updated over the network.
There is also the opportunity to create optimized machines for certain job categories. For example, an IT department at an insurance company could create a version of the “executive machine” for adjusters -- it would run the estimating program, collect digital photos, and securely email them off at the end of the day (or when it passes a WiFi hotspot).
Linux provides the technology to create a galaxy of new clients that do what people want, reliably and consistently, at a decent price.
Vendors could create differentiated products with longer lifecycles. SKU's would decrease, even as the number of models increase. Hardware companies could reclaim control over design points, regain customer mind share, and keep profits.
CIO's could order machines that did exactly what they needed, rather than adapt a machine and train the staff how to use it. Deployments would be easier. Support costs would drop. The Total Cost of Ownership would start to resemble the purchase price.
That sounds like a win-win-win for Linux, hardware vendors, and the business community.
Other events can drive change
One gradual, but significant, change happened in the middle of the Model T's production run, and accelerated toward the end: roads became paved.
That is really what created the opportunity for the new design points in cars in the middle '20s. Deep ruts in the mud were replaced with smooth, hard roads.
Broadband access to the Internet, easy and inexpensive (and wireless) networking, and the maturity of Linux and the GPL are creating a similar opportunity for client devices. Think of them as the “paving” of the Information Highway. Web services are becoming more widely available through Virtual Private Networks (VPN) and browser enabled applications, providing more reason for non-PC devices on the desktop.
Thinking beyond the desktop opens up new possibilities for the PC industry and its customers. New devices with new designs can provide machines that would be easy to use. They would increase productivity by reducing training and support costs. They would revitalize the hardware business – the network client solution space.
We need these new devices. Linux gives us the tools and technology to build these new devices, and the GPL enables us to do so. But, please – any color but beige!
About the author: Gregory S. Hopper is the founder of Netanium Strategic Consulting, LLC in Raleigh, NC, specializing in creating marketing strategies for emerging technologies. He is an original cofounder of Netpliance, and managed several strategy and development groups at IBM's Personal Systems Group. He serves on the Board of Directors of the Visiting International Faculty Program, and the Advisory Board of AnyWare Radio. He can be reached via email.
Copyright © 2003 by Gregory S. Hopper. Reproduced by DesktopLinux.com with permission.
Talk back! Do you have comments or questions about this story? talkback here
(Click here for further information)
|
|
|
|
|
|
|
|
