Digressions & Diatribes

by Kort E Patterson, Editor

Virtual Service Networks

Virtual Services and Networked Computers "for those who want to use a computer in their work, not work at using their computer".

According to the press releases of the tech industry, the potentials of personal computers continue to expand. Where once they were largely limited to word processing or number crunching, computers have today become gateways to exceptional quality multi-media entertainment, vast stores of on-line information, interactions with the global connected community, and an increasing variety of productivity tools.

However, the expanding potentials of networked computers also have a major downside. While computers are beginning to offer functions that would be useful to the general public, they're also becoming too complicated and technically demanding to be usable by the "appliance challenged" masses.

The administration and maintenance of modern networked computers requires a very different skill set than is required to use those same networked computers for productive tasks. The need to maintain specialized low level system skills distracts users from the high level tasks they intend to accomplish with their computers. It has also become a significant barrier to participation for those who have not yet joined the digital age.

I've been conducting techno-social experiments over the last several years, exploring the potential for providing individual users with on-demand virtual system administration and tech support services through secure virtual networks. The results of my experiments suggest a business model for a global virtual network of locally owned and operated small businesses that could open a new market for digital goods and services that is at least as large as the current customer base for computer products.

The key to unlocking this untapped market potential will be the broad deployment and consumer acceptance of the remote delivery of the support services required to provide the user with black-box "just use it" networked computer systems/services.

While it's admittedly an imperfect fit, the evolution of the automotive industry provides some useful similarities. I think it's fair to say that the majority of drivers on the road today are far more interested in driving their cars than in working on them. To most modern automobile users, a car is a black-box that performs useful functions. They may use their black-box on a daily basis, and even depend on it for their very survival, but they neither know nor want to know much about what goes on inside the box.

A substantial investment in training and tools is required to keep an automobile running properly over its expected operational lifetime. An early automobile user had to be a fairly competent shade-tree mechanic to keep his pride and joy running. In order for the automobile to became an integral part of our modern world, it had to be transformed from a tinkerer's obsession that was occasionally available for transportation, into an easy to use appliance that could be reasonably expected to run at the turn of the ignition key regardless of the weather, the time of day, or the phase of the moon.

Drivers have grown accustomed to thinking of their automobiles as "just drive it" black-box machines. If their pride and joy doesn't go when they turn the ignition key, they turn to a specialist with the knowledge and tools required to match their car's real world performance to their expectations.

An auto repair garage functions as a distributed-cost service. The garage's investment in knowledge and equipment relieves the individual users of the necessity of individually knowing how to keep their cars running. By applying the same investment to multiple automobiles, the garage distributes its total cost across its customer base. The greater the number of customers, the lower the percentage of its total costs the garage must recover from each individual customer in order to be economically viable. The pervasive presence of automobiles is testament to the viability of the distributed-cost service approach.

I've spent nearly twenty years as a software developer and advocate for making machines smart enough to be easy to use. During my nearly two decades "on the inside", I've become increasingly convinced that most of the computer industry is missing the point.

I realized early on that my software was only part of the service needed by my customers. A computer program isn't a static durable art form like a painting or sculpture. Most customers don't buy software just to own it, or to display it on a wall. Their intention is to purchase the functionality offered by the software – they buy software in order to use the software to serve other objectives.

When a user pulls up a webpage on his screen, his objective isn't to accomplish the technical details of transmitting network packets or displaying pixels on their monitor. His objective is to experience the contents of the webpage. All of the networked computer technology involved in delivering that experience is just the means to the desired end.

The essence of what most customers are seeking is a package of hardware, software, and human factors combined into a service that can accomplish their objectives. The real value of that service becomes the ratio of its ability to achieve the desired objective vs the total costs of the service.

The integration of networked computer services into small business and private households has slowed because the market segment with sufficient perceived need to justify the substantial learning curve and operating overhead costs that are currently required, has already been largely exploited. The monopoly dominated personal computer industry today has become largely dependent on selling "new and improved" replacement computers and software to those customers willing to continue paying the high added costs of an incomplete service package.

The remaining untapped market segments will only be receptive to the use of networked computer services if/when those services become cost effective (in other words, cheap and easy). They're unwilling or unable to make the investments in knowledge and tools that are currently required to access those services. Networked computers must become robustly reliable "just use it" black-box appliances that can be operated with minimal knowledge beyond basic language literacy, before they will achieve their full potential, becoming as comprehensively integrated into everyday life of the average individual as the "just drive it" black-box automobile.

The traditional free market solution to this situation has been the creation of a service industry to make the needed investments in knowledge and tools as a distributed-cost service to the user. The major obstacle to the traditional free market solution is also the primary source of the problem – the market manipulations of an abusive monopoly whose short-term interests are diametrically opposite the long-term interests of individual computer users.

Offering my software as a "pay-per-use" service package puts my company in the position of having to provide whatever tech support is required to deliver the desired service to our customers. We can't just talk at our customers until they're willing to hang up – we have to hang in there until the problem is solved and the customer can actually make productive use of our software. After all, if he can't use our software, he can't make any money – and if he can't make any money, we can't make any money. Our customer's problems become our problems.

As the front line tech support resource of choice for my customers, I see the full range of problems my customers encounter in their day to day operations. Over time, my company's tech support load has degenerated into mostly dealing with the endless disasters my customers suffer with a product I've taken to calling "windoze" by a company I now call "microscam". This off-loading of the real costs of the actual service expected by microscam's customers onto third parties like my company, has forced me to investigate alternatives to microscam's badly designed products for my customers. In the process I've come to understand that most of the obstacles to real progress today are the legacy of the abuses and failures of the past.

The first microcomputers were dismissed as toys by those accustomed to the "big iron" mainframe computers of the time. As microcomputers became Personal Computers (PCs), the processing power of these toy computers increased until at least the hardware now offers performance that was previously only available on million dollar mainframes.

However, the hardware is only part of the equation. While the processing power commonly available on today's PCs could support an entire office of productive users on mainframe terminals, the most common PC operating systems consume all of the available processing power just to inadequately/erratically support a single user. A PC operating system that must be rebooted several times a day as it struggles to support just a single user would be even more unusable as a true multiuser multitasking operating system.

Connecting multiple unstable inefficient single user PC's with a file sharing network overcame some of the limitations of using toy computers in real world applications. However, proprietary PC network products contribute substantial additional complications and sources of instability.

The "big iron" solution has been to turn to true multiuser, multitasking, networked operating systems built from the ground up to provide secure, reliable and efficient performance. Unix has been one of the most successful of these "industrial strength" operating systems.

Unfortunately, while Unix can do impressive things once it's up and running, configuring and maintaining a Unix system requires a substantial investment in knowledge and experience. The traditional Unix approach has been to rely on a dedicated IT department to take care of keeping the system running behind the screens, freeing users to focus on productive tasks. Since most small businesses and private households lack the resources to maintain a dedicated IT department, they've had to get by with PC grade operating systems.

The most common proprietary PC operating systems have attempted to improve the reliability and efficiency of their latest releases by basing them on Unix type kernels – with varying degrees of success. Apple has publicly announced that its OSX operating system is based on a BSD Unix kernel, allowing it to openly exploit the potentials of "Unix civilized by a Macintosh user interface".

Microscam emphatically denied that the latest windoze crashed less often than previous versions because they'd built this one on a stolen DEC kernel – until DEC proved it in court. Microscam has also energetically denied the existence of the secret "backdoors" and other violations of user's security and privacy built into windoze that have been documented by critics.

PC toymakers have attempted to address the problem of configuration and maintenance through simple minded "wizards" and flawed ideas like "plug 'n play" that create far more problems than they solve. However, even if the user manages to get his new toy PC running and connected to the Internet, he'll only have moved on to a new layer of problems. Many of the annoyances and outright dangers of computers and the global connected community only exist today because toy PC operating systems are being exposed to the adult real world.

For example, nearly all of the computer viruses that have caused so much damage target microscam products - a company that has abused its monopoly control over the PC marketplace to grow obscenely rich pocketing 84% profit margins, while the flaws in its products force its customers to pay millions for third party add-on products. It's now recommended that anti-virus software be updated at least weekly. Security specialists recommend that those foolish enough to try to use these badly designed toy products in a high risk commercial application should update their virus protection daily or even hourly.

Playing the endless damage-control anti-virus game is of course equivalent to trying to put bandaids on the most obvious symptoms while ignoring the actual cause of the disease. The single most comprehensive and effective anti-virus protection is to simply replace the frail susceptibility of windoze with the robust immunity of Linux. There aren't any Linux viruses. None at all.

I've experimented with remotely supporting various operating systems over the years, and Linux is the only technically viable candidate for a "just use it" machine. Linux is the natural choice for a remotely supported system since in many ways it's internally structured as a network, making sophisticated networking a native function.

Return to Port Of Call Home Page
Return to June/July 2003 Table of Contents