The Case for Server Side Digital Information Applications with particular reference to ‘Voice Out’

The Client Side Burden

2.0 Introduction

The first computers were, necessarily totally client side devices, supported by armies of technicians. This implicit self sufficiency was, ironically, confirmed by the military use of computing in the early 1970s to avoid a complete hiatus of one installation was hit. Networking dispersed computing power but it was still vital that each node should be self sufficient. The academic use of computing in the 1970s and 1980s involved collaboration but the hardware and software models were primarily client side.

The key event in the evolution of micro computing was the decision of a handful of major companies to manufacture and self-standing PC/Apple general purpose devices with their own site-installed software.

Key client side characteristics:

2.1 Unfitness for Purpose

Key factors in unfitness for purpose are:

Over engineering
Low mean time between faults
Unfamiliar responsibility
Provisional products
Training imposition
Exclusion
Cost instability
Over Engineering. PC/Apple devices were initially marketed for small business and domestic use, the former enhanced in the 1990s by the introduction of servers and networking software. With few exceptions, users only used approximately 20% of the features 'bundled' into systems, could not understand the manuals and were totally technician dependent. While this was an irritant to business customers it was a major problem for domestic users.
Low mean time between faults. Many PC/Apple sales locations implied that the goods were similar to consumer electronics but their performance and mean time between faults was much poorer.
Unfamiliar responsibility. The PC/Apple paradigm is the exception to the general experience of customers who take responsibility for hardware (receiving devices such as television sets, CD players, telephones) but expect suppliers (broadcasters, record companies and telephone companies) to be responsible for the integrity of data.
Provisional products. Many products were released in provisional form requiring modification resulting from user feedback.
Training imposition. PC/Apple devices, together with their associated 'bundled' operating systems and applications cannot be effectively used without extensive training. This is the result of over-engineering, poor design and sloppy functionality and was, in effect, a cost shift from the producer to the consumer; no similar problems have been experienced with digital television receivers and mobile phones.
Exclusion. Even with considerable training it has become clear that client side digital information systems exclude a substantial segment of the population, approximately 1/3 of the whole. By itself this is hardly a problem as many technologies exclude all kinds of people even after training; but the suppliers of information systems which 'shut out' 1/3 of the market were supplying the same systems to organisations with a universal remit such as Departments of Government when it was clear that such systems could not fulfil a universalist remit.

There are obvious specific examples of exclusion, quite apart from the general consideration of a lack of intellectual aptitude or the access to financial support for training, notably disabled people and the elderly with poor fine motor skills and dull vision and hearing.

Cost instability. The economic model for client side computing, unlike most consumer experiences, is so unstable that cost cannot be predicted. Users do not know how long their hardware and software will be 'supported' before an 'upgrade' with inadequate backwards compatibility is imposed. Maintenance costs for an existing system is impossible to predict but this is compounded by 'upgrading' and backwards incompatibility.

2.2 Poor Consumer Protection

Key factors in poor consumer protection are:

Licensing
Servicing restrictions
No protection for data loss.
Licensing. Most operating systems and software are sold under licence which means that they evade consumer protection legislation on such matters as fitness for purpose, incompatibilities, errors (bugs) and malfunctions (illegal operations).
Servicing restrictions. Licence conditions usually contain disclaimers against attempts to remedy faults through unrecognised operatives while licence issuers have often failed to provide support services or have retained the power to insist on expensive licensed operatives.
No protection for data loss. Most licences included no provision for data loss.

2.3 Maintenance Responsibility

Key factors in maintenance responsibility are:

Data quality
Incompatibility
Upgrading
Data quality. Again, maintenance responsibility has been an irritant to small business but a worry for domestic users. People expect to be responsible for devices such as televisions, heating systems and even motor cars but these generally have a long mean time between faults and their design changes infrequently (eg the almost static design of televisions between the introduction of colour and the introduction of satellite delivered services). Responsibility for material coming into a system was a totally alien idea which required but did not receive good marketing.

Consumers were not only responsible for intrinsic 'faults', they also had two additional problems, incompatibility and upgrading.

Incompatibility. Incompatibilities arise when different applications do not run effectively on an operating system or where additional layers of application do not work effectively together. The licences of various operating systems and software provide no proper recourse for consumers where incompatibilities occur. Customers are therefore left with systems that do not work with suppliers refusing to take responsibility, blaming each other and penalising the consumer for using his/her own resources to resolve the problem.
Upgrading. The production cycle of hardware and software also produced the consumer responsibility for upgrading. Technically, a consumer can refuse to upgrade but this means losing the whole investment in a system which is not backwards compatible. As already noted (2.1) one key factor in the upgrading model is uncertain cost but another is frequent loss of performance and additional installation and training costs.

2.4 Arbitrary Invasion

Key factors in arbitrary invasion are:

Provisional product
Junk mail
Malicious damage.
Provisional product. The phenomenon of provisional product and the need for user feedback has already been noted (2.1) and, together with the upgrading cycle, this has led to licence conditions including the licence issuer's right to 'invade' a system and 'upgrade' or 'patch' it. This frequently leads to the incompatibility of the 'upgraded system' with dependent peripherals such as printers and a need for additional maintenance to restore the whole system to functionality.
Junk mail. Whether or not ultimate responsibility for 'junk mail' should lie with the consumer or the internet service provider (ISP), operating system and application writers have signally failed to provide users with protective regimes. Filters within systems are an inadequate response.
Malicious damage. Many digital information systems, primarily those based on the PC model, have been subject to severe vicarious attack through viruses designed to damage Microsoft for which it has taken no responsibility. This is another area of indefinite cost and a threat to business viability.

2.5 The Oligopoly Ratchet

Key factors in the oligopoly ratchet are:

Lack of transparency
Market failure
Unstable cost.
Lack of transparency. Developments in the IT industry make it clear that there is a symbiotic relationship between hardwire and software suppliers: increasing the size and complexity of operating systems and applications necessitates hardware upgrades; such upgrades open the door to yet larger and more complex systems.

Although the business pages of newspapers cover this area in some detail, corporate relationships re not reflected in customer contracts so that 'glitches' between hardware and software are not acknowledged by companies which claim they have no relationship with each other.

This lack of transparency has made it easier for both hardware and software suppliers to deny responsibility for a lack of backwards compatibility.

Market failure. All the problems noted so far in this Chapter would have been overcome or at least mitigated were it not for the complete market failure in both hardware and software. The problem has been even greater in the latter than the former.
Unstable cost. When a customer embarks upon the use of a PC/Apple system the only predictable cost is the purchase of the initial 'bundle'. The following costs are unpredictable:
- Installation (if problems arise using the supplied installation procedures)
- Maintenance/repair
- 'Spam' filtering
- Virus protection and virus damage
- Enforced upgrading and file migration
- Systems failure.

When systems perform poorly, continuing to maintain them is rather like playing poker; there is a critical but not entirely rational decision to be made about whether to preserve or start again. That is why the only parallel to this consumer experience is gambling. From this perspective it is easy to see how any predictable subscription option with producer responsibility (server side) will provide competition for the client side model.

2.6 The Market Model

Finally, the market model for client side systems depends upon upgrading existing customers rather than trying to reach the bottom 1/3 of the market; and so the exclusionist model is economically entrenched. There has been a great deal of literature about a 'Business Case' for manufacturers providing more user friendly and even disabled accessible systems on the basis that the market segment contains 1/3 of the population; but this model only works in the unrealistic environment of unlimited capital, i.e. if there was unlimited capital there would be a 'business case' for making user friendly and accessible systems; but in an environment of limited capital where the quite proper obligation of the company is to achieve maximum shareholder value, user friendly and accessible systems always lose out in the competition for capital to developments which will produce a greater return such as upgrading high value customers and selling new product to people who regarded their systems as partial fashion or lifestyle statements.