Your mobile computer is a stationary computer

Steinar Kristoffersen and Fredrik Ljungberg

Norwegian Computing Center, Postboks 114 Blindern, N-0314 OSLO, Norway

steinar@nr.no

 

Viktoria Research Institute, Box 620, 405 30 Gothenburg, Sweden

fredrik@informatics.gu.se

Introduction

Mobile computing consistently fails to live up to expectations. Early adopters complain about the size and resolution of displays, awkward input devices and limited bandwidth . There is every reason to assume that the complaints will be exacerbated for mobile CSCW, since most collaborative computing is computationally intensive, often synchronous and visual. In addition, computer-mediated communications require minimal latency in order to be socially acceptable.

Many enthusiasts respond to this challenge by suggesting that new mobile technologies will be sufficiently powerful to meet the needs of CSCW. We assert, however, that expectations and requirements of the user community will proportionally increase by new advances in computing technology. The ante will, in a sense, be upped once more.

This paper provides an initial exploration of this problem, which represents a fascinating challenge given that today’s mobile devices are immensely more powerful than the desktop computers of yesteryear, but at the same time they are a far cry from what users want today. We believe this challenge is fundamentally conceptual, rather than technical; today’s mobile computing paradigm is simply not suited for handheld CSCW!

We have launched a research program, which aims to develop the conceptual foundations for mobile IT-use and CSCW more thoroughly. First, it is necessary to specify exactly how mobile IT-use differs from stationary computing. Next, a reference model is needed to guide fieldwork and provide a common vocabulary for designers.

Explaining the desktop ‘bias’ of mobile computing

Based on the premises offered above, the following argument can be given: Mobile IT is inferior to stationary computing in terms of performance and bandwidth. For example, bandwidth is limited, keyboards are awkward and there is no desk on top of which to put the devices when typing.

At the same time, mobile IT design is clearly stationary ‘biased’: de facto industrial standards have adopted the desktop metaphor and offer ‘pocket’ versions of familiar office applications. For example, there is Pocket Word (without styles) and Internet browsers (without support for Java and plug-ins), which, in addition, hardly display content but one line at a time.

It seems like this is a conceptual, rather than a technical cul-de-sac, since "users’ needs", in these terms, will always exceed what mobile computing can offer. But as the following argument demonstrates, this can be conceived as a result of a naïve design paradigm.

Is advanced document management or internet-based multimedia publication purposeful operations for mobile workers? Are there mobile use contexts where a typewriter-metaphor based terminal with a connected keyboard and screen will be useful at all. Consider the example of electrical maintenance workers equipped with a desktop metaphor-based device. When operating the device as afforded by its design, i.e. sat down in from of the user in one arms length distance, the objects of work, which are switches and power cables in a roadside cabinet, cannot be reached. If the device, on the other hand, is put down on the only other flat surface, which is the top of the cabinet, then the display cannot be seen when squatting to reach the switches and cables.

We assert that this is not an extreme example. Working with mobile consultants, journalists, surveyors and inspectors (typically mobile workers according to any definition) such work situations continually occur.

We suggest the following argument for coming to grips with the problem: The actual operations (achievements/purposeful action) of computing is the combination of functionality (provided in) a use context. For instance, ‘printing’ is the operation that one aims to accomplish, whilst ‘mapping text to postscript to driver’ can be seen as the necessary functionality.

The use context of stationary computing is stable. Technology is a central part of the use context. Within a stable use context, there has therefore been a tendency of confusing the functionality offered by the technology with the (achievable) operations. In our example, ‘printing’ becomes synonymous with ‘mapping text to postscript to driver’.

An important aspect of the use context is modality. When the modality changes, the use context also changes, and thus, the space of possible operations. The operations of one use context therefore differ from the operations of another use context. For instance, the mobile worker is likely to want to print documents at remote sites, for which printers the device is not prepared with the correct drivers (assuming that the cables are the same or all printers have identical infrared ports, which is, indeed, optimistic).

When the modality changes, and thus the use context, the functionality needed to achieve the desired operations are also likely to change. Accordingly, in mobile work, since the use context continually changes, the functionality needed often varies throughout sessions and it is likely to often be different from the functionality offered by desktop computing. The operation of ‘printing’, in our simple example, can no longer be supported with mapping text to postscript to driver’ Instead, what is needed is either ‘transfer text to local stationary computer for printing’ or ‘connect to a network site, which offers the relevant drivers for download’. The latter probably also requires one local stationary computer first to be equipped with file/program installation services for the handheld device, before the drivers can be downloaded and installed.

In order to make the mobile computer support mobile work it has had to be turned into a desktop computer by connecting it to the stationary network and funnel its functionality through a hosting PC.

The desktop metaphor in mobile computing is, thus, a symptom of the confusion of operations and functionality in a stable use context. Only when the use context is constantly changing, like in mobile computing, it becomes apparent that this contributes to a problematic design paradigm. In order to open up an alternative design space (or at least a place to turn), an alternative conception of the relationship between computer-mediated services and continually changing modalities is needed. We are currently developing a model that responds to this challenge by explicating the nature of mobile IT-use.

Modelling mobile work

The reference model for mobile work builds on fieldwork and discussions. Its core concepts are:

A mobile setting comprises at least one mobile session. The following figure summarises the model in relaxed UML notation:

Figure 1: Modeling mobile work

We have, briefly, identified and described three important modalities of mobile work:

Certainly, these are only ideal types; nevertheless they represent a useful conceptual topology which distinguishes mobile work from stationary work and contributes to developing a new research agenda for handheld CSCW. The following figure illustrates the changing modalities of mobile work:

Figure 2: Modalities of mobile work

Services and modality constitute central components of a mobile session, which due to its mobility is far more contingent and dynamic than mobile work. Thus, it points in the direction of new items for a mobile informatics research agenda when these dimensions are combined, such as in the following figure:

Figure 3: Changing services and modalities

Implications for research

Building on the framework outlined above, the following items for a mobile computing research agenda can be outlined:

Initialization

Main objective: Develop models and systematic support for establishing sessions in a mobile setting.

Some selected research problems:

Transformation

Main objective: Develop models and systematic support for sustaining services between modalities in a mobile setting.

Introducing the notion of changing modalities enhances the mobile informatics research agenda with an interest in how people can change between modalities whilst remaining connected to current services.

Some selected research problems:

As this list shows, some of the research problem that can be described as transformation from the originating terminal's perspective, are likely to be perceived as adaptation problems on the side of the receiving terminal.

Adaptation

Main objective: Develop models and systematic support for seamless use of new services within modalities in a mobile setting.

In this opposite of transformation, research interest is in how new services (for which the current modality may not be suited) can be adapted to function in a satisfactory manner.

Optimization

Main objective: Develop models and systematic support for improved performance of sessions in a mobile setting.

This item on the agenda is already well covered in the mobile computing area. There seems to be a tendency, however, of assuming that mobile technology will be almost as reliable as its fixed counterpart. We believe that more work is needed on how to complete business-critical tasks using mobile systems in case of unanticipated technological breakdowns.

We have selected one other area in which more work is definitely required: Simple input devices for mobile sessions.

New applications

We aim to combine empirical studies with the conceptual framework of the model to produce innovate, mobile-aware applications. By technical experimentation and empirical evaluation, such applications can inform re-design and improvement of the model, and, thus, the research agenda.

The remainder of this document briefly outlines some application ideas for extended mobile computing research.

MOTILE (MObile Tactile Input for a Lightweight Environment)

This application should offer the possibility of simple, tactile input for mobile devices. It can be conceived as a virtual keyboard, operated only with three buttons, and relying on the shared functionality of the network and target applications for advanced operations. A small prototype has already been developed.

SEA (Secure Easy Access)

This application should offer secure and easy user registration, directory services and authentication from remote environments that have not been prepared in advance, e.g., using web-based protocols.

MIC (Multimedia Interaction Console)

Managing the interaction itself is an integral and potentially heavyweight task for users of collaborative multimedia applications. This task is made genuinely more difficult in a mobile setting when the user may have to move between different environments. This application should offer a general console for operating and adapting mobile multimedia applications.

BEE (Best Effort Engine) Server

Since mobile environments cannot guarantee access or performance, whilst at the same time mainly hosting business-critical applications, we propose an application that performs rule-based scans of the user's local environment, in order to offer a distributed cache that can be accessed whenever the mobile system can get hold of it. The underlying idea is for no mobile host to ever run below its maximum capacity. The available storage space and processing power should be opportunistically used to anticipate problems following a breakdown in infrastructure or functionality.

Finally, one important rationale when selecting and outlining the four application ideas for development and testing in an experimental infrastructure is that, conceivably, they can also be offered as general components for further use in application development.

Conclusion

This paper has outlined a new research agenda based on a radical reformulation of the mobile computing design space. The core of the argument is that mobile computing is characterized by having to manage continually changing use contexts, whilst stationary computing deals with one fixed use context only. The operations of any computer are products of the artifact’s functionality and the use context, a point that is easily missed from stationary computing development since its use context is so stable. Thus, operations have been confused with functionality. When requirements and expectations of users naturally is determined by the performance of desktop computers, immensely more powerful than mobile computers, mobile computing designers have ended up trying to offer exactly the same functionality as in desktop computing, missing completely the point that their operation is so different. We assert that the offered functionality of mobile computing have to rely on an elaborate comprehension of the mobile use context, rather than the functionality of stationary computing.

Bibliography