Software Technology

---

Reductionist Culture and Instrumental Technology Perspectives

There are several possibilities in studying the relationship between culture and software that can be based on different perspectives on culture and technology. The viewpoint of most software designers and engineers can be tracked to both (1) the reductionist understanding of culture as language and other symbols, and (2) the instrumental theory of technology. In this context software is culturally neutral and can be adapted to every culture through the modification of its user interface. The modification, known as a locale, is then defined solely in terms of the requirements regarding the interface.

Consequently, the collection of symbols and conventions that characterize a particular culture or user community (e.g., transliteration, hyphenation, spelling, numbers, currency, time and date, colour) is perceived as being sufficient to adapt software to the given cultural context (Taylor, 1992; Sauter, 1997). This perception has resulted in the existing internationalization methods used for translating software from its source market to the target markets. Current internationalization methods enable the localization of user interface elements. These methods, while based on three presuppositions, namely, (1) the choice of character codes, (2) the use of locales, and (3) the use of resource files, led to the software internationalization architectures (Hall and Hudson, 1997).

The software internationalization architecture that is rooted in the reductionist perspective on culture and the instrumental theory of technology is equivalent to the user interface localization architecture. An example of user interface localization is given in Figure 1. Comparative studies help determine the set of primitives that are common to all cultures in which localization is considered. Primitives which are specific to a given culture (e.g., characters), interpretations of common primitives (e.g., of different colours and icons) and other constructs comprise this culture's locale. These primitives are used to define a common graphical user interface (GUI) toolkit.

Figure 1: User Interface Localization.

The abstract concept of the interface, indicated in Figure 1 as Text Window <<abstract>>, is a generalization of all possible user interfaces. To avoid extensive studies of different cultures, a reductionist approach is used whereby the text window is developed in Culture A, and the abstract text window is generalized by ensuring that it can support other cultures' locale. In particular, Culture B Text Window is obtained through the replacement of components of the Culture A Locale with components of the Culture B Locale.

Holistic Culture, and Substantive or Critical Technology Perspectives

In contrast to the above perspective on software internationalization, we propose to consider (1) the holistic understanding of culture, and (2) the substantive and/or critical theory of technology. From the holistic perspective of culture and one of the two theories of technology it clearly follows that software technologies are not culturally neutral. Software internationalization produces applications whose user interface is internationalized and whose core is not.

In Figure 2 we illustrate the results of software internationalization. Software X developed in Culture A is an artefact of Culture A. Cultural elements such as language, graphical symbols and conventions are used in the design of the software's user interface. Because of the internationalization requirements the interface is loosely coupled with the core. Loose coupling (indicated in Figure 2 with the uses arrow between software interface and core) allows for the replacement of an interface localized for Culture A with the interface localized for Culture B. This is an important distinction in the relationship between the user interface and the application core, paralleled by the relationship between surface and deep culture. The latter conjunction is tightly coupled.

Figure 2: Internationalization of Software Produced in Culture A.

The core of Application X embodies deeper elements of Culture A; those are models and procedures which are used to form messages, construct symbols, and create information and knowledge from data. Their specification and implementation is based on deep culture. Some of the models and procedures may be common to several cultures, their configuration, however, may reflect values and beliefs that are unique to a given culture.

One difference between the substantive and critical perspectives is in the impact of technology on society and its organizations. Observe that even in Culture A the production culture will be somewhat different from the deployment culture. (For example, the UNIX operating system was made by programmers for programmers—its learning curve is steep as a result of numerous conventions that are particular to computer programming culture.) Taking this into account we note that the substantive theory states that a new culture emerges through the implementation and use of technologies. This new culture is neither the production culture nor the deployment culture. In contrast, the critical theory states that the production culture modifies the deployment culture. In particular, the use of Application X in B implants values and norms of A in Culture B.

The goal of the internationalization architecture is, in general, to decouple culture-dependent elements from the culture-independent elements. As Hall states, "all the culturally and linguistically sensitive software components need to be separated from the core of the application" (Hall 1999, p. 298). This indeed may be the case but our short review of this architecture and software localization clearly assumes that—as the result of the reductionist understanding of culture—the only culture-dependent component is the user interface. In the following section we briefly discuss examples of applications that were developed for the same purpose. The applications were developed in different cultures, and as a result the user experience that they offer is fundamentally different.

Three Applications

There are many applications which implement the principles of problem understanding, decision-making, and communication. These principles are heavily culture-dependent (Steward, 1992; Lootsma, 1996). As an example, let us consider three well-known schools for problem solving and decision support: American, English and French, and, respectively, three examples of Decision Support Systems (DSSs): Expert Choice, Decision Explorer and Electre. These systems have been used at any organizational level and for every type of decisions, ranging from strategic to operational.

Electre

French culture, according to Hofstede (1997, p. 87), has high power distance and feminine orientation, while British and American have low power distance and masculine orientation. In France this leads to equal concern for quality of life, relationships and authority. High power distance leads to process standardization, bureaucracy and technocracy (Crozier, 1964; Hofstede, 1997). This would suggest support in which the consultant and the system provide technostructure and supply ideas leaving significant degrees of freedom to the decision makers accepting their authority (Mintzberg, 1979).

French Electre is based on the concept of reduction of incomparability between actions through the use of outranking relations that represent decision-makers' preferences. Effort is made to exploit the relations between actions gaining information allowing for the determination of their partial comparability. In this way the software provides support to define a partial ordering of alternatives. While it aims at providing structure and support similar to Expert Choice, it gives the decision-maker control over the results without the necessity of providing a justification. The recognition of incomparability which allows the decision maker to incorporate aspects cannot be measured, explained or even voiced. As a result this method provides greater distance between the consultant, system, and the decision-maker, than the other two methods described below.

Decision Explorer

Organizations in Britain use adhocracy and mutual adjustment in decision-making. The process is often informal and involves people from different areas who communicate in order to define and solve problems (Mintzberg 1979). The informal approach to solve problems through communication allows equal involvement of both decision makers and consultants, who have equal responsibility.

The English Decision Explorer is based on the assumption that the search for the solution of a complex problem is equivalent to finding its appropriate representation. The roots of this software are in the development of cognitive mapping used to map thought processes. Users of the Decision Explorer are often consultants who help decision makers to construct the map of the decision problem. The underlying assumption is that if the problem and its potential implications are well understood, and can be assessed from different perspectives and in varying levels of complexity, then the solution becomes obvious. Decision Explorer concentrates on the manipulation of symbols and logical analysis of the relationships. It facilitates informal communication and requires the involvement of all those who have information about the problem at hand.

Expert Choice

In the U.S. the divisional form and standardization of outputs play predominant in organizations. Consultants have a significant and active role in decision making; it is expected that decision makers can provide all information required to solve a problem and determine an optimal decision. Efforts are made to specify the desired results and use well defined coordination mechanisms in decision-making (Mintzberg, 1979; Hofstede, 1997, p. 152).

The Expert Choice application implements a "measurement philosophy" in which every complex problem can be represented with a structure, which in turn can be compared with another structure. This implies that decision makers can—possibly with the help of a consultant—determine their own subjective measures of goodness (utility) allowing them to compare and order decision alternatives. This seems to coincide with the key characteristics of American organizational culture and the principle of technology dominating the policy and social choices (Zysman 1999).

Expert Choice and many other similar utility theory-based systems assume, by the very notion of utility, a full comparability of alternatives. Decision Explorer alleviates the problem of comparability by supporting insight and narrowing the space of alternatives. The French Electre family of systems attempts to bridge the two approaches in acknowledging incomparability and seeking for comparable elements.

GIS and Other Systems

Claims have been made that formalized approaches to decision-making, many of which form a core of DSSs, do not differ and are not a function of culture (Neganshi, 1979; Al-Jafaray and Hollingsworth, 1983). Similar claims are made regarding geographic information systems (GIS), which—according to Brodnig and Mayer-Schomeberger—are culturally appropriate and provide a fair representation of local cultures (2000, p. 11). They state that, "certain purists consider GIS technology as a tool for epistemological assimilation and as such, as the newest link in a long chain of attempts by Western societies to subsume or destroy indigenous cultures. Spatial information technologies cannot capture... the cultural patterns imbedded in landscape and natural resources." (op cit., p. 11). In doing this Brodnig and Mayer-Schomeberger and several other authors seem to forget that mapping requires decisions about representations in which strong cultural bias has been shown (Harley 1990; Cogswell and Schiotz, 1996). Moreover, GIS and related technologies now incorporate models and procedures for symbol manipulation based on fuzzy logic, multiple criteria decision analysis, utility and optimization. For example, the DSSs reviewed above suggest that the use of utility has been accepted by the American culture but not by the French and English.

Some of the underlying reasons behind these surprising claims may lay in the American culture, which is highly individualistic, low-context, entrepreneurial and oriented towards tangible results rather than relationship, process and deliberation (Hall, 1976; Carmel, 1997; Hofstede, 1997). These attributes are particularly visible in the software industry as a whole, including the culture of designers and developers. The latter may be seen as being "hyper-American" (Keniston, 1997). The above three examples, other studies (Heaton 1998), as well as general studies of the philosophy of technology (Ferre, 1995) indicate that there are significant differences in the software developed in different national and organizational cultures.

---