PART I

IT and Praxis

Richard Coyne

Department of Architecture, University of Edinburgh, Scotland

Sidney Newton

Department of Design Studies, University of Western Sydney, Australia

Sally McLaughlin

The Waikato Polytechnic, Hamilton, New Zealand

This is a report on the dynamic relationship between information technology (IT) and architectural practice. The report summarises the attitudes and opinions of practitioners gathered through extensive recorded interviews, and compares these attitudes and opinions with the findings of other studies. The report is compiled from the point of view of an understanding of praxis-appropriating practice as preceding theory as the model for understanding. We thereby connect what is going on in IT with concepts currently under discussion in postmodern thought and in the tradition of philosophical pragmatism. We identify several of the major options identified by practitioners in their use of IT, including practicing without computers, substituting computers for traditional tasks, delivering traditional services in an innovative way through IT, and developing new services with IT. We also demonstrate how firms are changing and are being shaped by the market for architectural services. One of the major areas of change is in how IT and related resources are managed. We also consider how the role of the practitioner as an individual in a firm is changing along with changes in IT, and how different prognoses about the future of IT in practice are influenced by certain dominant metaphors. Our conclusion is that IT is best understood and appropriated when it is seen as fitting into a dynamic field or constellation of technologies and practices. Such an orientation enables the reflective practitioner to confront what is really going on as IT interacts with practice.

Information technology (IT) includes computers, such as graphics workstations and desktop and portable computers, communications tools such as telephones, cellular phones and faxes, and peripherals such as file servers, printers, plotters, digitisers, scanners and slide writers. It also includes databases, CAD systems, wordprocessors, electronic communications systems, multimedia and other data and programs. Information technology has introduced new dynamics into architectural and engineering practice. The influence of a technology is commonly seen in terms of impact, as though the technology is an isolated phenomenon that makes some particular operation more or less efficient. In this study we attempt to move beyond the concept of impact to address the complexities of IT in practice.

This study focuses on an interview survey of design practitioners that revealed certain practices, allegiances, tensions and contradictions about IT use. There appear to be two conflicting views presented by practitioners. One is to conceive of the impact of technology in theoretical terms, with reference to rigidly defined rationales, the expectation that things should conform to the dictates of logic, and that IT is provided to address well-defined problems. In tension with this view there is the pragmatic view-a recognition of the importance of praxis, of everyday socially situated practices in shaping the way information technology is adopted. The theoretical view masks the plurality of ways in which IT is taken up. In following the pragmatic strand within the survey responses we discovered that practitioners were adopting IT in four ways: (i) some were substituting automated tasks for traditional tasks, (ii) others were delivering traditional services in innovative ways made possible through the use of IT, and (iii) others were developing extended IT-based services. Furthermore, statistics from other surveys indicate (iv) that many architectural practices continue to operate without advanced IT such as computers.

Our survey also revealed that important trends are emerging as to how practitioners regard their market in an IT context. The attitudes of practitioners towards IT reflect existing divisions within the profession as to the role of the architect. Furthermore, a new IT based definition of the architect is emerging: the architect as a generator and coordinator of information.

Our interviews with practitioners revealed a number of contradictory opinions on the impact of IT on the day to day tasks of the architect. Practitioners are divided on the issue of whether the introduction of IT leads to deskilling or reskilling; on the role that technology plays in relation to design activity; and on whether IT facilitates autonomy, is an essentially isolating phenomenon, or reinforces the interdependence of a matrix of practices and players. With regard to management issues raised by IT, new practices are emerging around the issues of day to day management, IT selection, training and customisation.

Finally, the prognosis for the future of IT in practice is defined in terms of certain technical problems seen through a range of "metaphorical orientations." Where the computer is understood as a set of disparate technologies, technical issues are defined in terms of the quest for total integration and centralised comprehensive databases. A further metaphor is that of the computer as intelligent system, in which case the future lies in the codification of human expertise to make computer systems more effective. Where the computer is seen as a communications device, then the challenge is to develop systems that enable practitioners to exchange information and to collaborate. Where the computer is understood as a mass media technology then the challenge is how best to promote the firm with IT, and how to integrate video presentations, audio visuals and computerised multimedia.

This study is of relevance to practitioners in several ways. First, for those who already make use of computing we open up new ways of looking at that technology by drawing attention to assumptions at work. Second, for those practitioners still contemplating the move into advanced information technology we aim to convey a realistic sense of the state of play. Third, the study is also of relevance to those who supply the technology and technology based services: researchers, developers, consultants and vendors. One of the difficulties that has emerged in the field is a lack of appreciation by such "suppliers" of the way that practitioners actually work. This study presents opinions and attitudes evident amongst practitioners. These practitioners are from a range of backgrounds and have a diversity of technical expertise. The emphasis of this study is on elucidating the often hard won insights of these practitioners into the relevance, benefits, pitfalls, and techniques of information technology applied to their practices and their profession.

Unless otherwise indicated all quotations in this article are from interview respondents.

Models for exploring the implications of IT

There are a number of models of how organisations operate. These models impinge on how people understand technology. First, there is the rational decision making model of organisations. Here the emphasis is on detached principled decision making. Technology is regarded as a variable in a matrix of variables to be taken into account in making "rational" decisions about the future. Rational decision making does not rule out the possibility of other kinds of decision making (such as intuitive decision making), but it seeks to be clear about the distinction and the priority of each. Second, there is the model of technological determinism, of technology as determining the changes that confront practice (Forrester, 1989). Technology is regarded as a prime cause above anything else-social factors, work practices, etc. These two models seem to be privileged in technical and managerial theory, though not necessarily amongst practitioners themselves. Third, there is the praxis model in which technology is regarded instrumentally as equipment that we use to get things done in some specific work context. The emphasis is on human action in a context, and how technology impinges. On this view the introduction of a new technology, such as advanced IT, will affect and be affected by customary ways of doing and thinking about things. In this study we favour the praxis model as an alternative to the rationalistic-deterministic models that are so prevalent in thinking about technology.

The rational decision making model

The dominant characteristic of the rational decision making model is its grounding in detached analysis. On the rationalistic view, the management of a practice should undertake extensive analysis of the practice's operations, and of the market for their services, prior to the introduction of IT. Management are urged to quantify factors such as the volume of work undertaken, task completion times, and staff development costs. After the introduction of IT, emphasis is placed on "proving" or quantifying productivity gains (or losses) (Dill and Pitman, 1985; Architectural Record, 1990).

There are many limitations to the rationalistic model. First, it is very difficult to assess productivity, beyond what financial balance sheets indicate. If productivity is related to inputs and outputs, then inputs include investment in staff time and capital, and outputs include jobs completed, profits, kudos, greater potential for future work, etc. Both are difficult to measure. Second, the very premise of the rationalistic model-that new technologies are adopted because they increase productivity-has been thrown into question (Forester, 1989). Third, there has been a significant backlash against the application of "rationalistic" principles in business. Peters and Waterman (1982) observe that a narrow focus on "objective" analysis has led to the marginalisation of strategies that make the top American and Japanese businesses successful: customer service, encouragement of employees, making products that last and work, and relying on intuitive evaluations of the market. The evidence of our study is that practitioners do not really follow the rational decision making model, though it features prominently in the way some practitioners talk about their firms and IT.

Practitioners are adept at explaining their decisions in rationalistic terms, and then just as readily betraying that model by talking about the legitimacy of other modes of decision making, such as adopting CAD because everyone else is doing it, or doing what feels right. Tensions emerge between an underlying belief in the rationalistic principle that productivity gains and losses should be quantifiable, and the practical recognition that manual and automatic techniques aren't readily comparable:

It needs a bit of work to set up but once you're set up it goes much better. There are some things that you don't do manually because you haven't got the resources-it's not readily comparable.

There is also an emerging sense that information technology is essential technology, that extensive justification is unnecessary. A practitioner from a government department described this change as follows.

There was a policy once that everything you bought that had to do with computing had to go through a Board of Review (5 or 6 committees) [implying the need for justifications]. Now they regard them as typewriters-they're not interested.

It was clearly evident from our survey that practitioners are making technology related decisions on the basis of situated, context sensitive judgements rather than on the basis of detached analysis. We draw attention to the rational decision making model here as it is a prevalent view that appears to impede an understanding of what is really going on as IT is incorporated into design practice.

The model of technological determinism

Technological determinism may be defined as the belief that our future is being shaped by technology. Forester contrasts "determinism" with "free will," the belief that "we are to some extent free to choose and shape our future" (Forester, 1989, p.2). On the determinist view, the actions of individuals and groups are less influential in change than the enormous impact of technology.

Comments by practitioners provide evidence for the prevalence of the deterministic model. Some practitioners speak of a "natural progression" in improvements to machine speed, photo realism, user friendly interfaces, integrated building models and data exchange standards. Some think that obstacles currently impeding the technology will be overcome with technological solutions that will occur almost "automatically."

[CAD management] is a reducing requirement. In the next ten to fifteen years the increase in the use of AI [artificial intelligence] will mean that the management of systems in terms of making sure that you've got the right layer, with the right things on it, in the right place, in the right file, on the right job, at the right time-you are going to need that [AI]. A lot of the tedious stuff will be dealt with automatically by increasingly sophisticated software.

Within this model it is assumed that the groundwork for CAD has been laid. The future consists of more of the same with the impediments overcome. The incorporation of CAD into practice is progressive and evolutionary. CAD use runs according to a fixed course. As one practitioner put it, "there's no need to predict the future because it's already been predicted for us." In other words the future is determined by the technology.

Determinism is at play in the explanations that some practitioners give for their move into CAD, and into technology in general.

…we thought it [CAD] was inevitable. We wanted CAD as soon as we could get it. We wanted to be early. No matter what the price we figured you had to start.

You will get to a stage where unless you can operate these machines you won't get the jobs.

… you don't have a choice. If you want to survive you have to follow the technology. A lot of people feel that they're being pushed into it-which they are.

Within this model, the inevitable "progress" brought about by technology may be viewed in either a positive or a negative light. For some it is the portent of opportunity. Reading future trends and acting accordingly opens up the possibility of consolidating market share and establishing new markets. For others it poses a threat. They think that there will be change and that they will have to comply with that change. There is a sense in which they are being forced to comply with the dictates of the technology, dictates that are beyond their control.

In its most mundane form technological determinism takes it for granted that there is an incremental continuous path that is "progress." Symptomatic of this "continuous path" outlook is the emphasis placed on the "fact" that microchips double in speed and memory capacity every three years. They have done so since the late 1960s and, according to the incremental technological determinist, will continue to do so into the future. A variant on technological determinism is to hold that technological change is discontinuous. According to this view, technologies introduce radical discontinuities into our work practices. Thus practitioners speak of changes from a labour intensive office to a capital intensive office, from a craft based industry to a production line.

The issue ten years ago was whether you wanted to do it or not. The issue today is how. The issue in ten years time will be those who have and those that haven't. It's just the evolution over twenty years of going from a craft based office to a production line machine based office.

Implicit in such characterisations of radical change are calls to appropriate or adapt to the change; calls to rethink the nature of architectural practice and the implications of information technology.

The praxis model

The praxis model is characterised by its emphasis on everyday practices. According to this model we are socially situated. We are deeply embedded in culturally defined patterns of activity. These patterns of activity shape the way in which we take up new technologies. Furthermore, new patterns of activity, new practices, may emerge as new technologies are incorporated into our everyday ways of doing things.

In the praxis model we recognise that technologies and practices interact in complex ways. According to the praxis orientation the introduction of information technology into a firm constitutes a perturbation in the praxis of the firm. The introduction of information technology disrupts customary ways of practicing architecture. In the process the technology is also undergoing transformation.

The nature of praxis is evident as we consider the relationships between members of a firm, particularly those relationships that cross over the organisational structures and the formal roles of the participants: the tacit understandings of the boundaries to each others knowledge; knowing who to turn to for help; knowing who is reliable and trustworthy; and the teams of alliances that are set up.

Technologies are similarly connected to each other. They form parts of complex technological systems that include: physical dependencies, such as the dependence of computer systems on the electricity grid and the telephone system; metaphorical dependencies (CAD systems are related to manual drawing tools, multimedia is related to film and video, aspects of electronic mail are related to telex and CB radio, and so on); institutional dependencies (through industries, systems of distribution, regulations, training and education systems); and local dependencies (the computer requires the printer, the modem connects the computer to the telephone line, and so on).

These technological constellations are enmeshed with the complex constellation that is praxis: the role of the designer is meshed to the technology of the drawing board or CAD system, and their related technologies; the role of the manager is caught up with spreadsheets, the cellular telephone and the company car; secretaries and administrative assistants are involved with typewriters, wordprocessors and telephones. This interconnection of praxis with constellations of technology is particularly evident if there is a change in any of the components. A simple example is the introduction of word processing, which is gradually changing, and reducing in some cases, the role of the administrative assistant. But the effects are not just in one direction. Wordprocessing has been designed and developed through understandings of typing practice, and the technology is responding to changes in that practice as word processing incorporates more features of interest to management, such as the incorporation of dynamic charts and tables.

The praxis model recognises that new technologies reveal aspects of our practices: CAD reveals architectural practices as custodians of databases on buildings; computerised desktop publishing reveals architectural practice as a kind of publishing business; and multimedia reveals the firm as a dealer in advertising, presentation and even fantasy. So technologies do not merely meet needs, but are implicated in the defining and redefining of practice.

The praxis model critique of rationalism

From the point of view of the praxis model the rationalistic model presents too narrow a view of change. Whereas the rationalistic model emphasises change made on the basis of explicit decision making, the praxis model allows for the notion of change as a cultural phenomenon without a single rationale.

Very few professional designers have changed to date… This will change as a new generation of students come through that see the computer in the same way that the current generation regards a drafting board.

We spend 15 years learning to draw with a pencil-it's a language. You're putting things down for another person to understand your concept of a future reality. CAD just does that in a different way.

The praxis orientation is evident in the way that practitioners talk about taking up CAD on an "experimental" basis in the absence of any clearly articulated rationale.

I don't think that they [the firm] actively made allowances for anything that may or may not have happened. I think they just decided to get into information systems and see how it went and they had a management structure that coped …

Before I joined this company we had been on a bureau to a North Sydney company. We went into CAD in a non-committal way so we knew what we were doing when we finally purchased a CAD system five years ago. A lot of companies bought a CAD system without knowing what they were doing.

The motivation for a practice taking up CAD may have more to do with the expectations of clients and staff (implicit in the client and office cultures) than with "objective" estimates of productivity gain.

… the reason that any organisation gets into CAD is that there are staff that want it and they exert pressure on management. It's hardly ever taken objectively…

… 99% of architects go into CAD for the wrong reasons, the client expects it: "without CAD you're not up to date."

Clearly there is a tension here. The praxis model suggests that rather than dismissing "client expectations" as irrelevant, or irrational, we should recognise the importance of the expectations embodied in client cultures.

The praxis model critique of determinism

The praxis model makes sense of the experience of some practitioners that they are being forced into IT, and that "the future has already been predicted" for them. It does not, however, buy into the deterministic position. According to the praxis-based model action always takes place within a context of communally defined praxis. Responses to information technology take place within this context. Individual practitioners feel the weight of the evolving norms and conventions as they adopt new technologies.

All the firms are going into IT. So to keep up with other firms you have to start using it.

A lot of them [design practitioners] are taking the approach that CAD is coming so we might as well jump on the bandwagon now.

The praxis orientation distinguishes itself from determinism in that it leaves room for the fact that conventions change. In fact it recognises that change is the norm. Not only does adopting computers promote change, but it also reveals change. The adoption of a technology brings change into sharp relief.

People get the idea that once you get computers there's a set way of doing things-people are always changing the way that they do things and the way that you set things up on the computer has to be such that you can adapt to that change. It requires a great deal of management and computing skills to get those things working properly.

The shape of the future is not inevitable; nor is the way in which we will make use of technology.

According to the praxis-based model technology does not determine change. Technology is rather one of many interconnected factors in the constellation of actions that is praxis.

IT Strategies in Practice

The deterministic and rationalistic models foster simplistic readings of the "impact of technology." Commentators on technology put forward a view of the office of the future. It is implied or assumed that the majority of practices will move in this direction.

The praxis model, on the other hand, acknowledges the plurality of possible responses to technology. In the light of our survey of architectural practitioners we have found it useful to classify these responses as follows:

operating without advanced IT;

substituting IT for traditional activities;

delivering traditional services in new ways with IT;

extending the firms services through IT.

The response of an architectural practice to technology is the result of a complex interplay between available technologies and entrenched practices. None of these approaches is necessarily superior to the others. Different strategies will be appropriate for different practices at different times. Furthermore, any individual practice may adopt a variety of approaches. We also assume that a firm may be unaware that it is following a strategy at all.

Operating without IT

A cursory glance at the literature pertaining to the use of IT in architectural practices is likely to leave one with the impression that all practices are using IT, or will be doing so in the near future. While the data available indicates that a majority of architectural practices are using IT there are a significant number of practices that do not make use of the technology. Furthermore at the time of this study only a minority of practices in Australia, the USA and the UK use CAD.

Various studies have shown that CAD has become the norm in large practices but not in small practices. Why have small practices automated their business functions but not embraced CAD? Why are there a significant number of practices that do not use the technology at all? Various explanations are offered.

(i) Smaller practices consist of a fairly high proportion of professionals who are design oriented. It has not been the designers who have been using CAD.

Very few professional designers have changed [to CAD] to date … It's more the documenters and the managers that are taking it on at the moment.

There are a number of reasons why professional designers haven't taken on CAD: they don't understand it or understand what it can do for them, they feel that it will impose on their designs or on the ways in which they have become comfortable with designing, they simply don't like using CAD, they don't like the kind of output you get from CAD, they don't want to be classified as a computer operator, they don't want to be "at the mercy of technocrats," they feel that it will take too much time to learn, or that it will be too expensive. These attitudes are changing as computer systems become more oriented to designers, as graphic design and media people become more involved in the development and use of computers, and as IT becomes incorporated into everyday practices.

(ii) Smaller practices typically have different projects, different clients and offer different partial services than larger practices (Cuff, 1991, p.45; Dean and Olley, 1988). Whereas the clients of the larger practices (corporations and government organisations) expects the practices to have CAD, there is currently no such expectation on the part of the home owner or the smaller commercial client, both traditional small practice clients. Indeed automation may be perceived as detracting from the "personal" approach required by the small practice client (Architectural Record, 1984).

I think most small-office clients are primarily entrepreneurs who own or control their own business … It's one ego massaging another ego, and if we brought in a computer plot program, most of our clients would be outraged because they would feel we weren't giving our personal artistic or interpretative attention to their job. Our clients are terribly interested in their [architectural] programs, and they want to feel that the principals or the people that they are dealing with are working with them on a very personal basis.

The designer in the small practice is likely to be dealing with small scale, one-off projects, with minimal repetition of elements-factors which militate against the necessity of CAD. The small practice is more likely to offer partial services in the area of design or in a specific area of design expertise rather than in documentation or management (Gutman, 1988, p.22).

Interestingly, the primary projected advantage of a small firm embracing CAD is that it will allow the firm to undertake larger projects.

… the era of the backyard operator has arrived. … I know of several circumstances where two competent architects with a CAD system are doing large projects extremely competently and I would imagine extremely cheaply. So it empowers people with the skills to be closer to whatever it is that they want to be closer to rather than hanging around an office.

CAD is perceived as being a means of allowing the small practitioner to move into the kinds of projects that are highly coveted by the profession, without sacrificing the autonomy possible within the small practice. CAD is also regarded as a medium with the potential to provide a seamless interface between practices and consultants. From the point of view of the smaller practice this provides a new opportunity, allowing them to plug into larger projects.

(iii) A final factor at play in the reluctance of the small practitioner to embrace CAD is the extent of commitment required. Large practices can afford to have the technology in place just for those projects that are "good for CAD." CAD does not have to be integral to the overall operation of the firm. For small practices the decision to move into CAD is a commitment of the entire direction of the firm.

If we look at the applications that are taken up most extensively in architectural practices-word processing and spreadsheets-we find that they are very different types of applications to CAD. They are low cost, relatively easy to learn, offer instant gratification in the form of quality of output and flexibility of operation, and they don't impinge on core design skills. In fact applications such as these are considered to be instrumental in freeing up the designer from the more mundane aspects of practice so that she can concentrate on core activities. Furthermore these applications are widely used throughout the business world. They are already part of the small business culture.

Why has the widespread use of applications such as word processing and spreadsheets not seeded an interest in CAD. (According to Crawford and Huchinson's survey 45% of small practices (1-4 people), and 10% of medium practices (5-14 people), aren't using computers.) One factor at play is that practitioners simply do not know what is possible with computers. Olley (1991) found in a survey of 283 offices registered with the RAIA (Royal Australian Institute of Architects) Practice Division that there were many requests for the Practice Division to evaluate software and provide education and training, these requests were from offices that were, in the main, already using computers.

Another factor is the conception of computers and anything to do with computers as the antithesis of designerly ways of doing things. Computers do not fit in to the firm's understanding of its role, and the roles of its participants.

… for a lot of lay people the idea of a computer is completely tied up with the idea of cold, hard-headed rationalism, and this actually prevents a lot of people from accepting or using the technology because they would consider accepting computers to be the same as accepting a cold, hard-headed, rationalistic approach to life. If we could break down the concurrence in people's mind of computing and rationalism, a lot of interesting things could happen with computers in practice.

Finally, there may currently be no need for computers in the smaller practices. The "minor job" represents a significant sector of the architectural market. Furthermore, small practices may choose to specialise as "design architects" (Gutman, 1988, p.102) or offer other specialised service. Traditional office procedures may be entirely adequate in these markets.

Substituting IT based tasks for traditional tasks on a one-to-one basis

The prevalent strategy in practices that have adopted IT has been to substitute IT based tasks for traditional tasks on a one-to-one basis. The typist now uses a word processing package as a substitute for a typewriter; accountants use accounting packages as substitutes for file cards, Kalamazoo, etc; the draftsperson uses CAD in place of drafting board and pens; and the architect explores plot ratios using a spreadsheet rather than sheets of paper and a pocket calculator.

This substitution strategy affords a certain security to practitioners, particularly when it is phased in incrementally. Disruption of normal work patterns is minimised. Users are given an opportunity to become comfortable with the technology. Management are provided with an opportunity to experiment-to find out what works and what doesn't, and how people like to use the technology. Many practitioners are wary of becoming overly dependent on the technology-they don't want to have to employ dedicated technical personnel (systems managers) nor do they wish to be "held to ransom" by CAD operators. They don't want IT to impinge on the traditional values of the practice. Introducing the technology gradually would seem to be an appropriate strategy for keeping those values intact.

On the other hand there can be drawbacks. There are many who would argue, theorists and practitioners alike, that this kind of piecemeal approach will not allow the technology to be used productively (Harris et al., 1989, p.370). It is, in fact, a common perception amongst architectural practitioners currently using CAD that they are not using it productively. Furthermore the substitution approach can result in serious rifts between IT personnel and management. In the worst case scenario a practice may purchase a CAD system because their clients demand it and rely on junior staff to get it running, providing neither training nor management support. There is a lack of recognition that the new technology requires the development of new practices.

Initially, the technology may be introduced as a substitute for conventional tools and methods. The most successful marketing of computer systems appears to exploit this. The available technology may in fact lead to the construction of specific sub tasks within the firm. One of the practitioners interviewed identified a number of key roles for himself-secretary, accountant, graphic designer-roles aligned with the software tools incorporated in his practice.

Moving from senior management to a small business practitioner, I'm using all the tools to avoid putting on junior staff … You become a secretary, an accountant, a graphic designer because you're in the business of selling information and you have to present that information professionally. The level of information presentation is very high.

As the firm begins to accept the technology and as it becomes integrated into the firm's praxis, the more innovative aspects of the technology may emerge and be exploited.

Innovative delivery of traditional services

The culture of architecture is one in which innovation is highly valued (Cuff, p.52; Blau, 1987; Saint, 1983). Innovation is also an important aspect of management culture (Peters and Waterman, 1982). It is now a familiar line of argument that organisations will only reap the benefits of technology if they restructure their operations. Radical changes in the praxis of the firm are implied. Those practices that have developed innovative ways of delivering traditional architectural services are regarded as "creative leaders."

They [the practices that are experimenting with new ways of working and designing] are probably the same practices who would have been creative had they not had computers.

It is important for those architectural practices with corporate and government clients to be perceived as being progressive. Merely having computers has, in the past, been an important factor in such perceptions. As computing use becomes mainstream, "client perception" is likely to provide the impetus for new services. These services will exploit both traditional architectural skills and the possibilities created by computing technology. Our study of practitioners confirms several "standard" innovations that have been incorporated by practitioners.

(i) The development of databases of CAD details for use in multiple projects. Most practices that use CAD for detailed documentation establish libraries of details that can be pasted into any number of projects.

(ii) The development of specification databases. The query and report writing facilities of database programs make them suitable for the production of specifications for complex projects. These specifications may be used throughout the administration of the project, and in the post occupancy administration of the building. There have also been attempts to develop specifications for use in multiple projects. Examples of such standardisation include corporate space standards and "generic" specifications for complex building types such as hospitals.

(iii) The use of CAD databases in the ongoing administration of projects. CAD can be used for more than merely drawing up proposals. Queries can be made about dimensions and offsets and areas. This facility can be valuable during the course of administering projects.

(iv) The organisation of drawing production so that each documentation drawing is linked to related drawings. This is considered important in maintaining consistency between drawings when variations are being documented.

(v) The organisation of the tasks in such a way that CAD operators sit alongside the designers and model proposals as they are being developed. Designers thus become familiar with the technology and what it can do for them without distracting them from design.

(vi) Restructuring the production of correspondence (letters and memos) so that secretaries (or practitioners) are responsible for the text but a graphic artist or designer is responsible for the layout.

(vii) Using the resources of the accounting personnel in a more dynamic way, including requesting weekly rather than quarterly reports, and having management and practitioners work with spreadsheets.

(viii) The use of electronic communication facilities for remote collaboration on projects.

Practice innovations such as those outlined above may occur informally in smaller practices, but require a commitment from management and often the commitment of substantial resources in larger practices.

Amidst the futuristic outlook of the "technologists" and the promotion of innovation within architectural circles there is an emerging sense that it is not necessarily a good idea to be "leading edge." Investing in the development of computing tools can be costly. Projects and work practices change. Databases of details or specifications may only have a limited lifespan. Finally, the model of IT and practice held by the firm may be holding it back. One of the difficulties arising out of the rationalist model is that there is a tendency to look to technology for a "solution" to everything. The rationalist model conceives of the building industry as unnecessarily fragmented, documentation as unnecessarily ambiguous, and design as unnecessarily inefficient. Integrated building models, complete project descriptions and comprehensive suites of analysis programs are obvious "solutions" to these "problems." The "problem" from a praxis-based point of view is that the rationalist is overlooking the importance of everyday communication and work practices. The "problems" perceived by the rationalist are, by and large, not "problems" to be "solved" but realities of the practice of architecture and of the construction process and its participants-the total field of praxis.

Developing extended services

Information technology opens up significant opportunities for architects to extend the scope of the services that they offer. Some technology based areas into which architectural practices have diversified include desktop publishing, designing computerised slide show presentations, facilities management and computer consulting. The motivation for diversification include the provision of a comprehensive service to clients; exploiting existing investment in equipment and training; and countering the vagaries of a cyclical market for services. This type of approach seems to be viable in both large practices where there is management commitment and in small practices-indicating the scope that IT can offer the small practitioner.

Of course, diversification can also be perceived as detracting from core design skills. Cuff (1991, p.39) observes that architects have traditionally "sloughed off" constitutive skill areas, allowing them to become professions in their own right. Thus we have the emergence of the profession of quantity surveying to deal with building economics, structural engineering to deal with the mechanics of building design, and project management to deal with the coordination of the construction process. There is some evidence that the same is happening in relation to computing with the emergence of the CAD consultant.

IT and the role of the architect

The practice of architecture involves a broad range of competencies. What we do as practitioners is shaped by the roles that we "play out." It was evident from our survey that the existence of IT is extending the roles available to the architect. In particular we have the emergence of the notion of the architect as information generator. Another interesting phenomenon evident from our survey is that the roles with which practices and practitioners identify influence their response to IT. Allegiances to the roles of architect as form-maker, architect as generalist, and architect as business professional have an important bearing on those aspects of IT considered to be significant to the profession.

The architect as information generator and coordinator

Increasingly, CAD is being understood not just as an alternative to manual drafting but as a database. An opportunity exists for architects to recast themselves as generators and coordinators of building information. This is precisely what has been happening in the area of facilities management. Responses to our survey indicate the potential to develop this role throughout the design and administration of a project. Practitioners suggested that architects should promote the fact that CAD allows up-front checking for inconsistencies in documentation, immediate responses to queries, and fast propagation of design modifications throughout the documentation of a proposal.

To make a CAD drawing more complete a lot of the design work that may have been done on a piece of paper on the site has to be done at the documentation phase-for a correctly done project this may make the documentation stage longer than if done manually but it greatly reduces the construction phase-because everything is thought out it's very accurate because it's on CAD and there are less mistakes and less question marks-I think clients should be made aware of that fact more than they are.

If the state manager says "what's the gross area?" you can tell him in half an hour if it's on CAD. When it's done manually it will take you three days plus or minus 20%.

… one of the great attributes of our system is that we can write macros very easily. …[In one project] there were 260 client induced variations all of which were tracked perfectly in our written database and on our drawings. The job captain, who was a director of the company, knew at all times that those amendments were up to date. The speed with which we could get those amendments back to site was very fast. That endeared us to the process of construction which meant that everything went smoother. We now believe that we could not have done that project without CAD.

Concept designs tend to be on butter paper but they are transferred very quickly to CAD, some manipulation is done, then you spit it out on the laser [printer] so that you're actually dealing with real dimensions, and then work on butter paper again...

This change in understanding from CAD as an automated drafting tool to CAD as a powerful graphic and non-graphic database provides the incentive for designers, as well as just draftspersons, to start using the technology. Architectural practices become the source of information about buildings. The building database becomes a source of detailed information including information about site coverage, floor areas, dimensions, inconsistencies in the documentation, and 3D projections.

Foremost amongst the concern of practitioners is the need for improved data exchange standards. This reflects a recognition that there are significant gains to be made from pooling the output of the various professions involved, to create graphic databases describing the project. The fact that many facilities management consultants advise organisations that they should maintain CAD databases in-house reflects the increasing recognition that such databases are a valuable resource, both in the construction and in the ongoing administration of a building project. Practitioners consider the advantages of CAD databases over traditional documentation to be easy modification, and increased accuracy and flexibility in the way the data may be viewed and interrogated.

Photorealism: architect as form maker versus architect as generalist

Photorealism involves the production of computer images of CAD models that include realistic surface detail and lighting conditions usually generated through ray tracing and radiosity techniques. Architectural practitioners are divided in their assessment of the importance of photorealism. On the one hand a majority of practitioners felt that photo realism would be of great importance. On the other, a number of practitioners with extensive IT experience expressed reservations about the importance of high quality renderings:

We don't do full colour 3D modelling even though our system has ray tracing capabilities because we don't believe that those fully rendered 3D images necessarily communicate anything that the project is about more effectively.

We produce audiovisuals which are computer encoded slide shows that use six slide projections on the one screen to promote projects-the monorail, the festival marketplace-all were spawned in that way. So we've been a mixed media communications company as much as an architectural firm. We go from the dream making end of it all the way-which is where I think architects are using computers in really old fashioned ways. They think that clients are only interested in perspective views of what things look like and they cannot express what projects are about. That has meant that we've had a pretty wide exposure to the kinds of tools that are available, we've been producing our own media as well as our own technical documentation on large projects.

Attitudes towards high quality rendering tools reflect deeper divisions throughout the profession as to the role of the architect. On the one hand there are those who consider the task of the architect as first and foremost that of a form maker. Amongst this group the emphasis is on the artefact. Photorealism figures prominently as a potential technique for presenting the form of that artefact. On the other hand there are significant numbers within the profession who have long argued that the role of the architect should be much more comprehensive. For these practitioners issues such as communicating with clients, delivering expert advice, and being involved in the ongoing supervision of projects are high priorities. Photorealism is peripheral to the suite of IT tools relevant to these practitioners.

Interestingly, at the time of writing, there are few architectural practices that have taken up photorealism. There are a number of reasons why this may be the case: the quality of output, the specialised nature of the software and expertise required, and the amount of work involved. Creating high quality perspective renderings is a time consuming and highly specialised task, even on computer.

There is however another important factor at play. Rendered perspective images are not necessarily pivotal to either the design or communication of an architectural project. Designers are used to working with abstractions. Simple wire frame models which can be generated and manipulated quickly are often of greater benefit to the designer. Clients on the other hand are concerned with issues that extend beyond the appearance of the building. The client is interested in the real value of the project being proposed: image is important but so are construction, maintenance and ongoing operational costs. Designers concentrate on communicating how the proposal will be used, how it will be appreciated, and the commercial advantage that it will secure rather than simply what it will look like. Rendered perspective drawings are not necessarily the most appropriate medium for communicating this kind of information.

Clients are also interested in ensuring that they are receiving a high quality service. CAD is perceived by practitioners as being an important marketing tool but it is a misconception that this has to do with the quality of the images that can be produced. There are two principal ways in which CAD brings a marketing advantage to a firm.

(i) The mere fact that practices have CAD is a useful public relations exercise. The practices appear up to date.

(ii) It is necessary to have the drawings on CAD to collaborate with other client consultants, such as interior designers.

Finally, the communication involved in the development of a successful project is two way. One practitioner commented that even in the case of plans, highly finished computer generated output can inhibit the client from making valuable contributions to the process.

It's interesting what it [presentation techniques that simulate manual drawing] does for clients' perceptions of drawings-I laughed at it to begin with, I thought it was a real con … but there's a point to it-clients are much more inclined to have some input into drawings that look rough.

The production of highly sophisticated rendered images is likely to deter both designer and client from developing a design further. Interestingly, many practitioners feel that walk through animations will be important for client presentations. Animated sequences have the potential to communicate much information about how the building will be perceived (visually and kinesthetically) and used, without relying simply on the abstract sketches commonly produced by designers.

The architect as business professional (with IT support)

Jones (1982, p.173) provides an illuminating insight into the central role of information services in Australian society.

Australia is an information society in which more people are employed in collecting, storing, retrieving, amending, and disseminating data than are producing food, fibres and minerals, and manufacturing products.

An increasingly information intensive society inevitably leads to new ways of doing business and new expectations from clients. What we will see, and are seeing, is a shift away from the situation where the client relies on the understanding and recommendations of senior members of the profession. According to Jones it will become increasingly important to support these tacit judgements with the kind of research and modelling that is being adopted across the board in business.

The widespread availability of information technology entails increased expectations for high quality presentation at all levels. This applies at the level of the firms "paper work." The principle of re-using elements from a central database can be applied throughout a practice's operation. Cut and paste operations can be used to incorporate both text and graphics into a firm's "correspondence" thus facilitating the delivery of a consistent and distinctive image.

We took the attitude that every letter that we sent out of here was a desktop publication. Therefore, rather than expecting typists to have graphic skills, we liberated it completely from that and just asked them to put it in, then our graphic designer would manipulate it. Everything that we were writing became very important because ultimately we can cut and paste and re-publish in different formats. Your potency in all your submissions increases because of that-issues of policy and philosophy that can be stated universally and thought through carefully can be patched through all your projects as a quality control device. So you're delivering your own corporate promise from that database.

As management and accounting software becomes the norm in the business community architectural practices will be expected to keep pace. Olley (1991) found that 47% of 283 Australian architectural practices surveyed were using accounting packages and 28% of those practices were using project management software. Crawford and Hutchinson (1990, p.11) found that 34.9% of 152 New South Wales based practices surveyed used job costing software. Maintaining good internal management records will become increasingly important in an increasingly competitive and specialised environment.

IT and the micro practices of the architectural office

The day to day practices in which architectural personnel engage are important in determining the culture of a practice. Our survey revealed an interesting interplay between IT and those micro practices. In particular it reflected a diverse range of opinions on the following issues:

does IT lead to deskilling or reskilling?

does IT enhance or detract from the exercise of core design skills

does IT facilitate autonomy, is it an essentially isolating phenomenon, or does it reinforces the interdependence of a matrix of practices and players

The diversity of opinion with regard to these issues indicates the complexity of the way in which IT impinges on the practice of architecture. As might be predicted from the praxis model, it is not IT in itself that determines changes in the micro practices of the office, but the interplay of IT with entrenched attitudes and practices.

Deskilling or reskilling

The most significant item of software in architectural offices in terms of skills required is CAD. The principal determinant of whether involvement with CAD is perceived to lead to deskilling or reskilling is the degree of autonomy associated with CAD usage. Two strategies appear to have been effective in preserving the autonomy of the CAD user. (i) The designers use CAD in the course of working on their own projects. They maintain a sense that they have control over the design as well as documentation. CAD is thought of as a rich new medium for producing designs. (ii) An alternative strategy is to identify CAD operators as a distinct category of personnel and to restructure the way that work is carried out so that an appropriate interaction between designers and CAD operators is achieved.

What we found is that our designers were frightened of CAD. So we put them [designers and CAD operators] side by side and as concepts were generated, even though they might be crude, the CAD operators would model them. So pretty quickly these 3D models started to follow up the designer and the thing that he was thinking about yesterday was on his desk at lunchtime the next day in three dimensions. So the designer became interested enough in CAD to take the model over and to manipulate it slowly himself. That was a better investment than demanding that the designer get the CAD skills of a racing car driver [an experienced CAD operator]… CAD operators and designers operate perfectly well together if you leave them to do their job and if management understands the different things that they're doing on that job.

The perception of involvement in CAD as leading to tedious dead-end work appears to arise primarily when CAD operation falls to junior personnel by default. Where there is a perception that design is being monopolised by senior personnel, with junior personnel merely drawing up their ideas, then the use of CAD is thought of as tedious by both parties.

Unfortunately there has been considerable reluctance to recognise CAD operation as a distinct category of skilled activity. Shugg(1993), for example, argues that it is paradoxical that the most expensive item of equipment in the architectural office (CAD) is operated by the most inexperienced staff (draftspeople, junior personnel). From this perspective it seems imperative that designers, rather than less qualified or differently qualified staff, become the primary users of the technology.

There are, however, a number of problems with Shugg's position. First, CAD systems are difficult to operate. Drafting constitutes only one aspect of the day to day activities of the architect. The complexity of CAD software makes it difficult to marry skill in CAD with the more general spread of skills required to practice as an architect. Second, many practices have for years been able to make good use of a division of labour between designers and draftspeople. Since the principal productivity gains associated with CAD are in the areas of modification, re-use and database interrogation rather than in the initial creation of drawings then it makes sense to continue to allocate that task to draftspeople. Finally, it is true of all industries that mass data entry falls to less qualified personnel, no matter how expensive the equipment they are operating.

Enhancing or detracting from "design"

Design is a highly coveted activity amongst architects (Cuff, 1991). It is significant then to ask whether practitioners perceive that IT is enhancing or detracting from the exercise of design skills (and thus the quality of designs produced). Many predictions have been made as to how computers would enhance design. These include:

computers would generate designs automatically;

computerised procedures would force designers to be more logical and would thus produce better designs;

IT would free designers from more mundane routine aspects thus allowing them to concentrate on design;

IT tools would allow designers to produce multiple options thus enhancing their creativity.

We will consider each of these predictions in turn.

(i) Computers will generate designs automatically. The practitioners in our survey readily acknowledged the once prevalent notion of "automated design":

Interviewer: What do you recall were the major predictions regarding computers in design ten years ago?

Practitioner 1: Software that would design the whole building after entering a few options.

Practitioner 2: All that you would have to do is punch in all the parameters and it would go away and give you a complete office tower.

The practitioners also acknowledged that the promise of "automated design" had not been realised, and further held the firm belief that the automated generation of designs was not possible. Various reasons were offered as to why it was not possible to automate design.

An architect makes a snap decision that would require an expert system to sort through hundreds of variables, and the database has to be up to date. An architect might not make the correct decision but he makes it instantaneously taking all sorts of abstract concepts into account.

…the moment you try to create standards is the moment that people try to change them but there would be a lot to gain…

Despite the practical recognition that it was not possible to automate design there was an uneasiness amongst some practitioners about why this should be the case.

The once prevalent belief in "automated design" is a legacy of the early design methods movement, or more generally of rationalism. According to rationalism, design decision making is, or should be, governed by systems of rules. We should be able to make those rules explicit. We should be able to analyse a design situation objectively, identifying the variables and objectives that define the situation. We should be able to apply rules to develop a solution that is "the right solution" for a given situation.

At the heart of the problem with rationalism are assumptions about the nature of decision making. It appears that decision making is not a matter of following a set of context free rules. If it were, how would we know when and how to apply those rules? The kind of judgement required in decision making is necessarily prior to the formulation and application of rules. It involves taking into account an understanding of what is important in a situation, what is to be expected from others, what is an appropriate way of acting.

What is the basis of such understanding? To relinquish rationalism is not to assert that decision making is arbitrary. According to the praxis model decision making is grounded in patterns of thought and action that integrate a culture and define an individual's relation to it. Individuals are constantly being inducted into socially defined patterns of thought and action. These cultural norms are typically taken up unawares. They govern the way we dress, how closely we stand to one another, the tone of voice we use in a given situation, how we act out the role of architect, designer or client. They also govern the possible stance that an individual may take up in relation to science, politics and religion. They govern our work practices and are constitutive of the roles that we take up in society. These patterns are not static-they are always changing.

Decision making occurs in the context of these patterns of thought and action that have been ingrained in an individual by the culture or cultures in which that individual participates. Only in the most elaborately constructed situations could it be suggested that this decision making has the character of rule following.

(ii) Computerised procedures force designers to be more logical and thus produce better designs. One aspect of the commonly accepted view that rational decision making is a matter of following context independent rules is to brand designers as illogical. Computers are thereby seen as a competitor to designers-the superior capacity of the computer to consistently follow rules shows up the inadequacies of the designer.

It imposes stress because you are going from a craft based industry to a building production line process-you have to do things in a logical order, a lot of architects and people managing design don't think logically and they get caught out.

Adopting the rationalist stance implies an agenda for reforming practice: analysis programs should be run from the earliest stage of design; design expertise should be made explicit; designers must change their ways to conform to the "logical" mode of operation of the computer. The payoff for conforming to "logical" modes of practice would, according to the rationalist position, be better designs.

An architect makes a snap decision that would require an expert system to sort through hundreds of variables… An architect might not make the correct decision but he makes it instantaneously taking all sorts of abstract concepts into account.

There are a considerable number of processes that a designer goes through that are logical processes-they will never produce great designs, they will never produce the Opera House, but they will produce a lot better than the average design solution that you will get from the average architect, mainly because it was produced in a logical way.

Once again this position can be criticised from the point of view that rational decision making is not a matter of following rules. Rational decision making does not have the character of a computer algorithm. While there is no doubt that the introduction of IT into architectural offices requires the development of new organisational skills it is inappropriate to characterise these skills as more or less "logical" than those already in place.

Another effect of the existence of the rationalist stance has been to legitimate the category of the romantic. Foremost among the concerns of designers in making the move to CAD is the "mode of input"-they don't want to lose the primacy of working with pencil on paper. Many designers speak of a left-brain/right-brain conflict, particularly when first learning to use the technology.

While the difficulties that designers experience in adapting to the technology are important, couching these difficulties in left brain-right brain terminology implies a confrontation of two different modes of thought. Designers can, and do, master the technology. The difference would be better thought of as a difference in media. Designers can exploit the best aspects of both media, there is no requirement that they completely abandon one in preference to another. A practitioner in a small practice adopted just such a hybrid approach.

Concept designs tend to be on butter paper but they are transferred very quickly to CAD, some manipulation is done, then you spit it out again on the laser so that you're actually dealing in real dimensions, and then you work on butter paper again.

(iii) IT will free designers from more mundane routine aspects of practice thus allowing them to concentrate on design. This belief appears to be manifest in two different ways. On the one hand there are those practitioners who believe that some of the "thought processes" that a designer uses can be captured as computer algorithms. By capturing those "thought processes" those "logical" parts of the design process could be undertaken by the computer leaving the designer free to concentrate on the more creative parts of design. This approach is subject to the same problems as those of "automating design."

Alternatively there are those who use tools such as spreadsheets and word processing packages as design tools.

I use a Mac all the time now. When I feel like doing things, I flick through things I haven't thought about for a month and think more about it-your ideas don't just drift away.

(iv) IT tools will allow designers to produce multiple options thus enhancing their creativity. One of the earliest predictions made in relation to CAD was that it would allow the designer to explore multiple options. Considerable effort has been expended in developing programs that generate design options. It was evident from our survey that practitioners have not taken up these programs. The principal benefit of CAD from the practitioners' point of view is that it allows them to modify proposals. This use of CAD concurs with findings of Darke (1979), Foz (1973) and Hillier et al (1972) as to the nature of the way designers develop proposals. Designers "generate" an initial scheme from the earliest stages of design and modify that scheme in the light of the possibilities and problems that it raises. They do not "generate" multiple schemes. They do not need to "model" a variety of proposals in order to come up with one that is appropriate.

Autonomy or isolation of the individual

Our study of practitioners confirms that most hold to one or several of the following understandings of the role of the computer.

(i) The computer is equipment that demands concentrated attention. Its use leads to isolation from other aspects of office practice.

(ii) The computer is a tool that enhances the capabilities, and thus the autonomy, of the user.

(iii) The computer is ubiquitous technology. The computer fades into the background-it's just part of the furniture, part of the equipment that you use to get the job done.

(iv) The computer is a window to information and data via electronic mail, on-line catalogues and collaborative environments.

(v) The computer increases our dependence on a technological matrix, including our dependence on those who are responsible for delivering the technology.

These positions are not mutually exclusive. Practitioners tend to fluctuate between these various attitudes towards technology. Each will be discussed in turn.

(i) The perception that computer use can lead to isolation from other aspects of office culture has a number of origins. Practitioners find that there is a level of engagement when working at computers that may be different to working with pen and paper or at a drawing board.

When you're at a drawing board you get distracted more easily than when you're at a computer. It's just because you look up and have a look at what other people are doing, whereas the computer somehow has this effect where you can stare at the computer for ages ... I think you need more attention because you're not just drawing a line. You're thinking about how long this line has to be, and you're thinking about the X and Y coordinates continuously.

Drawing boards are "more physical" than computer terminals, you are also working with a larger surface. One strategy that practitioners have implemented in an attempt to preserve some of the "drawing board office" qualities is to have display screens that are large enough for two or three people to sit around. Another is to avoid creating a work pool of CAD stations by distributing the computers throughout the office. Interestingly, the reduced work surface does not seem to be a problem for individual users, even though they must pan and zoom to negotiate their way around a drawing.

Also at play in user's perceptions of computers as a source of isolation is the '60s and '70s image of the automated office as a soulless environment where people are there to "serve" the technology. This mainframe aesthetic is gradually being displaced as computing technology becomes commonplace. But residual fears remain. This is particularly true of those practitioners that are not yet participating in the use of advanced IT.

For a lot of lay people the idea of a computer is completely tied up with the idea of cold, hard-headed rationalism, and this actually prevents a lot of people from accepting or using the technology because they would consider accepting computers to be the same as accepting a cold, hard-headed, rationalistic approach to life. If we could break down the concurrence in people's mind of computing and rationalism, a lot of interesting things could happen with computers in practice.

(ii) Computers have the potential to enhance the autonomy of the designer in a number of ways. Computer applications can extend the skills of the designer.

You become a secretary, an accountant, a graphic designer because you're in the business of selling information and you have to present that information professionally. The level of information presentation is very high.

I couldn't possibly perform a design role without the full range of technology, I can no longer draw with a pen and pencil. Anything that I do in the design area is either in my head entirely or it's entirely on the screen. For me it's been wonderful: I could not sketch in 3D; to be able to model certainly enhanced my design skills.

Computers are eliminating middle management, thereby enhancing the autonomy of those who are left .

There are two schools of thought: there are the professionals that resist doing those sorts of tasks; and there are those that embrace it because it allows you to do what you want to do without having to transfer the information to somebody else. I think, like most organisations, we're probably pushing out the paper shuffle of middle management. People who cannot add knowledge are just disappearing.

Finally, computers appear to facilitate the execution of mundane tasks, freeing up the designer's time so that they can concentrate on other activities.

I do not think that design has become more complex than it ever was, but I do [think] that computers allow us, by taking care of the complex tasks, to concentrate on other ones.

Of course, some of these tasks are outside of the traditional areas of architectural expertise, and now include page layout, formatting and IT trouble shooting.

Computers are still considered to have an obtrusive presence in architectural practices. This can be both a positive and a negative phenomenon. On the one hand lack of familiarity with the technology can interfere with designing.

One of the problems initially when you start working with CAD is the way that the information is input compared to the way that we design. Design is essentially a right brain, pattern making activity and the way that we input, even with a mouse, tends to be left brain dominated because even with a mouse you're doing keyboard things and certainly in the initial stages until that becomes fully automatic what tends to happen is that you're in a right brain spatial mode and you have to switch to a left mode to input information.

On the other hand computers can be seen as special, and access to the technology may be regarded as a privilege.

Today we are talking from the "inceptional" end of CAD. Ten years down the track it may be completely different, where CAD operators and designers may feel like data processors feel today. At the moment there is a novelty associated with CAD where people feel they have achieved if they end up on a machine.

(iii) As familiarity with the technology grows computers are becoming part of the day to day life of design practices. However, the involvement of aspects of architectural practice in an anti-technological romanticism may mean that practitioners follow rather than lead in the appropriation of information technology.

A lot of architects don't like the technology of their practice (even drawing boards). I think we need an understanding of the role of technology, and in the past architects have been pretty bad at catching up.

(iv) The notion of the computer as a window to information is at the forefront of predictions about the future but, for the moment, is largely unrealised in practice.

I do not think that we are barraged with more information than in the past, but the opportunity is there to access information and it is just not being presented. I think it is insane that you cannot dial up a database and find all the different types of glass available on the market.

To date there are only a few companies that have developed databases of information relevant to designers. Some practices are using electronic mail to communicate between offices in different cities but these practices are the exception rather than the rule.

(v) As architectural practitioners incorporate IT into their firms they become increasingly dependent on both the technology, and upon those who supply, maintain, and coordinate the use of that technology. This dependence can be unsettling.

I am more removed from the process than I otherwise would have been; more dependent on others than I would like to be.

Many practitioners have actively resisted being placed in a position of dependence on technology and "technocrats," by insisting that architectural personnel take on responsibility for the technology rather than employing technology specialists.

IT and emerging management practices

The management practices which have evolved as architectural offices embrace IT stand as testimony to the weight of existing norms and practices. The transfer of "architectural" values and modes of operation are evident in phenomena such as the tendency of many offices to favour "studio" models of training for IT, and the reluctance of many offices to employ dedicated non-architectural technical staff. Our survey revealed interesting trends emerging in the following areas:

day to day management

IT selection

training for IT

customisation

Day to day management

The introduction of IT is changing the way in which architectural practices operate. This is particularly the case in relation to CAD. The most intensive use of CAD to date has been in the area of documentation. It is not surprising then that this is the area that has generated new management tasks. New management tasks common to most practices using CAD include the following.

(i) The scheduling of drawing production

With CAD drawing elements can be re-scaled, and detail can be added to them. It often makes sense, therefore, to complete smaller scale drawings (1:200s, 1:100s) before more detailed drawings. Furthermore, base drawings can often be directly referenced so that information is drawn once only rather than duplicated across drawings. This can be of considerable assistance in ensuring that modifications are reflected throughout the entire documentation.

(ii) Scheduling the provision of information

The accuracy of CAD can impose a requirement that accurate information about a proposal be supplied from the earliest stages of the documentation process. Inaccuracies that formerly may not have been picked up until construction had commenced show up in the documentation phase. The change that this introduces is that designers may now have to specify information earlier than they did using manual methods. For example, a designer may have to indicate the precise dimensions of the site or the building envelope at the sketch design or development application phase.

(iii) The maintenance of office standards

Critical to the effective operation of CAD is the establishment of appropriate layering standards, title blocks, libraries of elements and file structures. A coordinated approach is required to avoid idiosyncratic systems of operation emerging between CAD operators. The need for coordination is more important where there is the customisation of menus and macros.

(iv) Coordination with other offices

Apart from the difficulties involved in translating files from one CAD program to another, the principal difficulty in sharing CAD files is that different practices employ different layering standards and naming conventions. Furthermore, there are potential difficulties associated with establishing protocols for data transfer between offices.

(v) Maintaining backups and archives

Instituting reliable backup procedures is critical to the success of any IT based activity. This can be a difficult thing for practitioners with limited exposure to IT to grasp.

(vi) Tracking versions of files

Different users may independently make changes to a file. Individual users may store several versions of a file. Keeping track of the latest version of a file is a significant management task. There are several software packages available that assist in this.

(vii) Controlling access to equipment

In many practices terminals are shared. Furthermore, determinations have to be made as to which projects and which parts of projects are undertaken on CAD.

(viii) Technical support

Some larger practices consider it necessary to employ a dedicated technical person ("systems manager"), usually with a technical computing background. The small and medium firms want to avoid the situation of having to employ dedicated technical personnel. They continue to rely on vendors, bureaus and their own knowledge for technical support.

In some practices the incorporation of these management tasks can change the way work is undertaken in those practices. For example, one of the practitioners interviewed regarded the management changes as indicative of a transition from a craft based industry to a production line. In some practices the management of drawing production already exists as a distinct task (independently of the introduction of CAD).

Designers are designers and then there's another breed of people called design managers. They're the people who have to help produce the project.

Where this is the case we expect that the management of the production of CAD drawings would simply become an extension of the "design management" function. Finally, small practices may assimilate the changes required without dramatically changing the way that they operate.

In all but the smallest practices the introduction of CAD has entailed the nomination of a "CAD manager." The qualifications and status of this person within a practice varies considerably. The CAD manager may be appointed from the ranks of designer or design manager, or they may simply be the firm's most proficient operator. In most cases this person is expected to combine the role of "CAD manager" with their original function.

IT selection: the emergence of standards

There is an emerging sense amongst practitioners that the use of computers in the industry has its own momentum. Making technology related decisions is not a matter of "rationally" analysing the nature of the practice and the various hardware, software and training systems available. It is rather a matter of deciding what is the standard system in use in the industry. Such standards are often "de facto." They emerge within the tacit conventions of praxis that are the results of persuasive marketing, products that are general and robust, and a culture of use that develops where practitioners can share their understandings of the systems. Hardware is subservient to software in this regard. By always selecting standard software the practitioner does not have to research what different vendors have to offer. Practitioners are in a position to appreciate what the software can do for them as others are already using it. They expect to be able to gain such an appreciation without being immersed in technical detail. The industry tends to consolidate its expertise and practices around particular software, for example, AutoCAD.

Many practitioners recognise that these "standards" are not necessarily the best systems available. Some practitioners consciously depart from the norm selecting software that they feel to be appropriate to their own particular practice. Others adopt the "standard" to be compatible with other firms or because of the quality of support (maintenance, training, troubleshooting) available.

Training

Training is a deliberate intervention into practice to accommodate the changes experienced with the introduction of new technology. Of central concern to practitioners is the issue of IT related training. This reflects both the state of the software and the degree to which the profession has adopted IT. The industry is in the process of building up a pool of professional and technical staff trained in the use of the technology. The attitude of management towards training can be critical to the successful implementation of IT.

Practitioners are buying into two traditions with regard to training: (i) the business model of short formal training courses, and (ii) the studio based model with its design orientation and emphasis on tacit understanding. The business model is mainly adopted by the larger practices for training in areas such as word processing, spreadsheets and technology related management issues. This short formal course training may be provided through external courses or courses run in-house. Apart from short initial training courses provided by the CAD vendors, "studio" based training is the preferred option in relation to CAD. Many practitioners feel that CAD training is only really effective in the context of working on a real project. Most, however, are critical of leaving inexperienced operators to their own devices. "On the job" training coupled with support from experienced operators is considered to be the most effective approach.

The fact that the industry is currently in short supply of trained personnel creates problems. While there is some concern that it is difficult to retain trained staff, many practitioners take the view that it is short sighted not to train people for this reason.

If you train people they recognise that you have the right attitude.

Some highly trained operators have been able to demand high salaries because such expertise is in short supply. This is particularly the case with the use of CAD to develop photo-realistic renderings. Training is clearly an issue when it comes to selecting hardware and software. Some practices make exclusive use of Macintosh computers, which are understood to require little training, and to avoid confining expertise to a few experienced operators.

The dominant value of the architectural office is an emphasis on design. Architects have been reluctant to erode the standing of design values through an emphasis on technology. This has had both positive and negative implications in terms of the training of personnel. In many cases it has meant that training of both operators and other relevant staff is inadequate. Rifts have been created between design, management and technology personnel. On the other hand, the values of the profession have remained intact. The operator with a good knowledge of construction or who is capable of using CAD as a design medium is highly valued. Simply gaining computing skills, however, does not ensure increased standing for the individual within the profession.

Customisation

The attitude of practitioners towards customising their own software varies dramatically. On the one hand it is recognised that customisation can be expensive, and there is the risk of developing idiosyncratic systems which make the transfer of files and expertise difficult. However, in many cases it has been necessary to customise systems in order to make them more productive. Without customisation some systems are too complex to use. The command set may be too large or difficult. Customisation can simplify the way that files are organised, or drawings are layered. Customisation may also be necessary to save operator time and to minimise repetitive tasks.

Two conflicting views emerge with regard to customisation: (i) that it is necessary to customise in-house; as opposed to (ii) that you should wait until third party products are available-that it doesn't pay to be "leading edge." It is common for in-house developments to be shared amongst practitioners through user groups (Sidhu, 1990).

Many practices undertake some system development and/or enhancement work, mainly writing macros and customising menus. Few practices, however, have a full time staff member dedicated to system development and enhancement work (Sidhu, 1990, p.36).

Praxis and the Future

The praxis model suggests that our practices are carriers of culturally defined metaphors. These metaphors shape our attitudes towards the future. In assessing the future role of computers commentators invariably invoke some overriding metaphor of what the computer is-a disparate set of databases, a form of intelligence, a communications tool, a medium. Out of each of these metaphors arise scenarios that accord with the metaphor-the future lies in integration, intelligent assistance, collaborative design systems, or an adventure in a multimedia cyberspace.

Integration

"Integrated CAD" involves the pursuit of a single building model from which practitioners such as engineers, quantity surveyors, and facilities managers can automatically extract the data that they require. Vendors and researchers emphasise the importance of this development.

This is where it's all going-getting my system to talk to the consulting engineer's systems to talk to the client's real estate agent's system, his property manager... It's all about communication, communication protocols and data protocols-this is where the excitement is.

The major prediction for the future from one consulting engineer was of

huge integrated databases and whole buildings designed with one system. Most have not come true and we are still predicting them.

Fully integrated CAD so far appears to be elusive in practice.

We're currently experiencing that problem with the children's hospital where we're using a lot of consultants with different CAD systems.

There's currently no way of transferring 3D-we've just had a job where it would have been very nice to get the engineer's 3D.

According to one consultant

Even with our client base we deal with about five different CAD systems.

There are problems with the idea of full integration. No matter how comprehensive is the building model there is always a need to augment that model or to change the way that the building is described (Rooney, 1987). For example, the issue of whether dimensions should be taken from the outside walls, inside walls or the centre lines of walls will vary according to who is using the model and for what purpose. As a further example, the definition of floor to ceiling height changes where there is a light well passing though the floor.

It's a myth this full 3D model in which everyone whips in and takes data out of it, but a disintegrated model is quite feasible and I think that's the way we're going.

Responses to these kinds of problems fall into two categories. On the one hand there are those who see lack of integration as a problem with the initial description. The CAD system should be modified so that it forces the user to supply the required information at the outset of the entry of the building model. For example, the CAD description used to produce drawings for building development approval should also include information about the type of floor material, which may be necessary in the calculation of costs later on. According to this view the CAD system should be modified so that the CAD user has to specify the information about materials at the outset. As a second response the facility can be provided in the CAD system to allow a user to specify the relevant information or to tag entities (such as floors and walls) with such information only at the relevant stage of the design and documentation process.

The advantage of the second approach is that it recognises that within the construction industry sophisticated communications conventions are already in place. Architects, builders, engineers and quantity surveyors are trained in the interpretation of working drawings. Working drawings do not constitute a three dimensional model of a building. They are slices through the building that are sufficient to communicate to the relevant professionals the information that they require. There are strategic moments when it is appropriate for the designer to make decisions about materials and finishes. Designers should not be forced to commit themselves to such specifications prematurely. Rather, CAD systems should allow the relevant professionals to specify and extract the information that they require as they require it.

A similar problem to the timing of decision making arises in relation to the transition between conceptual design and working drawings. Some designers have found computer based 3D modelling to be useful for conceptual design. The traditional approach to supporting such exploration has been to allow the designer to construct building models from standard and pre-defined elements such as walls, floor slabs, doors and windows in such a way that the "model" can be automatically converted into a set of working drawings. The problem here is that designers do not normally work with such "primitives." Abstract three dimensional graphic primitives such as cubes and spheres are more appropriate to work with, and provide greater flexibility, at the sketch design phase. "Conversion" of such three dimensional models into working drawings may involve taking slices through the model as an underlay for the production of plans and sectional working drawings.

A corollary to integrated building models is the view that it is possible to create data exchange standards that will make possible the transparent transfer of data between different CAD systems or specialised application programs. The problem with this view is the same as the problem with the assumption that we need integrated building models-different representations are good for different purposes. To hold that all applications of a model should be based on the same underlying representation, or should be able to be converted into a "standard" representation, would be to hold that all applications programs have similar data requirements. Such a position denies that differences in the functionality of applications are significant.

At the moment it is impossible to communicate between systems even with DXF [a common data exchange standard].

The difficulties that participants in the STEP project, a project aimed at establishing a standard representation for CAD conversion, have in agreeing on a common representation is testimony to the efficacy of different representations for different purposes.

The moment that you try to create standards is the moment that people try to change them but there would be a lot to gain.

The belief in integrated CAD also underlies certain misgivings about the technology-that the computer will be pervasive and indispensable, it will be a tool of total and integrated control of the design, construction and management of the building, and eventually of all the players in the design and construction process. The greatest power will reside with whoever controls the database.

The impact on the built environment is going to be enormous-it's going to come through integrated databases. The design professional will leave the contractor behind in this regard.

Intelligent systems

The goal of creating "intelligent" systems permeates thinking at many levels of computing technology. At its most extreme it has led to attempts to uncover algorithms for design or algorithms for creativity (Simon, 1975; Stiny and Gips, 1978; Stiny and Mitchell, 1978). The initial enthusiasm within the profession for expert systems follows a scenario emerging from the metaphor of computers as "intelligent systems."

Information exchange means by default that you have access to a lot of information. Manipulating it is a province of expert systems. Expert systems haven't taken off anywhere near where people thought they would, but that's partly because you need an environment rich in data.

Attempts to convert scanned documents (raster drawings) into CAD formats (vector drawings) provides an instructive example of early attempts to apply "automated intelligence" to solving geometrical problems. Teicholtz reviews the state of the art of scanning technology.

Every year someone says that accurate raster-to-vector conversion is only a few years away, and while this software is maturing, a perfect "blackbox" conversion may never exist. … The ineffectiveness of early "vectorizing" software turned some potential users off, but then a new concept emerged: scan overlays. (Teicholtz, 1991, p.55)

The "problem" of raster to vector conversion is one of converting a matrix of pixels (dots) into lines and arcs. Automatic conversion typically yields vector drawings made up of a great number of lines that fail to reflect the geometric construction of the original drawing. As Teicholtz points out much effort has, and continues to be, expended on improving automatic conversion techniques. The alternative is to use the raster drawings as base drawings within CAD systems. They can simply be stored in raster format and be used, for example, as base plans for furniture layouts. Alternatively they can be converted into vector format by manual tracing on the computer screen. The advantage of the latter alternative over automatic raster to vector conversion is that the user is in control of the geometric entities created.

It is significant that the notion of "scan overlays" emerged as a "new concept" in the wake of earlier (disappointing) attempts to automate raster-vector conversion. The pursuit of "intelligent" systems is often construed as being a harmless, if not imperative, exercise. It is becoming increasingly evident, however, as vendors and researchers fail to deliver the promised "intelligence," that the pursuit of such systems has resulted in the neglect of the often competing requirements for the provision of facilities that allow the user appropriate control. For example, if the orientation is towards creating "intelligent" systems then emphasis will be placed on creating the right database structure. If, however, the orientation is towards allowing the user to interactively control the process of constructing and manipulating databases then the environment which the user is exposed to will be completely different. In the case of the "intelligent" systems approach the user is primarily required to manipulate parameters. In the absence of "intelligent" systems the user must be provided with facilities for defining data and constructing relationships between elements of that data.

Notwithstanding the difficulties there is optimism amongst practitioners about the ultimate feasibility of expert systems.

I think a lot of the stuff which has been done is nonsense, but we'll get there eventually.

They're going to happen but the results are going to be a fair way away because of the complexity, especially considering the fact that we as a profession can't agree on a set of layering standards. An architect makes a snap decision that would require an expert system to sort through hundreds of variables, and the data base has to be up to date. An architect might not make the correct decision but he makes it instantaneously taking all sorts of abstract concepts into account.

We could probably come halfway, but I do not think we can replace a designer with a computer. They will eventually be very helpful in the decision-making process. Computers are good at number crunching and storing information, and the extent of its use will be as a code checker.

Collaborative IT

The computer is being used increasingly as a communications device. The metaphor of the computer as communications tool provides the impetus for recent developments combining the computer with electronic communications. Electronic communication involves the exchange of digital information across a computer network. Information held in one computer user can be transferred to any number of other computers. This exchange can be instantaneous (synchronous), can include various types of media (text, graphics, audio, video), is possible across different makes of computers, and can be as efficient between international sites (wide area networks such as Internet) as it is within a single office network (local area networks such as Ethernet or Appletalk).

Computer-supported collaborative work (CSCW) and computer-mediated collaborative design (CMCD) builds on the potential of electronic communications to allow individuals at remote locations to work together. Technology is already in place to enable the real time transfer of high resolution images (such as CAD graphics and video) across existing local, regional, national and global telecommunications networks. Such developments make it possible for multiple users in remotely located offices to interact in real time while using the same item of CAD software. There are several possible scenarios: interactive access to multimedia databases including promotional examples of successfully completed buildings, details of current work in progress or potential development sites, "best practice" information on innovative construction technologies, illustrated industry standards and CAD databases of standard details; electronic couriering of large text or graphic files such as specifications or entire CAD models; videoconferencing and basic telephonic communications entirely via the computer workstation; designers working directly with CAD, supported interactively by specialist CAD operators; the "virtual" design team or site meeting; and a "virtual" desktop comprising a wide range of CAD applications, available and running over a network of different machines, with each machine supporting the specific application to which it is best suited. Computer-supported collaborative work is implicated in the re-shaping of the geographical and organisational structure of a practice and its client and consultant base.

As yet practitioners do not make extensive use of computer-mediated collaborative work, but many can see the potential.

We're going to need exactly that in a couple of months with a job that we are doing with a large city office but it's not available so we're going to have to get in our cars and drive in-and I resent that.

The grounds for optimism about the effects of extensive electronic communications reside with the success of the fax machine.

Can you imagine how we used to live without a fax? The challenge is not the immediacy of the updates but how you manage the distribution of them.

But there are reservations about even the feasibility of widespread use of electronic mail.

There is no doubt that fax make it much easier to get hard copy from one place to another but the use of electronic mail in most architectural offices is non-existent, to make it work properly you would have to have a large organisation with a high percentage of workstations to employees or it's self defeating. The exchange of information between various disciplines (eg. the architect and the engineer) is almost a non-event because nobody has the same system, even if they have the same system they don't have the same hardware or the same disc or tape drive.

There are two sets of problems that emerge from the scenario of computer-supported collaborative work. First, the realisation of this opportunity is pursued from the point of view of traditional CAD "problems." The key issues are seen to revolve around the control of consistency and data continuity (whether to have a centralised or distributed data storage and processing resource), on replicating the full richness of face-to-face communication (such as having visual and audio links, and working around a common desktop), data exchange standards, and auditing (such as ownership of ideas and copyright). In each of these problem domains the general approach is to encode a set of procedures that accord with existing design practice. Data consistency is seen as a security problem to be solved through more effective control (such as file locking and user access privileges), face-to-face communication is seen as a problem of channels to be solved by including audio and video exchange (with communication protocols similar to CB Radio or aircraft control), data exchange is seen as a standards problem to be solved through a common building description for the industry, auditing is seen as a quality assurance problem to be solved by logging and recording details as changes are made (details including the user responsible for the change, when the change was made, and who agreed to the change).

This approach is not dissimilar to the one applied in the development of CAD itself, where the intention has largely been to encode a set of procedures that accord with manual drafting practice. In every case the major difficulty would appear to be in prescribing the full range of uses for a particular feature of the computer system. Data exchange standards are very useful, but the search for a single industry standard denies the value of features of individual CAD packages that makes them attractive. Total data consistency would appease most CAD managers, but may be an impractical goal in the context of the dynamics of design. Wherever standards, protocols, controls, checks and systems are implemented, there are particular instances where such procedures do not readily apply and where individuals will therefore seek to subvert and undermine those procedures.

The second set of problems pertaining to computer-supported collaborative work recognises that electronic communication will rapidly become ubiquitous. Rather than attend primarily to standards and protocols, the issues of concern then become issues related to the integration of this technology within a constellation of praxis. How might practitioners make effective use of electronic communications?-not to replace existing modes of communication (such as surface mail and face-to-face conversation), but to augment existing modes of communication (to work away from the conventional office, at home or interstate). How will the practice of design engage the technology? How will it be used and what new design services will it enable? What new insights to design practice will this shifting emphasis on communications bring to light?

Multimedia and Mass Media

There is a growing sense that the computer and the mass media are merging. The metaphor of the computer as medium, in the sense meant by Marshall McLuhan, has been strongly promoted by computer systems developers such as Kay (1991), and this conception is influencing how some design practitioners see their firm.

Computerised communications networks are now assuming many of the attributes we normally associate with the mass media-radio, television and newspapers. We can characterise the mass media as what becomes of writing and the pictorial arts under the influence of mechanical and electronic reproduction-or mass production. The mass media are concerned with reproduction. Information is generated, copied and reproduced many times over. It is also concerned with dissemination-information is distributed widely through media networks. The mass media also involve editorial control. All information received is not distributed indiscriminately but is moderated, sorted and censored by editors, editorial boards, or referees. The mass media are also ubiquitous. Their artefacts are to be found in homes, work places, and as part of the landscape as billboards, video displays, and radio and television broadcasts picked up on car radios, walkmans and portable televisions. The mass media are also public-information is generally distributed indiscriminately, or to "news groups," rather than used privately. Some aspects of the mass media display the property of portability-products of the media (newspapers, magazines, portable radios, etc) can be acquired and carried around by individuals. The mass media are also ephemeral-the content of the mass media often has a short life, perhaps it is archived or repeated, but generally it is never to be revisited (newspapers have a life of one day, radio broadcasts a life of several minutes). Finally, the mass media are impressionistic. The mass media, at their most characteristic, generally trade in quick impressions rather than precisely crafted and accurate information.

The confluence of computerised communications and the mass media brings about an entity with certain emerging characteristics. The mass media have been interactive for a long time-with letters to the editor, talk-back radio, and the broadcasting of home videos on television programs-but now there is the possibility of sophisticated synchronous (real time) interaction through computer networks-such as "on-line bulletin boards" and Teletext (Rafaeli and La Rose, 1993). The electronic mass media have the potential therefore to be highly interactive. The electronic mass media now rely substantially upon computer networking. Computer networks, telecommunications and media networks are merging (Weiser, 1991). The new mass media also increasingly rely upon digital data manipulation-desktop publishing, digital image capture and manipulation, multimedia systems, talking books, and the use of CD ROMs (digital compact disks).

According to some commentators, the mass media traditionally operated as a "one-to-many" form of communication-one source, many recipients. The new emerging mass medium is "many-to-many"-any individual can instantly communicate with another individual, or to a group. As such, the new mass media are thought to have the potential to re-vitalise the democratising power once accorded to newspapers and broadcasting (Rheingold, 1993). As one practitioner put it:

information equals power; until such time that enough people have that information or access to it; then the power ceases.

Electronic communications and the mass media are also implicated in new conceptions of the computer as providing a total immersion environment (Helsel and Roth, 1991; Rheingold, 1991). Experiments in "virtual reality" are thought to herald a possible future in which our senses are plugged into a virtual world of information. This world has features of the physical world. It is said that we will enjoy the physical sensation of moving through simulated physical environments (buildings and landscapes), move through information as though it were transparent to all of the senses, and interact with other people as though we are with them, and assuming whatever form suits us. These scenarios implicate design practice. Benedikt (1991) suggests that there will be specialised "cyberspace architects" who deal with space no less that do traditional architects, but their medium will be information.

As this aspect of computing is new to practice there was very little understanding amongst survey participants of what it entails. Though for one practitioner interviewed the media angle pervaded the whole practice. This was mainly in the area of the promotion of early design proposals.

So we've been a mixed media communications company as much as an architectural firm. We go from the dream making end of it all the way.

Multimedia is a fundamental part of our practice-it's as important as all the other architectural processes and I think it's something that architects have really sold themselves short on. You have to communicate in the realm of concepts and ideas competently … The truth is that there's so much media around that you can muster to give full expression to your ideas, and these decision makers are subjected to communication like that all the time now, everybody gets presentations on videotape-if you're not competent in those communications fields you're going to be stuck in a sort of paper technological field.

Summary

We have favoured the praxis model for our analysis of the experience of practitioners. According to this model the members of a firm are players in an interconnected whole that involves people and their relationships (organisational and informal work relationships), ways of working, technologies and systems of equipment and their interdependencies. By accepting this interconnected whole as the starting point in understanding the firm we are able to pick up on experiences that might go unnoticed in a strict systems or analytical approach. From our study it is evident that there are different strategies through which computers are embraced, or even rejected, by practice. Amongst firms who accept computers there is the strategy of substituting one component (an item of equipment or a task) in the practice for another, such as replacing drawing boards with CAD systems. There is also the strategy of restructuring the firm's activities to carry out traditional tasks in an innovative way, such as turning the documentation of projects into a database management exercise. There is also the strategy of extending existing services, such as moving into the media production business in some way. Whatever strategy is adopted, the introduction of computer technology brings about changes in how the practice is conducted, the nature of the market place, and the role of the individual practitioner. We also discussed some of the management implications of the introduction of CAD and some major metaphors through which practitioners construct scenarios of the future.

These include the use of the computer as heterogeneous equipment to be integrated, the computer as intelligent assistant, the computer as a communications device, and the computer as mass media. Recent developments in computer technology ensure some currency to the last two metaphors, which will soon make their impacts felt in the way practices see their future.

Acknowledgments

This work is supported by an Australian Research Council Small Grant. Interviews were conducted by Ahmed Jumani. Assistance was also provided by David Marchant and John Rollo.

References

Architectural Record (1984) Round Table: Computer use in the small architectural firm, Architectural Record, May: 39-53

Architectural Record (1990) Beyond Computers as Pencils, Architectural Record, November: pp 28-31.

Atkin, B. and Gill, M. (1986) A Study of CAD in the U.K. Construction Industry, Construction Industry Computing Association, Cambridge, England.

Aungles, S. ed. (1991) Information Technology in Australia, New South Wales University Press, Sydney

Australian Bureau of Statistics (1991) Architectural Services Industry, 1987-88: Supplementary Statistics for Architects, Belconnen, ACT.

Badham, R. (1991) Computers, Design and Manufacturing: the challenge, Australian Government Publishing Service, Canberra.

Benedikt, M. (1991) Introduction, Cyberspace: First Steps, ed. Michael Benedikt, MIT Press, Cambridge, Massachusetts, pp.2-25.

Bernstein, R. (1991) The New Constellation: The Ethical-Political Horizon of Modernity/Postmodernity, Polity, Cambridge.

Braverman, H. (1975) Labour and Monopoly Capitalism, Monthly Review Press, New York.

Cowdroy, R.M. (1985) Clients' Perceptions of Architects' Performance, Board of Architects of N.S.W., Sydney

Coyne, R.D and Snodgrass, A. (1991) Is designing mysterious? Challenging the dual knowledge thesis, Design Studies 12 (3): 124-131

Coyne, R.D and Snodgrass, A. (1993) Cooperation and individualism in design, Environment and Planning B: Planning and Design 20: 163-174

Coyne, R.D., McLaughlin, S. and Newton, S. (1994) A survey of computers in design practice, Report, Department of Architectural and Design Science, Unviersity of Sydner, NSW, Australia.

Crawford and Hutchinson (1990) A Study of Means of Maintaining and Developing Professional Competence Throughout a Career as an Architect, Board of Architects of N.S.W., Sydney.

Cross, N. (1984) Developments in Design Methodology, Wiley, New York pp. 245-264

Cuff, D. (1991) Architecture: The Story of Practice, MIT Press, Cambridge, MA.

Darke, J. (1979) The primary generator and the design process, Design Studies 1 (1): 36-44

Dean, J. and Olley, H. (1988) The Market for Architectural Services-An Initial Investigation, Royal Australian Institute of Architects National Education Division, Red Hill, Australia.

Dill, J.C. and Pitman, J.H. (1985) Computers: Evaluate your Options, Architectural Record October: 45-47.

Doubilet, S. (1987) P/A Reader Poll: Computer Use, Progressive Architecture, 68 (8), August: 15-18

Dreyfus, H.L. (1979) What Computer's Can't Do: The Limits of Artificial Intelligence, Harper and Row, New York.

Dreyfus, H.L. (1991) Being-in-the-World, MIT Press, Cambridge, Massachusetts.

Drucker, P.F. (1968) The Practice of Management, Pan Books, London.

Fish, S. (1989). Doing What Comes Naturally: Change, Rhetoric and the Practice of Theory in Literary and Legal Studies, Duke University Press, Durham, Southern Carolina.

Forester, T. (1989) Computers in the Human Context: Information Technology, Productivity and People, Basil Blackwell, Oxford

Foz, A. (1973) Observations on designer behaviour in the parti, Design Research and Methods 7 (4): 320-323

Gadamer, H.-G. (1975) Truth and Method, Sheed Ward, London.

Goldman, G. and Zdepski, M. (1991) Using computers before the working drawings, Progressive Architecture January: 131

Gomez, P. (1983) Architecture and the Crisis of Modern Science, MIT Press, Cambridge, Mass.

Gordon, D.E. (1989) When will CADD come of age?, Architecture, December: 95-99

Gutman, R. (1988) Architectural Practice: A Critical View, Princeton Architectural Press, New Jersey

Harris C.L. (1989) Office automation: making it pay off, in Forester, T. (ed.) Computers in the Human Context: Information Technology, Productivity and People, Basil Blackwell, Oxford pp. 367-376.

Heidegger, M. (1977) The question concerning technology in Krell, D. (ed.) Martin Heidegger: Basic Writings, Harper, San Francisco. pp.287-317.

Heim, Michael (1993) The Metaphysics of Virtual Reality, Oxford University Press, New York.

Helsel, S.K and Roth, J.P. (eds) (1991) Virtual Reality: Theory, Practice and Promise, Meckler, Westport, Connecticut.

Hillier, B., Musgrove, J. and O'Sullivan, P. (1972) Knowledge and Design, in Mitchell, W.J. (ed.) Environmental Design: Research and Practice, University of California, Los Angeles.

Jackson, D. (1993) The future for architecture: a new blueprint for Britain, Architecture Australia, May/June: 72-73

Jay, M. (1984). Adorno, Harvard University Press, Cambridge, Massachusetts.

Jones, B. (1982). Sleepers, Wake! Technology and the Future of Work, Oxford University press, Melbourne

Kay, A. (1991). Computer, networks and education, Scientific American, September, pp.100-107.

Kindon, D. (1979) Marketing Architectural Services, Part 1: The Marketing Audit, Royal Australian Institute of Architects Information Division, Manuka, Australia.

Kish, E. (1991) CAD and the Profession, Progressive Architecture, May: 142-143

Kranzberg, M. (1989) The Information Age, in Forester, T. (ed.) Computers in the Human Context: Information Technology, Productivity and People, Basil Blackwell, Oxford. pp. 19-32

Kuhn, T.S. (1970) The Structure of Scientific Revolutions, University of Chicago Press, London.

Laurel, B. (1991) Computers as Theatre, Addison Wesley, Reading, Massachusetts.

Leebaert, D. ed. (1991) Technology 2001: the future of computing and communications, MIT Press, Cambridge, Masachusetts.

Lischewski, H. (1987) Computers: practical tips for CAD selection and management, Architectural Record, August.

Meyrowitz, Jushua (1985) No Sense of Place: The Impact of Electronic Media on Social Behaviour, Oxford University Press, New York.

Mitchell, W. J. (1987) Reasoning about form and function in Kalay, Y.E. (ed.) The Computability of Design, Wiley, New York pp. 89-98.

Olley, H. (1992) 1990 Profile of the Architectural Profession, Royal Australian Institute of Architects: National Division, Canberra

Peters, T. and Waterman, R. (1982) In Search of Excellence: Lessons from America's Best-Run Companies, Harper & Row, New York.

Pitman, J.H. and Dill, J.C. (1984) Computers: the need for graphic standards, Architectural Record February: 23-27

R.I.B.A. (1992) Strategic Study of the Profession, Royal Institute of British Architects.

Radford, A.D. (1988) Computers in Australian Architectural Practices, RAIA Education Division, Red Hill.

Rafaeli, S. and La Rose, R.J. (1993) Electronic bulletin boards and "public goods" explanations of collaborative mass media, Communications Research, Vol.20, No.2, pp.277-297.

Rector, R.A. (1990) The economics of rationality and the rationality of economics in Lavoi, D. (ed.) Economics and Hermeneutics, Routledge, London.pp. 195-235.

Rheingold, Howard (1991) Virtual Reality, Summit Books, New York.

Rheingold, Howard (1993). The Virtual Community: Homesteading on the Electronic Frontier, Addison Wesely, Reading, Massachusetts.

Rooney, J. (1987) Representing objects, in Rooney, J. and Steadman, P (eds.) Principles of Computer-aided Design, Pitman, London pp.13-37.

Saint, A. (1983) The Image of the Architect, Yale University Press, New Haven.

Shugg, C. (1993) Gaining a positional advantage with CADD, Architecture Australia, Jan/Feb: 61-63

Sidhu, J. (1990) The Use of CADD in Architectural Practices in Sydney, Unpublished masters thesis, Department of Architectural and Design Science, University of Sydney, Australia.

Simon, H.A. (1975) Style in design, in Eastman, C.M. (ed.) Spatial Synthesis in Computer-Aided Building Design, John Wiley, New York.

Stevens, G. (1991) CAD and the Profession, Progressive Architecture 72 (5), May: 141-142

Stiny, G. and Gipps, J. (1978) Algorithmic Aesthetics: Computer Models for Criticism and Design in the Arts, University of California Press, Berkley

Stiny, G. and Mitchell, W.J. (1978) The Palladian grammar, Environment and Planning B 5 (1): 5-18

Teicholtz, E. (1991) Scanning the Scanners, Progressive Architecture (4), April: 55-57

Teicholtz, E. (1992) Computer-Aided Facility Management, McGraw-Hill Inc., New York.

Thompson, W. (1991) Computer techniques for making photo-realistic presentations, Progressive Architecture February: 39-41

US Department of Commerce, Bureau of the Census. Census of Service Industries, 1982. Subject Series.

Weiser, M. (1991) The computer for the 21st century, Scientific American, Vol. 265, No. 3, pp. 66-75.

Weiser, M. (1991) The computer for the 21st century, Scientific American, Vol.265, No.3, pp.66-75.

Winograd, T. and Flores, F. (1986) Understanding Computers and Cognition: A New Foundation for Design, Albex Publishing Corporation, New Jersey.