Jump Over Left Menu
The Workshop arose from a concern that there was a lack of a definition of window management and a lack of focus for research activities in this area in the UK. The main impetus for the Workshop came from the UK Alvey Directorate's Man-Machine Interface Director. To set the Workshop in context it is appropriate to give a brief overview of the UK Alvey Programme, its origins, aims and current status.
The Alvey Programme grew out of a report produced by a government committee under the chairmanship of John Alvey. The report, entitled A Programme for Advanced Information Technology, was published in 1982 . The Alvey Programme can be seen as the UK response to the challenge of the Japanese Fifth Generation initiatives and as a recognition on the part of UK government of the need to invest in Information Technology (IT). The Alvey Report called for support for Information Technology in four main technology areas: Very Large Scale Integration (VLSI), Software Engineering (SE), Intelligent Knowledge Based Systems (IKBS) and Man-Machine Interface (MMI).
The government set up an Alvey Directorate in June 1983 to implement the recommendations of the Report, under the leadership of Brian Oakley. Directors have been appointed for each of the four enabling technology areas and strategy statements have been published for each. A key aspect of the Alvey Programme is that it is a collaborative programme of pre-competitive research. Research projects are carried out by consortia normally involving at least two industrial partners; academic and government research laboratories are included in many of the projects.
Exploitation of research is also a key concern.
Funding for the programme is around £350M over five years. Such funding comes through three government agencies, the Science and Engineering Research Council, the Department of Trade and Industry and the Ministry of Defence; industry itself is expected to contribute 50% of their costs in collaborative projects.
The MMI sector of the Alvey Programme is headed by the MMI Director, Chris Barrow.
The Alvey MMI strategy was published in August 1984 . The MMI programme recognizes the need for research in the MMI area if UK products are to remain competitive in the IT market place. The feeling is that the UK has lost ground in the MMI area over the last decade and that a serious effort is needed to restore the UK position. MMI is seen as playing a vital role in making products from other technology areas usable.
The objectives of the MMI programme are two-fold:
- to raise the level of UK user interface design, in terms of innovation and design methodology, so that industry can compete effectively in world markets;
- to improve UK capabilities in pattern analysis, to make possible the use of advanced speech and image techniques in the user interface.
The user interface design objective is being addressed through a broadly-based programme of research in the three closely related areas of user interface design, human factors and design methodology. New techniques for using new workstation architectures and input devices are a component of these activities. Thus there is a lot of interest in using high performance workstations and window management systems. Single user systems are used extensively by the other parts of the Alvey Programme and are a major part of the infrastructure provided through Alvey.
The need for this workshop arose from the recognition that window managers are still at a relatively immature stage and different window managers take very different approaches. Also there appeared to be little consensus on window manager design and no forum in the UK for the discussion of the issues.
Further details of the Alvey Programme are contained in the initial Annual Report .
1.2 GOALS OF THE WORKSHOP
The Alvey MMI Director requested that the workshop should address the following requirements.
- Define what should be done in the short term particularly to meet the infrastructure needs of researchers in the Alvey Programme. The Alvey Programme has made a substantial investment in Unix systems, and hence window managers under Unix systems need to be addressed.
- Define what needs to be done in the long term; it was thought that this might well have implications for the hardware architectures of future workstations, for instance.
- Look at standardization issues. Is it timely to embark upon standardization activities, and if so what should be the scope of such activities? Should activities be confined to the UK or take place in a broader arena?
The specific goals of the Workshop were to:
- reach a better understanding of the issues in window managers;
- establish where, if anywhere, there is a consensus;
- identify significant topics for further research;
- produce a written record of the proceedings.
The Alvey MMI Director asked Bob Hopgood of the UK Science and Engineering Research Council's Rutherford Appleton Laboratory (RAL) to organize the Workshop. The Organizing Committee for the Workshop was:
- Bob Hopgood (Rutherford Appleton Laboratory)
- William Newman (Alvey Software Engineering)
- Austin Tate (Alvey Intelligent Knowledge Based Systems)
- George Coulouris Alvey Man Machine Interface)
- Ken Robinson (Rutherford Appleton Laboratory)
- Tony Williams (Rutherford Appleton Laboratory)
Attendance at the Workshop was by invitation only. At an early stage in planning the Workshop it was realized that there was insufficiently broad experience in the UK of actually building window managers and hence it was felt to be essential to invite some participants from the USA. The participants reflected a wide range of interests, including academic, government and industrial backgrounds. Prior to the Workshop, participants had been told that the aims of the Workshop were to:
- identify issues in the design of the user interface, such as interaction styles, methodology needed, target populations and so on;
- identify requirements and problems in the programming interface, such as the sharing of tasks between the window manager and the application program, levels of functionality to ensure portability of applications, etc;
- identify constraints and requirements of the underlying hardware, including the division of functionality between hardware and software;
- define, as far as possible, requirements for a standard window manager, suitable for use on a range of equipment, both to provide a vehicle for research work and also to provide target software for manufacturers to supply and support.
Participants were asked to provide a short (1 page) paper describing their interests, background and the issues in window management that needed to be tackled.
From these papers, lists of issues in the user interface, program interface and architecture areas were extracted and circulated prior to the meeting (see Chapter 14 for more details). Participants also received copies of background material, references [7, 13, 15-17, 19, 20, 34, 39, 42, 44, 47, 51, 58, 64-67] from the bibliography. Thus participants came to the Workshop with at least this background in common.
1.4 STRUCTURE OF THE WORKSHOP
The structure of the Workshop followed the same pattern as the historic workshops at Seillac, France, on Methodology of Computer Graphics  and Methodology of Interaction . A number of the participants had in fact been involved in these workshops.
The first day was given over to invited presentations from seven speakers. These were intended to set the scene, both in terms of historical perspective and current research activities.
After the invited presentations, the Workshop split into three Working Groups, initially focussing on the Application Program Interface (API), User Interface (UI) and Architecture of Window Managers respectively. The Working Groups were intended to meet in parallel sessions during the second day, with plenary sessions at the middle and end of the day for reporting back. It was envisaged that some regrouping and redirection might be necessary in the light of progress made and indeed, as recorded in Part III of this book, this turned out to be the case. The format for discussing issues in the Working Groups derives from work in the standards area surrounding GKS and other graphics standards.
The third day was set aside for preparation of Working Group reports and formulation of conclusions. It is in the nature of a workshop that it is unwise to make too rigid plans in advance for the end of the workshop and in the event further Working Groups were formed which met on the final day and there was a lengthy plenary session to discuss the progress made and formulate conclusions. The issues lists presented in section 14.5 formed the basis for the Working Group discussions.
The invited presentations form Chapters 2-8 of this book and all but Chapters 3 and 5 were produced from transcripts of the presentations. Presenters were given the opportunity to correct these before publication. Chapters 9 to 13 are papers submitted by participants which were not actually presented, but were circulated as additional background material. Colin Prosser had originally intended to present the paper in Chapter 10, but at the last minute decided instead to address the issue of introducing window managers to Unix systems, because he felt that the issues raised in Chapter 10 would be better addressed by the Working Groups. His actual presentation forms Chapter 2 of this book. It is worth noting that although the paper is written with Unix operating systems in mind, the issues raised apply much more generally than to Unix systems alone, and turned out to be topics around which a lot of discussion centred.
Throughout the Workshop SUN, PERQ and Whitechapel MG-1 systems were available for participants to explore their respective window managers. The SUN system ran both SunWindows and Andrew; the PERQ ran the PNX window manager, while the Whitechapel ran their new window manager.
A number of terms are used in different ways by different authors and speakers in this volume. For example, the term user can mean the applications programmer or it can mean the workstation operator.
Terminology within window managers is also a major problem in that each system seems to have its own terminology at present. Towards the end of the Workshop a Task Group was set up to look, inter alia, at this issue.
In this volume, the authors' and speakers' own terminology has been used, no attempt has been made to incorporate a consistent terminology. It was felt that since terminology was an issue at the Workshop it would have been entirely wrong to factor in a post hoc rationalization of terminology.
A glossary of the more important terms is presented in Chapter 25.