Split Window View and Modification Director

Split Window View and Modification Director: Innovative Awareness Mechanisms in Real-time Collaborative Writing

Keywords: Awareness mechanisms, Split Window View, Modification Director, CSCW, real-time collaborative writing.

Group awareness is a vital feature improving the usability of real-time collaborative writing systems. Unfortunately, existing awareness mechanisms were derived from a solely “bottom-up” approach, without prior research on what awareness information users really need. Previous research fails to provide a comprehensive set of awareness mechanisms, and some mechanisms were even implemented in an ad-hoc manner. Applying the principle of user-centred design, this research exploits a “top-down” approach in identifying awareness mechanisms by conducting a laboratory-based usability experiment with REDUCE (Real-time Distributed Unconstrained Cooperative Editing). This research has identified quite a few innovative mechanisms, such as Split Window View (SWV) and Modification Director (MD). SWV provides key information about others’ working and viewing areas. MD conveniently and instantly notifies users when their work is modified by others.

Real-time collaborative writing systems allow distributed authors to work on documents at the same time. Examples of these systems include GROVE (Ellis et al., 1991), REDUCE (Yang et al., 2000), SASSE (Baecker et al., 1993) and ShrEdit (Dourish and Bellotti, 1992). Maintaining group awareness (GA) is a pivotal feature enhancing the usability of real-time collaborative writing systems. GA provides users with sufficient knowledge about the status of a document itself and all activities other users perform upon the document. GA plays an essential and integral role in cooperative work by simplifying communication, supporting coordination (Ellis et al., 1991), assisting “anticipation” (Gutwin and Greenberg, 2002) and supporting “convention” (Grudin, 1994).

In face-to-face interaction, it is naturally easy for people to know who is present, what others’ responsibilities are and what others are doing. However, when group members are geographically dispersed, supporting spontaneous interaction is much more difficult. To enrich GA in real-time collaborative writing, various awareness mechanisms such as telepointers (Greenberg et al., 1996), miniature and radar views (Gutwin et al., 1996) and multi-user scrollbars (Baecker et al., 1993) have been used. Unfortunately, these mechanisms were implemented in editors without prior research on what awareness information users need when writing collaboratively, i.e. a solely “bottom-up” approach. Consequently, some awareness mechanisms are implemented in an ad-hoc manner and some may even generate contradictory results. Although using this solely bottom-up approach has exposed some relevant awareness mechanisms, some other possible mechanisms are easily overlooked or require excessive experimentation to be found. For instance, the emergence of a radar view reported in (Gutwin et al., 1996) is the result of a long process of improvement and enhancement, which could have been naturally avoided if the process producing the radar view initiated from end-users. Furthermore, by using only bottom-up approaches, previous research does not provide a comprehensive set of awareness mechanisms; designers are often left without any clear sense of which awareness mechanisms should be implemented in real-time collaborative editors to support GA.

To produce a usable system, a designer must be directed by the principles of user-centred design (Baecker et al., 1993; Mayhew, 1999). Applying the user-centred design philosophy, this research exploits a “top-down” approach in the study of users’ needs and identifying highly suitable candidates for awareness mechanisms. The top-down approach involves conducting a laboratory-based usability experiment with REDUCE¾Real-time Distributed Unconstrained Cooperative Editor (Yang et al., 2000). REDUCE is a real-time and multi-user text editor allowing geographically distributed people to interact synchronously upon the same document without constraints (Figure 1). Unlike the solely bottom-up approach used by previous research in finding awareness mechanisms, the top-down approach effectively ensures that awareness mechanisms are developed in the light of users’ needs. By using the top-down approach, problematic concerns are identified at early stage of development; therefore, excessive evaluation normally required in the solely bottom-up approach is now avoided.

Some novel awareness mechanisms have been discovered gradually by this research using the top-down approach. After the first few usability experiments, this research has identified two awareness mechanisms, namely Dynamic Task List (DTL) and Toggle Multi-user Scrollbar (TMS) (Tran et al., 2001). Further experiments have led to the discovery of Split Window View (SWV) and Modification Director (MD). These mechanisms were proposed by end-users and have not yet been implemented. Hence, the mechanisms presented in this paper are mock-ups and still require further evaluation and improvement before being implemented in any real-time collaborative writing systems. The main emphasis of this paper is to describe SWV and MD, and discuss their strengths and weaknesses in enhancing GA for real-time collaborative writing. This paper begins with examining limitations of some major existing awareness mechanisms. It then describes a research methodology, and a top-down usability experiment used by this research.

The growing interest in GA results from the fact that awareness support has increasingly been identified as a crucial part of successful collaboration (Dourish and Bellotti, 1992; Rodden, 1996; Gutwin and Greenberg, 2002). Perceiving and understanding the activities and intentions of other members of a collaborating ensemble is a basic requirement for human interaction. As in any groupware system, GA is a major concern in real-time collaborative writing. A number of different awareness mechanisms supporting collaborative writing have been used by previous research. This section examines limitations of three commonly used existing awareness mechanisms:

· Telepointers: First, telepointers are only capable of providing other users’ mouse positions when all users locate at the same portion in the document. Consequently, if telepointers are removed from a local user’s view, it is extremely hard for the local user to gauge a location at which a remote user is working and how active that user is. Second, another issue in implementing telepointers is disruption; telepointers could be distracting to a local user, especially when remote users perform a substantial amount of movement. Third, in collaborative writing, telepointers fail to convey information about sections in a document on which a remote user are currently working, since the position of a telepointer is often not the same as a position where a remote user is typing.

· Radar view: Although a radar view has proven to be useful in maintaining group awareness (Gutwin et al., 1996), two major associated problems need to be overcome: low-fidelity and disruption. The fidelity of the representation affects the visibility of contents shown in a miniature view. The major problem with a miniaturisation technique is that it has limited scalability; a miniature view of an extremely large data space contains too little detail to be useful. Additionally, a radar view can be distractive. For example, when other users perform a substantial amount of movement, the views of rectangles in the miniature view change constantly, causing a local user’s screen to become overly-busy.

· Multi-user scrollbars: In the literature, there are two different variations of multi-user scrollbars: version 1 in (Baecker et al., 1993) and version 2 in (Gutwin et al., 1996). In version 1, each remote scrollbar is located in a different vertical region; however, in version 2, all remote scrollbars are located in the same vertical region. The major problem of multi-user scrollbars occurs when a large number of users are working in the workspace. In version 1, the display of a large number of remote users’ scrollbars cause space constraints; it forces the area of the document portion viewed to be reduced. In the case of version 2, when views of more than two users intersect, it is hard to know exactly the location of remote users because many remote scrollbars overlap one another. Another issue with multi-user scrollbars is identification; to identify which remote scrollbar belongs to whom, a user must map a colour with a user. Mapping is very ineffective when many users are in the workspace, as a user must memorise many pairs of users and their scrollbar colours. If a user’s name is attached to an associated remote scrollbar, it is much easier in identifying the owner of a remote scrollbar (Tran et al., 2001).

To design usable awareness mechanisms that are enthusiastically accepted by users, designers should not start from functions that designers want to provide. In contrast, designers should start from the study of users’ activities and needs, which are only identified by a user-centred approach. The next section describes a research methodology, an integrated approach¾a top-down approach merged with a bottom-up approach, which is used by this research to determine awareness mechanisms in real-time collaborative writing.

To produce a usable system supporting collaborative writing, much research has exploited the user-centred approach in the study of how people write together (Posner and Baecker, 1992; Baecker et al., 1993). Similarly, to provide usable awareness mechanisms for real-time collaborative editors, a designer must be directed by the principles of user-centred design. Applying the user-centred design philosophy, this research manipulates a “top-down” approach to determine mechanisms supporting group awareness in real-time collaborative writing systems. The essence of the top-down approach is to determine users’ needs and identify highly applicable candidates for awareness mechanisms by performing a laboratory-based usability experiment. In applying the top-down approach, this research conducted an experiment with REDUCE. At present, REDUCE supports almost no GA features; hence, conducting the experiment with REDUCE allows determination of awareness information users really need to perform a collaborative writing task effectively and efficiently. The results of the experiment lead to the discovery of applicable awareness mechanisms that are highly capable of supporting GA in real-time collaborative writing.

Given awareness mechanisms derived from the top-down approach, this research will then implement necessary awareness mechanisms in REDUCE. After this is carried out, a bottom-up approach will be used to evaluate the effectiveness of these mechanisms in supporting GA and to determine further awareness support. The bottom-up approach involves carrying out another series of laboratory-based usability experiments with REDUCE, but in this case with the mechanisms implemented. However, this paper focuses on the work of the top-down approach in determining awareness mechanisms.

The usability experiment involved ten pairs of subjects working on three writing tasks, including creative writing (CW) (e.g., writing short essays from scratch), document preparation (DP) (e.g., writing research papers) and brainstorming (BS) (e.g., generating ideas). This research used these three categories for two main reasons. Firstly, these categories represent a wide range of collaborative document editing tasks. Secondly, the categories require different styles of collaboration. The types of awareness mechanisms that are needed in different contexts of collaborative writing are found by using these varied tasks. The ten pairs were allocated to perform the three tasks in order: 4 pairs worked on CW, 3 pairs worked on DP and 3 pairs worked on BS.

Subjects included lecturers and Ph.D. students in Information Technology (IT) who are familiar with cooperative work and have well-established computer skills, but are not acquainted with group awareness. Intentionally, the research selected lecturers and Ph.D. students in IT to be experimental subjects because their experience in cooperative work and their well-established computer knowledge would give worthwhile comments and feedback. Since these subjects have an excellent technological understanding, they are capable of advising highly applicable awareness mechanisms. Particularly, they are able to provide mock-up user interfaces of proposed awareness mechanisms, and describe how the mechanisms function; this ability is hardly provided by people without good technological understanding. Determining innovative awareness mechanisms is the major contribution of this research; hence, choosing technologically-trained experimental subjects is key to the experiment. Subjects worked in pairs and were located in two separate subject rooms. Each pair participated in a two-and-a-half hour session that included:

· Training (30 minutes): Each subject was fully trained in using REDUCE to ensure that they were familiar with the system and confident in collaboration.

· Experiment (1 hour): Subjects worked in pairs to work on one task with verbal communication for thirty minutes and on another task without verbal communication for thirty minutes. Conducting the experiments with and without support of verbal communication allowed identification of problems users had and the workarounds users resorted to when verbalisation was absent.

· Questionnaire and interview (1 hour): Subjects filled in a questionnaire, which included twenty five-point scale questions and seventeen open-ended questions. Subjects also took part in an interview to discuss awareness information and awareness mechanisms they needed when perform writing tasks collaboratively.

The next two sections describe two innovative awareness mechanisms identified by this research: Split Window View and Modification Director. Although the research experimented with a group size of two, the mock-ups of SWV and MD show what happens with groups consisting of more than two users.

In collaborative writing, authors’ working areas can be different from their viewing areas (Benford et al., 1995). Such awareness mechanisms as multi-user scrollbars, radar views and telepointers show users’ current viewing areas rather than their working areas. The results of analysing the five-point scale questions show that it is important to provide a user with information about other users’ working and viewing areas. Interestingly enough, information about others’ working areas was considered slightly more important than about others’ viewing areas as shown in Figure 2 and Figure 3. As a result, it is crucial for real-time editors to have mechanisms that provide users with such information. Split Window View (SWV) enables a user to find quickly both other users’ viewing and working sections (Figure 4). When remote users change their viewing or working regions, their views displayed in the split window are updated immediately. In SWV, remote insertion cursors are also added to the display. Remote insertion cursors assist a local user in tracking down easily and precisely other users’ working areas. Figure 4 shows a mock-up of Peter’s screen including: Peter’s editor (Peter is a local user), Jun’s view (where Jun is currently looking at in the document), Kim’s editor (where Kim is currently working in the document), and Jim’s view and Jim’s editor.

Figure 2. Distribution of subjects’ responses to question, “Knowing parts of a document on which other users are currently working”

Figure 3. Distribution of subjects’ responses to question, “Knowing parts of a document at which other users are currently viewing”

Information overload is a general problem in an electronic environment; if too much information is presented in interfaces, it becomes difficult for users to concentrate on the essential aspects of their work. Not only does SWV allow a user to observe more than one users’ work at the same time, but it also enables the user to request explicitly whether or not the user wants to retain other people’s views by clicking a button on the user panel. By doing so, the presentation of awareness information is controlled by users.

Although SWV allows a user to watch other collaborators work simultaneously, this mechanism raises two problems in display and design issues: space constraints and display fidelity. First, since the visible size of a screen is limited, a problem occurs when having multiple views in a single window screen at the same time. The more users that are viewed, the smaller the size of the main editor. This limits the local user’s main view of the workspace. Second, when both viewing and working areas of a particular user are displayed (e.g., Jim’s view and Jim’s editor are viewed at the same time in Figure 4), SWV might need to implement a low-fidelity presentation of the document in order to fit the entire view in a window of half normal size. Depending on the fidelity, the contents in miniature views can be difficult to read and to understand. Besides that, with low-fidelity views, users might have difficulty seeing the remote insertion cursors. If a low-fidelity presentation is avoided, horizontal scrollbars in each view are added.

In many cases, subjects did not realise that the other subject modified their document. Furthermore, the results of five-point scale questions show overwhelming support for the need to be aware of other users’ current actions: 95% of the respondents considered it important (Figure 5). These results strongly support the need for an awareness mechanism like the Modification Director (MD).

Figure 5. Distribution of subjects’ responses to question “Knowing what actions other users are currently taking”

Assisting users in monitoring who edits their work and which parts of their work are being modified is the inspiration for the emergence of MD. Figure 6 shows a mock-up of MD, including a local user’s (Kim’s) screen. On a user’s screen, there is an additional window, called Users showing:

· users’ coloured icons, which are the same colour as the users’ corresponding text colours, and

· information buttons¾a user can click an information button to view information about a particular user such as their geographical location, duration in the workspace, or work background.

Figure 6 illustrates how MD functions by showing an example when one user’s (Kim’s) work is modified by another user (Tom). When Tom modifies Kim’s text, a corresponding text colour icon that indicates Tom is modifying text flashes instantly on Kim’s screen. To view an area which is modified by Tom, Kim can click the flashing icon to open up a read-only window. This read-only window contains the modified area. The pop-up window also contains Tom’s insertion cursor, which allows Kim to find the exact location of the modification easily.

The major advantage of MD is that it notifies users immediately whenever their work is modified by somebody else. In addition, MD allows authors to know which part of their work is altered and who makes the modification.

Although a pop-up window is used in MD, the window is controlled by users; the window only pops up when a user clicks a flashing icon. Hence, disruption is significantly minimised, because a user is aware of the presence of the pop-up window.

A design issue needs to be addressed when implementing MD; if user A modifies more than two different sections of user B at the same time, which section should be shown in the pop-up window? One solution to resolve this problem is to organise the pop-up window in page format, which allows users to go backward and forward amongst the modified areas.

This paper has described the authors’ research in determining mechanisms supporting group awareness in real-time collaborative writing based on a “top-down” approach. Using a top-down approach in identifying awareness mechanisms has been ignored by previous research. This paper has described two novel awareness mechanisms: Split Window View (SWV) and Modification Director (MD).

In real-time collaborative writing, none of the existing awareness mechanisms distinguish between a user’s viewing and working areas. Although viewing and working areas are usually the same, in certain cases they can be different. SWV supports users in gauging easily other users’ viewing areas and working areas in the shared document. SWV also allows a user to choose explicitly specific remote users whom they want to observe.

MD supports group awareness (GA) by notifying a user instantly whenever their work is modified by other people. This awareness support is also not provided by existing mechanisms. MD allows a user to view a modified area simply by clicking on a flashing icon. In synchronous collaborative writing, one author’s modification of another author’s work often occurs; therefore, having such a MD is vital because this mechanism allows an author to know if and how their original work is changed by other authors.

Since previous research exploited the solely bottom-up approach in finding awareness mechanisms, none of that research is able to differentiate the importance of different awareness elements. For example, knowing information about others’ working areas is more important than about others’ viewing areas. However, by using the top-down approach, this research has contributed to understanding of GA by successfully distinguishing the importance of different awareness elements.

Most existing awareness mechanisms do not provide an ability to control the visibility of mechanisms. However, mechanisms identified by this research allow users to control the visibility of the mechanisms. Therefore, users are not distracted from their work by excessive presentation of the awareness mechanisms. The awareness mechanisms in this paper both justify the strong need for GA and reveal highly useful mechanisms that have not been found from past research.

In the future, a conceptual framework for GA will be formed based on the results of the usability experiment. The framework will be a guideline to transform awareness needs into software modes and interface tools. A “bottom-up” approach will be exploited to clarify the efficiency and effectiveness of those awareness mechanisms in supporting collaborative writing by conducting a further set of usability experiments.

We sincerely thank Professor Penny Sanderson, The University of Queensland, for her support and valuable insight of this research. We gratefully acknowledge support of the usability experiment from John Craick, the manager of SCHIL, and Professor Chengzheng Sun, Griffith University.

· Baecker, R. M.; Nastos, D.; Posner, I. R. and Mawby, K. L. (1993). The User-centred Iterative Design of Collaborative Writing Software. Proceedings of InterCHI’93, Amsterdam 24-29 April, pp. 399-405.

· Benford, S.; Bowers, J.; Greenhalgh, C. and Snowdon, D. (1995). User Embodiment in Collaborative Virtual Environments. Proceedings of the ACM Conference on Computer-Human Interaction CHI'95, Denver, Colorado, USA, ACM Press, New York, pp. 242-249.

· Dourish, P. and Bellotti, V. (1992). Awareness and Coordination in Shared Workspaces. Proceedings of the ACM Conference on Computer Supported Cooperative Work CSCW’92, October 31 – November 4, Toronto, Canada, ACM Press, New York, pp. 107-114.

· Ellis, C. A.; Gibbs, S. J. and Rein, G. (1991). Groupware: Some Issues and Experiences. Communications of the ACM, 34(1), pp. 39-58.

· Greenberg, S.; Gutwin, C. and Roseman, M. (1996), Semantic Telepointers for Groupware, Proceedings of OzCHI ’96 the Sixth Australian Conference on Computer–Human Interaction, IEEE Computer Society Press, pp.54-61.

· Grudin, J. (1994). Groupware and Social Dynamics: Eight Challenges for Developers. Communications of the ACM, 37(1), pp. 92-105.

· Gutwin, C. and Greenberg, S. (2002). A Descriptive Framework of Workspace Awareness for Real-Time Groupware, Computer Supported Cooperative Work, the Journal of Collaborative Computing, Kluwer Academic Press.

· Gutwin, C.; Roseman, M. and Greenberg, S. (1996). A Usability Study of Awareness Widgets in a Shared Workspace Groupware System. Proceedings of the ACM Conference on Computer Supported Cooperative Work CSCW’96, Boston, MA, ACM Press, New York, pp. 258-267.

· Mayhew, D. (1999). The Usability Engineering Lifecycle: A Practitioner’s Handbook for User Interface Design. Morgan Kaufmann, Publishers Inc, San Francisco, CA.

· Posner, I. R. and Baecker, R. M. (1992). How People Write Together. In Baecker, R. M. (Ed.) Readings in Groupware and Computer-Supported Cooperative Work: Assisting Human-Human Collaboration. Morgan Kaufmann, San Mateo, CA, 1993, pp. 239-250.

· Rodden, T. (1996). Population the application: A Model of Awareness for Cooperative Applications. Proceedings of the ACM Conference on Computer Supported Cooperative Work CSCW’96, Boston, ACM Press, New York, pp. 87-96.

· Tran, M. H.; Raikundalia, G. K. and Yang, Y. (2001). Usability Experiments for Determining Document-Related Awareness in Real-time Cooperative Editing. Proceedings of Sixth Australasian Document Computing Symposium, Coffs Harbour, 7 December, pp. 95-98.

· Yang, Y.; Sun, C.; Zhang, Y. and Jia, X. (2000). Real-time Cooperative Editing on the Internet. In IEEE Internet Computing, 4(1), pp. 18-25.