Awareness as a mechanism for supporting the learning of database browsing techniques

Dr M.B. Twidale, Dr. J.A. Mariani, Dr. T. Rodden, Dr. P. Sawyer

Introduction

The proliferation of databases both in academic and commercial environments has led to a demand for skills in navigating and manipulating them. This demand can only grow in the future as the number and quality of databases improves and they become a standard information resource for all organisations. These skills will be needed both by specialist information scientists but also by all professionals, whose daily work will increasingly require the navigation of computational information spaces. Databases are a key information technology, and interfaces that permit their more effective use are a vital component.

The Higher Education sector needs to address the teaching of database browsing techniques for two interlinked reasons. Firstly, databases will be increasingly used within the sector to improve the quality of learning and research. This applies not just to postgraduate students undertaking literature searches (perhaps the most common activity at the moment) but also at the undergraduate level over most subject areas. Secondly, industry and commerce will be using databases more and will expect graduates (who may be deemed as general information professionals) to possess transferable skills in database navigation just as they are currently expected to possess such skills in critical analytic reading, writing and presentation of arguments and ideas.

Furthermore, the rapid increase in academic and commercial interest in Computer Supported Cooperative Work (CSCW) indicates the growing realisation that most activities have a cooperative aspect that is not allowed for in existing computer applications. Similarly the growing interest in group-based learning in the Higher Education sector reflects the awareness that the transferable skills involved in cooperative learning are ones that we should be encouraging students to acquire both for their future careers and also to enhance the quality of their current learning. Our proposal is to perform a preliminary analysis of the difficulties of novice users of databases and to develop a prototype system to support the collaborative browsing of databases made possible by awareness of others' activities in the information space. This development will use an iterative design methodology involving rapid prototyping and continual evaluation of the usefulness of the evolving system by learners from a variety of backgrounds. Although the issues in database browsing support are domain independent, we believe it is vital to use concrete examples both for clarity of exposition of our proposal and for our own benefit in order to avoid the temptation to obscure real-world problems by overly generic and ambiguous thinking. Therefore throughout the rest of this proposal we will take as our example database a library information system. Such a system already exists at Lancaster and will provide us with data for our preliminary analysis of novices' difficulties. The advantage of such a computer-based database is that it can be browsed at one's desk away from the physical information repository (the library). The disadvantage is that unless reproduced in the virtual information space, we lose some of the advantages of a physical information space including ease of use and awareness of other users.

Illustrative scenarios

We envisage a number of scenarios in which the system might be used in an educational context:

  1. Expert consultancy. A learner may be browsing the database and decide that she is not making the progress she would like. If in a physical library she might go and talk to the subject librarian. For the database use, she may wish to communicate by telephone or email (or even ultimately by video link). By use of the awareness features the expert could call up on the screen a visualisation of the student's search so far and so compose suitable advice, both specific to solving the task in hand and also explain some generic browsing techniques that the learner might consider for future tasks, as well as correcting any misconceptions made apparent in the search attempt. The expert can pass a search procedure to the student by showing on her terminal how to do this search, or how the expert would go about attempting a similar search by means of the same awareness mechanisms. Note that this interaction may be synchronous (both the learner and expert working at the same time) or asynchronous (learner and expert leaving messages for each other using both text and the database awareness features)

  2. A 'language laboratory' for database skills. A number of individual learners are browsing a database as part of a practical class on database browsing skills. They may all be in the same teaching room, or working remotely on their own terminals elsewhere. The expert uses her terminal to get an awareness of the browsing activities of an individual student. If that student seems to be progressing well she moves on to the next student. When she comes to one having difficulty she can choose to intervene either via the awareness mechanism or, if they are in a teaching room, face to face.

  3. Collaborative browsing for learners. Pairs of learners (or other small groups) who have a similar browsing task may make use of the awareness mechanism to monitor the progress and activity of their peers. This is known to be a useful learning activity both for the questioner (who asks questions such as "Why are you doing that?") and for the respondent who has to reflect on her action in order to generate a suitable explanation. Again there are two possible modes of use, synchronous and asynchronous.

  4. Serendipitous meetings. If the groups in the above scenario are made larger, database awareness can be a useful tool for supporting serendipitous meetings of the kind that naturally occur in physical libraries where strangers meet by browsing the same bookstack and finding common interests for collaboration, or when finding that a book they want is already being used by someone else, who therefore might be worth talking to. With the computational tool one could become aware when others are browsing the same parts of the database.

The system to be developed

We propose developing a system to promote awareness of activity in the cooperative use of a database. Until recently, the development of database systems has focussed on providing an interface for the user such that she need not (indeed cannot) be aware of anyone else use of the system either currently or previously. Even if many other people are accessing the system simultaneously, the user is given the illusion of having the entire database to herself. Although this may be very useful in many circumstances, work in the domain of CSCW has shown that there are cases where it is useful or even necessary to be aware of what one's co-workers are doing [Rodd92].

We wish to develop a system and in particular an interface that enables the users of a database to be aware of the activities of others for educational purposes such as to learn more about the activity of database browsing by watching others in action. In the general case, one would wish to be aware of when data was changed by others as well as when it was being browsed, but for this project we shall only consider issues relating to the latter.

Awareness issues

A previous study of the use of information within a collaborative setting [Hugh93] and a requirements analysis of the information storage mechanisms within CSCW systems [Mari91] form the starting point for this project. They reveal the following issues for supporting and representing awareness.

Reactive data storage:
A central feature of cooperative applications is the propagation of activity between users. In the case of shared information systems an important part of this is an awareness of when information is being used. Embodying the information store with reactive properties allows the promotion of awareness.

Awareness Levels
Different kinds of user may need to be aware of each other to a different degree, according to the nature of the task. While aware of the desirability of an easily reconfigurable system, for this study, we shall touch on awareness levels by use of user types in different learning scenarios.

Displaying awareness information
A key activity will be the investigation of suitable ways of displaying awareness information. This will rely on iterative development and frequent testing.

A general problem in CSCW systems is information overload. Clearly, users will not (and should not) want to be swamped with awareness information for every event that occurs within the database. Among other problems, they may miss events which are of particular importance to them. Following from existing solutions we propose a filtering mechanism. For this project we will provide filters for users dependent on their background (eg novice or expert) and the nature of the activity to be undertaken. This will provide information for the eventual development of user- configurable filters.

Synchronous and asynchronous working
We must provide awareness mechanisms for both. Asynchronous awareness is a particularly interesting problem. It will require temporal information to enable users to better understand the meaning of the actions of others. We propose (using the example of shared editing [Fish88]) a smooth animated play-back of the changes made while the user was absent. Again, however, we face the problem of information overload, and again we propose to use filters. Play back is also useful for educational purposes such as running through an example of an expert's behaviour or encouraging reflection and hence learning from one's previous performance.

Supporting Informal Communication:
Experience in existing CSCW systems suggests that strictly prescribed communication mechanisms amongst collaborators are not popular; indeed they are often subverted by the users who quickly migrate to the simplest (perhaps incorrect within the context of the mechanism) way of communicating. We would like to support a communications channel outside of the system rather than a rigid method of co-ordination. Again, at the simplest level, this could be text-based but we would like to investigate at least an audio channel, and eventually possibly video (we are fortunate that Lancaster has excellent multimedia resources with the recent installation of a campus-wide ATM switch). Of importance here is the inter-relation between the communication medium and the information stored and the level of support required within a cooperative system.

In summary, the purpose of the proposed research is to develop techniques to support interaction with a shared object store by exploiting the notion of cooperative browsing.

Background

The authors are unaware of any work which is specifically looking at the synergistic combination of the different perspectives on database browsing to be examined within this project. However, a variety of work within various research communities have examined a number of issues associated with each of the areas to be tackled. Work of particular note to this project is briefly reviewed below.

Cooperative interfaces to shared information
A wide range of multi-user interface architectures and systems have been developed. These systems adopt an approach which is either cooperation transparent or cooperation aware [Lauw90]. A number of recent system have extended the facilities by cooperation aware applications [Rein91] to develop frameworks which allow a range of applications to present different views onto shared information views. Systems adopting this approach include the Rendezvous system [Patt90] and the applicants' own work [Bent92] which provide the ability to support different user views on application objects and interactions with these views. In both case this is enabled by a separation between underlying objects, which contain the application semantics and interaction objects, which maintain details of users' views on these objects. One project has considered how co-authors maintain a consistent database: their bibliographic references [Musl92].

Reactive databases.
The augmentation of data bases with rules which react to and constrain changes to data has been a focus of much database research. Such active database rules are frequently characterised as constraints (rules which generate some exception on the occurrence of some condition - e.g. attempting to modify an immutable attribute) and triggers (rules which initiate some action when a condition becomes true - e.g. to automatically derive attribute values). Several investigators [Bloo87, Mede91] have investigated the integration of rules into the database object model where they may be invoked when, for example, an attempt is made to update an object attribute.

Active database rules are usually proposed as a means of maintaining database consistency. It is argued that such consistency constraint mechanisms are better defined at the data level than in the applications which manipulate the data. We propose to apply them to the enabling of awareness using the same argument; where objects are shared, rules can be employed as a mechanism for signalling users' activity.

Information browsing activities
A number of studies have been made of the browsing behaviours of various kinds of users with numerous information systems eg. [Schl86]. In particular a considerable body knowledge has been accumulated on the problems and strategies of browsing hypertext systems, for example; [Egan89, Remd87].

Impact of the programme on educational content and delivery

The evolving system will provide a rich supportive learning environment affording a more thorough understanding of the nature, purpose and mechanisms of database browsing. An interface that allows the articulation and subsequent play- back of the process of browsing can facilitate reflection; known to be important in learning but often very hard for students to undertake. A sophisticated interface can provide a microworld learning environment to facilitate learning and transfer to the less supportive environments of conventional database interfaces.

We envisage the teaching of browsing techniques to be a key part of the study skills that students will learn irrespective of subject domain. Furthermore, spin-offs from the main activity of the project will include recommendations for more effective teaching of the use of traditional database query systems, and recommendations for more appropriate interfaces of such systems. Easier to learn and use database systems are more easy to incorporate in high quality learning activities devised by subject domain practitioners, and to distract less from the particular discipline being studied. Besides the educational benefits, the project will advance our knowledge in the development of more effective database systems, better interfaces for browsing and their use in cooperative settings.

Proposal of work

We believe the proposal to have many novel features and to raise intriguing research issues. Therefore we feel it necessary to have a very cautious work strategy in order to avoid being seduced by the challenge and excitement of developing a very innovative system. One of the lessons of the history of funding for AI and for many forms of educational technology is that in order to secure funding researchers were tempted to make extravagant claims that inevitably proved impossible to fulfil. As a consequence we propose a series of activities that can yield a stream of results within the framework of the admittedly ambitious envisaged system.

Firstly, it has been noted that research in intelligent tutoring systems can also yield spin-off results in terms of recommendations for more effective traditional computer based teaching and even conventional face-to-face teaching. In a similar manner the results of our analysis of database querying and continual evaluation of the evolving system mean that we will be able to make suggestions for more effective ways of teaching general browsing skills on existing systems.

The same analysis will also lead to recommendations for more effective interfaces and query options for databases, even those that do not support awareness. This can immediately feed into the continuing evolution of such interfaces by developers such as those of the library information systems developers at Lancaster. The massive increase in databases available mean that such developers often do not have the time to analyse learner's and user's interactions and difficulties in order to improve the interface. As a consequence, such interfaces can be quite poor. This was less of a problem when the only users of such systems were skilled computer and information scientists. However, the more widespread availability creates a need for much more supportive and easy to learn and use interfaces for databases, similar to the way that interfaces to word processors have improved, accompanying their more widespread usage.

The first version of these recommendations can be achieved in the first year of the project and naturally these results will be straightforward to make available to other HE institutions by the conventional means of information transfer for ideas, namely journals, conferences etc.

Although ultimately a completely generic domain independent system is the goal of any computer systems research project, we believe that for this proposal, less ambitious but more realistic immediate goals are more appropriate. Hence we propose incrementally developing a system that is capable of handling a growing number of usage scenarios. The experience gained from this activity can be used to inform a subsequent project to develop the generic, configurable version.

Local Baseline

This proposal draws on the wide range of interests of the proposers. It involves the use of ideas from CSCW, Database design, AI+Education and Interface design. It builds upon the emerging results from a number of related projects at Lancaster:-

References

[Bent92] Bentley, R., Rodden, T., Sawyer, P., Sommerville, I.: An architecture for tailoring cooperative multi-user displays. In: Proc. CSCW92, Toronto, 1992.

[Bloo87] Bloom, T., Zdonik, S.: Issues in the Design of Object-Oriented Database Programming Languages, in Proc. OOPSLA '87, October 1987, Orlando, Florida.

[Egan89] Egan, D.E., Remde, J.R., Landauer, T.K., Lochbaum C.C. & Gomez, L.M.. 1989. Behavioral Evaluation and Analysis of a Hypertext Browser. In Proceedings of ACM CHI'89 Conference on Human Factors in Computing Systems, 205-210.

[Fish88] Fish R.S., R.E. Kraut , M.D. Leland and M. Cohen. 1988. Quilt: A collaborative tool for cooperative writing. In COIS88 Published in SIGOIS Bull. 9(2-3):30-7.

[Hugh93] Hughes J., M. Twidale, J. Mariani, T.Rodden and V. King "Paperwork and its lessons for database systems" , Proceedings of 12th Schaerding IFIP Workshop on Design Of Computer Supported Cooperative Work And Groupware Systems, June 1-3 1993

[Lauw90] Lauwers, J. C., Lantz, K. A., 'Collaboration awareness in support of collaboration transparency: Requirements for the next generation of shared window systems', in Proceedings of CHI '90, pp 303-311.

[Mari91a] Mariani, J.A., Rodden, T., "The Impact of CSCW on Database Technology", pp 146-161, International Workshop on CSCW, K. Gorling and C. Sattler (Eds), Informatik informationen-reporte, 4/1991, April 9-11, 1991, Berlin.

[Mari92] Mariani, J. Oggetto: An Object Oriented Database Layered on a Triple Store. The Computer Journal, 35 (2), April 1992.

[Mede91] Medeiros, C.B., Pfeffer, P.: Object Integrity Using Rules, in Proc. ECOOP '91, July 1991, Geneva.

[Musl92] Musliner, D.J., Dolter, J.W. & Shin, K.G. in Proceedings of ACM CSCW'92 Conference on Computer-Supported Cooperative Work 386-393 1992.

[Patt90] Patterson, J. F., Hill, R. D., Rohall, S. L., Meeks, W. S., 'Rendezvous: An architecture for synchronous multi-user applications', in Proceedings of CSCW '90, October 7-10, Los Angeles, Ca., ACM, 1990.

[Rein91] Rein, G. L., Ellis, C. A., 'rIBIS: A real-time group hypertext system', International Journal of Man Machine Studies, 34(3), March 1991, pp 349-368.

[Remd87] Remde, J.R., Gomez, L.M. & Landauer, T.K. 1987. SuperBook: An Automatic Tool for Information Exploration - Hypertext? In ACM Hypertext'87 Proceedings, 175-188. [Rich90] Richardson, I. In-house library systems development at the University of Lancaster. Vine 81, Dec 1990, 24-26.

[Rodd92] Rodden T., Mariani J., Blair G 'Supporting Cooperative Applications' International Journal on CSCW, Vol. 1, No. 1, October 1992, pp 1-27.

[Schl86] Schlager, M.S. & Ogden, W.C. in Proceedings of ACM CHI'86 Conference on Human Factors in Computing Systems 107-113 1986.

[Sawy92] Sawyer, P., Colebourne, A., Sommerville, I., Mariani, J.: Object-oriented database systems: a framework for user interface development. In: Proc. IDS'92, Glasgow University, July 1992.

[Twid93a] Twidale, M. B. (1993a). The use of multiple instantiations in interactive learning environments. Proceedings ICCE '93, Taiwan.

[Twid93b] Twidale, M. B., Rodden, T. & Sommerville, I. (1993b). Developing a tool to support collaborative dialogues and graphical representation of ideas. To appear in Verdejo, F. (Ed) Collaborative dialogue technologies in distance learning, the proceedings of the NATO ARW, Segovia, Spain

[Twid93c] Twidale, M. B. (1993c). Redressing the balance: the advantages of informal evaluation techniques for Intelligent Learning Environments. Journal of Artificial Intelligence In Education. 4 (2/3) 155-178.