A FORMAL AWARENESS MODEL FOR 3D WEBBASED COLLABORATIVE ENVIRONMENTS Pilar Herrero Computer Science School Universidad Politécnica de Madrid Campus de Montegancedo s/n 28660 Boadilla del Monte Madrid (Spain) E-mail: Phone: + 91 336 74 56 Fax: + 91 336 65 95
ABSTRACT
The importance of an awareness model for a Virtual Environment has been underwritten by many studies and very interesting research has been undertaken in this CSCW research area. In this paper, we give a new interpretation of the set of key concepts and ideas, which have been used to define conventional awareness models in CVE’s, for the purpose of building a formal model of awareness for 3D Web-based Collaborative Environments. At the same time, we will point out some limitations of current awareness models and we will propose some extensions, like Visual Acuity and Internal Filters. We will also provide an interpretation for these concepts in the context of Web applications. Keywords: CSCW, CVEs, Awareness, Virtual Reality, Visual Acuity, Internal Filter, Sense Transition Region. INTRODUCTION
Examining the current approaches to Web-based communication and collaboration, we can identify two very different solutions: Asynchronous communication and Synchronous communication. The first situation happens when different users share information through a web application without the requirement of being connected at the same time.
Angélica de Antonio Computer Science School Universidad Politécnica de Madrid Campus de Montegancedo s/n 28660 Boadilla del Monte Madrid (Spain) E-mail: <
[email protected]> Phone: + 91 336 69 25 Fax: + 91 336 69 17
In this case, the user is not aware of other users that might be connected, the feeling of immersion and shared presence is null and the degree of interactivity is very low. The second situation happens when all the users meet and interact in a multi-user collaborative environment. Some of these environments take the form of three-dimensional virtual worlds where every connected user is represented by an avatar. There can be autonomous agents populating the world too. The concept of awareness of other users takes very different meanings depending on the situation. In 3D web-based collaborative environments, awareness of other participants can take a physical interpretation, while awareness in non-graphical environments must be interpreted in a more abstract way. This paper describes the work that is being carried out at the Universidad Politécnica de Madrid with the aim of formalising the concept of awareness. The first section presents the state of the art in formal models of awareness for multi-user environments. Next, we introduce an extension of the Spatial Model of Interaction based on sense acuity, that allows a more realistic model of awareness, and an extension based on internal filters, that allows the consideration of personal traits and their effect on awareness. Finally, we discuss the possible abstract interpretations for the concepts of the awareness model in the context of asynchronous WEB applications. RELATED WORK
Perhaps the most well known awareness model for virtual environments is "The spatial Model of Interaction". This model was developed between 1991 and 1993, as a way to control the flow of information about the environment in CVEs (Collaborative Virtual Environments). To carry out it was necessary to mix the researches led by Professor Steve Benford at the School of Computer Science and Information Technology, in the Nottingham University, the researches led by Lennart E. Fahlén at The Swedish Institute of Computer Science (SICS) and the researches led by John Bowers at Royal Institute of Technology (KTH) in Stockholm (Sweden)
This model allows objects in a virtual world to govern their interaction through some key concepts: aura [Fahlén 92], awareness, focus, nimbus, adapters [3] and boundaries [5]. In 1992, Fahlén and Bowers defined aura as the subspace which effectively bounds the presence of an object within a given medium and which acts as an enabler of potential interaction [Fahlén 92]. Once aura has been used to determine the potential for object interactions, the objects themselves are subsequently responsible for controlling these interactions. Objects carry their auras with them when they move through space and objects typically have different auras in different media. “When two auras collide, interaction between the objects in the medium becomes a possibility” [3]. Perhaps, the main concept involved in controlling interaction between objects is “awareness”. One object’s awareness of another object quantifies the subjective importance or relevance of the other object. “The awareness quantifies the degree, nature or quality of interaction between two objects” [10]. “The awareness relationship between every pair of objects is unidirectional and is specific to each medium” [3]. This is achieved on the basis of quantifiable levels of awareness between them [2] Awareness between objects in a given medium is manipulated via focus, “the more an object is within your focus the more aware you are of it” [3], and nimbus, “the more an object is within your nimbus the more aware it is of you” [3]. Object's aura, focus, nimbus and hence awareness can be modified through some artefacts called adaptors [3].
introduced by Greenhalgh in 1997 [10] which is called "Third Party Objects". Third party objects and other improvements were introduced in MASSIVE-2 . Sandor developed a reinterpretation of the COMIC Spatial Model and its concepts in 1997 [15]., which was called Aether (Awareness Engine THeory and Experimental Realization). Perhaps the most recent work until this moment is QoS Architecture for Collaborative Virtual Environments, focusing on the management of streamed video within shared virtual worlds [11]. INTRODUCING SENSE AWARENESS MODEL
ACUITY
INTO
For MASSIVE-1 a standard function was defined for evaluating focus and nimbus [10] with some controlling parameters. This function (Fig. 1) depends on the location and orientation (Fig.2) of the object and uses the parameters shown in table 1.
Focus / Nimbus 1.0
Vc Vb
r0
r1
distance
Fig.1- Focus and Nimbus function relative to the distance.
Finally, aura, focus and nimbus may be manipulated through boundaries in space. For Bowers, boundaries have great importance in structuring social interaction [Bowers, 1993]. Between 1994 and 1996 The University of Nottingham continued working in this topic. Some of the concepts above mentioned were studied and redefined, for example, the concept of boundary. For Benford, “boundaries divide space into different areas and regions and provide mechanisms for marking territory, controlling movement and for influencing the interaction properties of space” [4]. More specifically, boundaries can be thought of as having different kinds of effects, which can be of four sorts: obstructive, non-obstructive, conditionally obstructive and transforming [4]. After some experimental implementations of the spatial model [7] MASSIVE-1 emerged as a “Model, Architecture and System for Spatial Interaction in Virtual Environments”, an experimental distributed virtual reality system intended to support collaborative activity [8]. A lot of researches have been carried out starting from this model; in this way Rodden in 1996 developed a new model of awareness for Cooperative Applications [13] but perhaps the more fundamental extension was
THE
Transition region (value=V c)
Foreground region (value=1.0)
Background region value=Vb
φ
θ
Forward direction
Fig.2- Orientation component
Name
decrease in the focus value is faster when the distance is around XVA.
Meaning
θ
Conical angle of foreground region
φ
Conical angle of transition region
Vb
Focus and Nimbus value of background region
Vc
Value for transition region
ro
Radius of near-field
r1
Radius of transition region
Table 1- Parameters of the standard focus/nimbus function. In our opinion these functions do not properly reflect real life for two reasons. The first one is that the focus function should be different to the nimbus function, because these two functions have different meaning. For example, if we are in a visual medium visual focus is dependent on forward direction whereas visual nimbus is the same independent on direction. The second and very important reason is that in real life we have "Sense Acuity", that is, each sense’s specific ability to resolve fine details. Perhaps the most well known is the Visual Acuity (VA), which is a measurement of the eye's ability to resolve fine detail and is dependent upon the person itself, the accommodative state of the eye, the illumination level and the contrast between target and background. If we bear in mind this concept, it is easy to see that the focus function should depend on VA. Moreover, if we consider the minimal distance XMDD necessary to see an object with fine detail, for example to read a paper, we will have three regions in our visual field (Fig.3): the first one goes from our eye (in fact from the nodal point) to XMDD. In this region (R1) it is impossible to see with clarity an object because it is too close to our eye and we are not able to mark out the object's details. This distance is a function of the object's size. The second region (R2) goes from XMDD d to XVA (the distance marked by our VA) in which we will see the object with all its details. The last region (R3) will expand from the previous distance (XVA) to the distance beyond which the visual focus value will be zero. In this region (R3) we will see the shape but it will be impossible to see the fine details. The visual focus in this region will decrease fast with the distance.
We have calculated some data for XMDD and XVA in a Distributed Virtual World (which was built with World Up) using the Sense8 system unit for size, which we will call u.s. (unit of size), considering two objects. The first object was the image in a poster of 9,27 u.s. (high) x 7,95 u.s. (wide) and the second object was the message which was written in the lower part of the poster the size of which was 3,09 u.s. (high) x 7,95 u.s. (wide). The last one contained different types of letters, but we have selected only the capital letters in bold that were written in font "Times New Roman" with size 12 points, which is equivalent to 0,5 u.s. (high) x 0,15 u.s.(wide). The data obtained for the first object are: XMDD= 100 u.s. and XVA=230 u.s. Between this two positions, the observations confirm that the focus value decreases gradually. For the same reason we can prove that for X> XVA, with X close to XVA, the decrease is bigger. INTRODUCING INTERNAL FILTERS INTO T H E AWARENESS MODEL
In real life we can have an extensive focus which contains a lot of objects, but we really concentrate our attention only on some objects which appeal to us, that is, we receive inputs from objects (we perceive them) but we select (and we are aware of) only those that we are interested in. This selection can be a consequence of the object's appearance, if we are interested in one concrete kind of objects or in one concrete external characteristic, for example, if we are looking for yellow objects. A user’s internal model also will permit us to select an object taking into account our personality, our mood, our attitudes, our intentions [12]. This selection can also be influenced by Social Trends, it is possible that our interest in an object is a consequence of social conventions or fashions [14]. In this way, the inputs that come to our perception based on the awareness model should be subsequently filtered, as it is shown in figure 4. INTEREST ON APPEARANCE U SER DIRECTIONS
focus
INTERNAL MODEL
1.0
SOCIAL TRENDS fm
IN TER NA L F I LTER distance XMDD
X VA
Inputs
Selected Inputs
X NV
Fig.3- Focus function relative to distance. In the Fig.4 we propose a new focus function. For time being, we know that the focus function and nimbus function are different and both depend on object's size and the distance. We also know that
Fig.4- Internal Filters the the the the
AN ASYNCHRONOUS INTERPRETATION OF AWARENESS
If the set of concepts and ideas which have been used to
interpretation in the context of an asynchronous collaboration, we will be able to build a formal model of awareness for this kind of Web applications. Awareness: This concept will quantify the degree, nature or quality of asynchronous interaction between a user and the WEB-based environment. Focus: It can be interpreted as the subset of the web space in which the user has focused his attention. It can relate both to content and to other users. Regarding content, it can be computed by collecting information about the set of places that the user has been visiting while he was navigating through the Web and the set of resources that have been used. Regarding other users, it can be computed by collecting information about areas of common interest and past effective interactions. Nimbus: It is the user’s projection over the WWW space. It can be defined as the set of owned resources that the user is interested in sharing with others and the kind of other users that could or should be informed about the user's activities. Aura: As it happened in CVE’s, this concept will be used to determine the potential for user’s interactions. Boundaries: They are used to divide the web space into different areas and regions and provide mechanisms for marking territory, controlling movement and for influencing the interaction properties of the web space. Sense Acuity: This concept will be used to limit the depth in the search for interesting contents or users and the kind of information that the user can receive from the web site. The maximum number of links to be traversed and the format of this information can be established The concept of Visual Acuity, which has been used in CVE’s, can be interpreted as the degree of restrictions on the visual information that the user can receive from the web. A maximum acuity value will authorise the user to obtain all kind of visual information (images and videos) from the web, while a minimum value will forbid him to acquire visual information. In the same way, the Sound Acuity can be interpreted as the level of permission to receive sound effects from the information that is exposed in the web site. Just as it happens in UNIX with its files and directories, it could be interesting to define a set of permissions to control the reception of information from the web:
T (General Acuity): Permit just access to text information. V xxx (Visual Acuity): Permit xxx types and amount of visual information S xxx (Sound Acuity): Permit xxx types and amount of sound effects Internal Filters: Focus and nimbus could be restricted by the user's internal state and desires. For instance, focus could be restricted through potential collaborator's profiles and through content filters. We will only be aware of those users that are within our focus and fall under our defined profiles. The history of previous interactions and their effects on our mood or internal state can also restrict our focus or nimbus. Thus, a successful interaction will increase our level of attention towards users or contents that fall under a similar profile. CONCLUSIONS
In this paper we have shown how to build a formal model of awareness for 3D Web-based Collaborative Environments, emphasising in how to give an abstract interpretation of a set of concepts and ideas, which have been commonly used to define awareness in CVE’s, in the context of an asynchronous collaboration. Moreover, we place emphasis on those problems that "The Spatial Model of Interaction" does not solve, like Visual Acuity, and we also consider the possibility to modify our awareness as a consequence of our internal state. Our asynchronous interpretation of the awareness concepts is currently being implemented in a prototype system to be used for training purposes. REFERENCES
1. Bailey, I.L. and Lovie, J.E., New Design Principles for Visual Acuity Letter Charts Am J Optom Physiol Opt 53, 740-745 2. Benford, S. Prinz, W. Mariani, J. Rodden, T. Navarro, L. Bignoli, E. Grant Brown C. and Naslund, T. MOCCA - A Distributed Environment For Collaboration, Available from the MOCCA Working Group of Co-Tech. 3. Benford, S., and Fahlén, L.E. A spatial model of interaction in large virtual environments, in Proc. Third European Conference on Computer Supported Cooperative Work (ECSCW'93), Milano, Italy. Kluwer Academic Publishers, pp. 109-124. 4. Benford, S., Fahlén, L. E. and Bowers, J. M., Managing Mutual Awareness in Collaborative Virtual Environments, in Singh, G.,Feiner, S. K., Thalmann, D. (eds.) Proc. ACM SIGCHI Symposium on Virtual Reality Software and Technology (VRST'94), Singapore, pp.223-236, August 1994, World Scientific: Singapore 5. Bowers, J., Modelling Awareness and Interaction in Virtual Spaces, Supplement to Proceedings of the 5th MultiG Workshop, Stockholm-Kista, May 17, 1993, pp. S9-S24.
6. Fahlén, L. E. and Brown, C.G., The Use of a 3D Aura Metaphor for CompterBased Conferencing and Teleworking, In Proc. 4th Multi-G Workshop, Stockholm-Kista, May 1992, pp 69-74. 7. Greenhalgh, C., An experimental implementation of the spatial model, in Pehrson, B., Skarback, E. (eds.) Proc. 6th ERCIM workshops, Stockholm, pp.53-71, June 1994. 8. Greenhalgh, C. M., and Benford, S. D. MASSIVE: A Virtual Reality System for Tele-conferencing, ACM Transactions on Computer Human Interfaces (TOCHI), 2 (3), pp. 239-261, ISSN 1073-0516, ACM Press, September 1995 9. Greenhalgh, C. Dynamic, embodied multicast groups in MASSIVE-2, Technical Report NOTTCS-TR-96-8, Department of Computer Science, University of Nottingham, UK, 1996. 10.Greenhalgh, C., Large Scale Collaborative Virtual Environments, Doctoral Thesis. University of Nottingham. October 1997.
11.Greenhalgh, C., Benford, S. and Reynard, G., A QoS Architecture for Collaborative Virtual Environments, ACM Multimedia (MM'99), Orlando, Florida, November, 1999, ACM Press 12 .Imbert, R., Sánchez, M.I., de Antonio, A., Segovia, J. The Amusement Internal Modelling for Believable Behaviour of Avatars in an Intelligent Virtual Environment. ECAI-98. 13th European Conference on Artificial Intelligence. Proceedings of the Workshop in Intelligent Virtual Environments. Brighton. United Kingdom. August 1998. 13.Rodden, T., Populating the Application: A Model of Awareness for Cooperative Applications, in Proc. ACM 1996 Conference on Computer Supported Cooperative Work (CSCW'96), November 16-20, 1996, Boston, Massachusetts, USA, ACM Press, pp. 87-96. 14.Sánchez, M.I., Imbert, R., de Antonio, A., Segovia, J. Modelling and Evolution of Social Trends in Virtual Environments. SAB98. Workshop of Socially Situated Intelligence. University of Zurich, Technical Report, pp. 80-88. CPM Report No.: pp. 98-45. August 1998. 15. Sandor, O., C. Bogdan and J. Bowers, Aether: An Awareness Engine for CSCW, in Proc. of ECSCW'97: Fifth European Conference on Computer Supported Cooperative Work, H. Hughes, W. Prinz, et al. eds., Lancaster, UK, Kluwer, 1997, pp. 221-236
