Top Banner
Presented at 1 st International Conference on Engagability & Design, Birmingham Institute of Art and Design, Birmingham, July 2004 Mobile Group Music Improvisation N. Bryan-Kinns IMC Research Group, Dept. of Computer Science, Queen Mary, University of London [email protected] Abstract Collaborative improvisation, composition, and performance of music constitute a basic and distinctive form of human interaction. Across cultures, the production and enjoyment of music is typically an open, collaborative, activity more analogous to informal conversation or story telling than to a formal lecture. We know very little about how to design for such situations even though we have seen a blossoming in the range of mobile interactive devices. In this paper we present a design for a remote group music improvisation environment and a evaluation framework based on psychoanalytic theories of attunement between mother and baby in early development. The applicability of such evaluation to joint creative play in adults is demonstrated through observation of the use and development of the remote group music improvisation environment using mobile tablets Key words: play, collaboration, music, design, evaluation, psychoanalysis, user interface Introduction We take group music improvisation as a vehicle in which to explore group creativity as it constitutes a basic and distinctive form of human activity. This contrasts current approaches to understanding collaborative work which have focussed on task, role, and knowledge oriented issues. In turn these have typically lead to the development of systems concentrating on the logistics of collaboration – how to share data, how to negotiate roles, and how to make knowledge available in different contexts. We need to move beyond this to consider the glue that makes collaboration engaging, enjoyable, and rewarding for the participants. We argue that the mutual engagement of participants is fundamental to the efficient and satisfactory completion of group activities and moreover becomes increasingly important for more creative and innovative the activities where participants understand, appreciate, and in some cases anticipate each other’s actions and requirements. It is proposed that from notions of individual’s engagement with interactive experiences (Douglas and Hargadon, 2000) mutual engagement can be characterised as a point at which participants feel engaged with the production in hand (i.e. able to change and appreciate changes in the form) and engaged with other participants (i.e. able to understand their role in the collaboration, the possible next steps of others, and others’ perceptions of their possible next steps). This can be viewed a state in which participants experience flow (Csikszentmihalyi, 1991) as a group (Sawyer, 2003) – highly focussed activity pushing at the boundaries of a shared understandings of experience and expectation. It is argued that whilst collaboration can be undertaken without high levels of mutual engagement, it is vital to enjoyable and high quality collaborative activities. Mutual engagement itself is a state of being within a social construction, and as such we cannot directly observe or induce mutual engagement in people. Instead we can look for indications that mutual engagement has occurred in traces of interaction, and we can try to develop tools which are designed to remove barriers to mutual engagement. For example, we might try to reduce the need to become
8

Mobile Group Music Improvisation - EECSnickbk/papers/nickbk_eandd_full.pdf · Mobile Group Music Improvisation N. Bryan-Kinns IMC Research Group, Dept. of Computer Science, Queen

Mar 18, 2020

Download

Documents

dariahiddleston
Welcome message from author
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Page 1: Mobile Group Music Improvisation - EECSnickbk/papers/nickbk_eandd_full.pdf · Mobile Group Music Improvisation N. Bryan-Kinns IMC Research Group, Dept. of Computer Science, Queen

Presented at 1st International Conference on Engagability & Design,Birmingham Institute of Art and Design, Birmingham, July 2004

Mobile Group Music ImprovisationN. Bryan-Kinns

IMC Research Group,Dept. of Computer Science,

Queen Mary,University of London

[email protected]

AbstractCollaborative improvisation, composition, andperformance of music constitute a basic anddistinctive form of human interaction. Acrosscultures, the production and enjoyment of musicis typically an open, collaborative, activity moreanalogous to informal conversation or storytelling than to a formal lecture. We know verylittle about how to design for such situationseven though we have seen a blossoming in therange of mobile interactive devices. In this paperwe present a design for a remote group musicimprovisation environment and a evaluationframework based on psychoanalytic theories ofattunement between mother and baby in earlydevelopment. The applicability of suchevaluation to joint creative play in adults isdemonstrated through observation of the use anddevelopment of the remote group musicimprovisation environment using mobile tablets

Key words: play, collaboration, music, design,evaluation, psychoanalysis, user interface

IntroductionWe take group music improvisation as a vehiclein which to explore group creativity as itconstitutes a basic and distinctive form of humanactivity. This contrasts current approaches tounderstanding collaborative work which havefocussed on task, role, and knowledge orientedissues. In turn these have typically lead to thedevelopment of systems concentrating on thelogistics of collaboration – how to share data,how to negotiate roles, and how to makeknowledge available in different contexts. Weneed to move beyond this to consider the gluethat makes collaboration engaging, enjoyable,

and rewarding for the participants. We arguethat the mutual engagement of participants isfundamental to the efficient and satisfactorycompletion of group activities and moreoverbecomes increasingly important for morecreative and innovative the activities whereparticipants understand, appreciate, and in somecases anticipate each other’s actions andrequirements.

It is proposed that from notions of individual’sengagement with interactive experiences(Douglas and Hargadon, 2000) mutualengagement can be characterised as a point atwhich participants feel engaged with theproduction in hand (i.e. able to change andappreciate changes in the form) and engagedwith other participants (i.e. able to understandtheir role in the collaboration, the possible nextsteps of others, and others’ perceptions of theirpossible next steps). This can be viewed a statein which participants experience flow(Csikszentmihalyi, 1991) as a group (Sawyer,2003) – highly focussed activity pushing at theboundaries of a shared understandings ofexperience and expectation. It is argued thatwhilst collaboration can be undertaken withouthigh levels of mutual engagement, it is vital toenjoyable and high quality collaborativeactivities.

Mutual engagement itself is a state of beingwithin a social construction, and as such wecannot directly observe or induce mutualengagement in people. Instead we can look forindications that mutual engagement has occurredin traces of interaction, and we can try todevelop tools which are designed to removebarriers to mutual engagement. For example, wemight try to reduce the need to become

Page 2: Mobile Group Music Improvisation - EECSnickbk/papers/nickbk_eandd_full.pdf · Mobile Group Music Improvisation N. Bryan-Kinns IMC Research Group, Dept. of Computer Science, Queen

Presented at 1st International Conference on Engagability & Design,Birmingham Institute of Art and Design, Birmingham, July 2004

immersed in the logistics of collaboration byexplicitly supporting cues to mutual knowledge(cf. Healey and Bryan-Kinns, 2000).

As we are considering collaboration in someway supported by tools, the features of acollaborative situation which relate to mutualengagement are twofold – the tools themselves(production mechanisms, co-ordinationmechanisms), and social constructs – these arebriefly outlined in the following sections.

Production mechanismsThe aim of creativity is to produce somethingnovel, so we need to design interfaces whichsupport and enhance play rather than hinderingit. We have developed a set of design criteria foridea exploration and formulation (Bryan-Kinnset al., 2004) based on Tabor’s notion of a ‘spacefor half formed thoughts’ (2002) which consistsdesigning for a spatial metaphor, multimodality,emergent patterns of interaction, and variablefocus allowing shifting of attention.

Co-ordination mechanismsOf course, we are interacting with others incollaborative situations so we need to reduce theemphasis on the logistics of interaction –knowing who is doing what, when, and where.To this end we developed a set of design criteria(Bryan-Kinns et al., 2003a; 2003b) for efficientcollaboration from features of humaninteraction: Localization within the artifact beingproduced, mutual awareness of actions, mutualmodifiability, and shared and consistentrepresentation.

Social constructsWhat we believe about activities, others, andourselves is socially constructed (Burr, 1995)and frames our action. Moreover, eachinteraction contributes a change to the socialconstructs which will in turn change the way weview the world. Within social constructs theremay be many factors influencing mutualengagement in a collaboration e.g. personalcharacteristics, experience, social setting,cultural conventions; the focus of our research ison the effect of technology on mutual

engagement, and how that effect could beassessed. In this paper we focus on identifyingindicators of mutual engagement - suchassessments could be used to compare systems,or to generate guidelines for systemdevelopment.

Mutual EngagementIndicatorsThere may be many indicators and measures ofmutual engagement which would be applicablein different situations. For example, we mightrate the quality of the end product, or judge thecreative ‘prowess’ of the activity as it unfolded,or ask participants to rate their enjoyment andengagement with others in the process. Theseratings would tend to be both subjective andhighly domain dependent, making them difficultto usefully apply across domains. Instead it isargued that by examining the communicativeinteraction involved in collaboration we canidentify key features of the collaboration whichmight indicate levels of mutual engagementbetween participants.

The approach detailed in this paper is inspiredby psychoanalytic descriptions of attunement(Stern, 1985) between mother and baby in earlychildhood (see Wright, 2000 for a introductionto Stern’s work in the area and its relation tocreativity in adult life). Such a view describesthe interaction between mother and baby wherein a normal situation, the mother responds notonly to their infants’ major emotions, but also ina moment-by-moment way to smaller changes –‘she reads her baby by every possible non-verbalmeans and intuitively senses the changingpatterns of its feeling state’. Moreover, ‘in thenormal situation, the mother does not merelyregister the infant’s state in an ongoing way: sheengages in responsive displays of her own,which, in one modality or another, reflect thepatterns and rhythms of the baby’s’ (Wright,2000). We draw inspiration from the approachand focus on three levels of attunement:

Page 3: Mobile Group Music Improvisation - EECSnickbk/papers/nickbk_eandd_full.pdf · Mobile Group Music Improvisation N. Bryan-Kinns IMC Research Group, Dept. of Computer Science, Queen

Presented at 1st International Conference on Engagability & Design,Birmingham Institute of Art and Design, Birmingham, July 2004

1) Acknowledgement – participants show thatthey are aware of the contribution of another

2) Mirroring – participants mirror, or reflect,others’ contributions thus demonstrating thatthey themselves are able to produce it

3) Transformation – participants transformothers’ contributions, indicating a high level ofmutual engagement

This framework has strong parallels to thealignment of words and gestures in everydayconversation which indicates a level ofinteraction and engagement with each otherwhere joint meaning and understanding is beingcreated (Tabensky, 2001). However, byfocussing on attunement of action, we do notlimit ourselves to verbal or gestural interaction.

Remote Group MusicImprovisationAs mentioned previously, we take group musicimprovisation as a paradigm example of groupcreativity and play. Titon (1996) providesethnomusicological discussion of music andimprovisation from rain forest cultures such asthe BaAka people to classical Indian musicwhere performances are a balance betweenprecomposed and improvised music. State of theart music technology for individual and groupmusic production focuses on two issues: noveltools for improvisation, and support forcomposing and editing music. New forms ofinstruments aimed at individual improvisationare exemplified by systems such as the ElectricCircus (Coady, 2002) which is played byjumping on a floor mounted ‘keyboard’. Newmeans of individual composition are exemplifiedby the mixed reality techniques employed by the‘augmented composer’ (Berry et al., 2002)where computer recognisable cards are arrangedon a table to create musical phrases. From agroup music perspective composition issupported by commercial systems such asRocket networks (cf. Hall, 2002) where supportfor group composition is limited to file sharingacross networks, not the process of collaborationper se, and research into the effects ofrepresentations on temporal location in musicduring group composition (Nabavian, 2002).

Work on group improvisation technologies isless advanced. Research such as FMOL (Jorda,2001), Webdrum (Burk, 2000), and Metatone(Leach, 2001) have begun to explore the groupimprovisation in geographically remotelocations. This typically involves developing ashared visualisation of the music being producedand some communication support, but they arebased on a very limited analysis of the characterof mutual engagement in musical collaboration.Without this perspective we are left to designdevices by intuition – hoping that we will createnew forms of interaction that somehow are moremusically and socially engaging.

Design of a novel group musicenvironmentDaisyphone is an on going design project(Bryan-Kinns et al. 2003; 2004) whose aim is tosupport remote group music improvisation. In itscurrent form, up to 10 remote participants cancreate and edit a short shared loop of musicsemi-synchronously – typically updates takeunder one second to be shared. This providessupport for a form of remote group musicimprovisation whilst requiring little networkbandwidth which is important when designingfor mobile devices which may have limitedprocessor power and network bandwidth –approaches such as BT Music On-Line (BT,2004) which rely on high speed networks,processors, and video, would not be appropriatefor such ad-hoc, low cost, and informalcollaboration. As with other remote groupimprovisation tools such as WebDrum (Burk,2000), Daisyphone works by clients sharingindications of musical contributions via a centralserver through the internet so providing a sharedand consistent representation of musical loopsbeing constructed.

We have used our design criteria for productionand co-ordination mechanisms outlined earlier toinform the design of Daisyphone. There is noownership in Daisyphone – people can edit eachothers’ notes and play the same instrumentswhich we hope will start to breakdownconventional playing patterns such a audiencevs. performer. As well as sharing musicalcontributions, Daisyphone also shares graphical

Page 4: Mobile Group Music Improvisation - EECSnickbk/papers/nickbk_eandd_full.pdf · Mobile Group Music Improvisation N. Bryan-Kinns IMC Research Group, Dept. of Computer Science, Queen

Presented at 1st International Conference on Engagability & Design,Birmingham Institute of Art and Design, Birmingham, July 2004

annotations on and around the musiccomposition space. This annotation is intendedto support both localization within thecomposition, and social and discursiveexchanges. Moreover, it provides a ‘messy’space in which to create and play which we hopeengenders exploration and fun interaction (cf.Fass et al., 2002). Finally, in order to be usablein a variety of form factors and devices, thedesign is scalable and does not rely onconventional user interface widgets such asmenus or pointers, and instead relies on truedirect manipulation of the representation.

The Daisyphone user interface is illustrated infigure 1. Notes are lower in pitch towards theedge of the circle. As the grey arm rotatesclockwise, the notes underneath are played, soeach of the spokes represents notes played at thesame time. Hues of notes indicate whocontributed them (each participant is allocated aunique hue), and intensity of colour representsthe volume of the note. Volume and instrumentare modally controlled from the four centralspokes.

Figure 1: Daisyphone interface

Previous studies (Bryan-Kinns et al., 2003;2004) have identified several design issues withDaisyphone and its support for group creativity.In this paper we present analysis of logs of theuse of Daisyphone to illustrate our analyticframework for identifying points of mutualengagement in collaborations.

Study backgroundLogs of the use of Daisyphone were gatheredfrom a study involving 10 school pupilsinteracting with Daisyphone in a semi-remotesituation using wirelessly connected mobilegraphical tablets. The participants were given abrief introduction to how the Daisyphoneworked covering: how to set notes, how to unsetnotes, how to select different instruments, andthe shared nature of the representation. Theywere then asked to work with remote co-participant(s) to try to create a recognisable tunee.g. a TV theme tune, or a tune they liked.

Participants had up to 15 minutes to completethe task. During the session participants couldask for a new Daisyphone session giving them ablank canvas to work with. This was negotiatedbetween the participants of the group. Theycould return to previous work if they wished;none of the groups revisited old work.

The aim of the physical setup was to providesemi-remote collaboration where non-Daisyphone interaction between remoteparticipants could easily be identified andrecorded. Two desks were set 10m apart toreduce audio contact between the tables asillustrated in figure 2. Visual contact could beachieved by turning away from the table, andvocal communication achieved throughshouting. Each desk had on it: a tablet runningDaisyphone, an instruction sheet, and a pair ofspeakers connected to the tablet.

Figure 2: Study setup

Page 5: Mobile Group Music Improvisation - EECSnickbk/papers/nickbk_eandd_full.pdf · Mobile Group Music Improvisation N. Bryan-Kinns IMC Research Group, Dept. of Computer Science, Queen

Presented at 1st International Conference on Engagability & Design,Birmingham Institute of Art and Design, Birmingham, July 2004

From logs of the interaction contributions werecategorised in two primary ways: inferredcontribution to graphical form, and inferredmotivation for contribution which aredecomposed below and discussed in thefollowing sections.

Graphical form was categorised in four ways asillustrated in figure 3: filling of the space (F),drawing straight lines (L), drawing curved lines(C), and individual dots (D) not forming part ofa discernable and contiguous geometric shape.

Figure 3a: Filling b: Straight lines

Figure 3c: Curves d: Individual dots

Inferred motivation was categorised in six ways:apparently random (R), experimental (E), andintentional contributions which are necessarilyrelated to other contributions in some way asthey make a sequence or graphically align.Intentional contributions may either be fitting totheir own contributions (S), or others’contributions. Fitting to others’ contributionsindicates a form of attunement – transformation(T), mirroring (M), or acknowledgement (A) ofothers’ contributions as outlined below:

Acknowledgement – participants madecontributions which took into account otherpeople’s already present contributions by notwriting over them, or by creating chords usingpart of someone else’s contribution. This showsthat they were aware of others, and in the case ofchord creation, had high levels of musicalability. Figure 4 illustrates a situation in which

the two participants implicitly acknowledge eachother’s existence by working in their own part ofthe Daisyphone space – playing in their ownspace and not encroaching on the other.

Figure 4: Acknowledgement

Mirroring –participants may start to mirrorpatterns or tunes created by others. For example,two users (Lindon and Curtis) in the same spaceinitially started making different geometricshapes with Lindon sticking to lines and Curtissticking to curves as illustrated in figure 5a.Later on in their third of session Lindon thenstarted to move on to making curves – mirroringCurtis’ work (illustrated in figure 5b). In thestudy mirroring typically focussed on thegeometric patterns (lines and curves) rather thanon mirroring musical melodies which indicates alevel of mutual engagement, but a low level ofengagement with the music itself – they areplaying with each other through visualinteraction, not musical interaction as we hadhoped.

Figure 5a: Pre-Mirroring b: Mirroring

Transformation – there were very fewinstances (discussed later) of participants takingother people’s contributions, playing with them,and creating their own versions of them. This is

Page 6: Mobile Group Music Improvisation - EECSnickbk/papers/nickbk_eandd_full.pdf · Mobile Group Music Improvisation N. Bryan-Kinns IMC Research Group, Dept. of Computer Science, Queen

Presented at 1st International Conference on Engagability & Design,Birmingham Institute of Art and Design, Birmingham, July 2004

a indication of higher levels of mutualengagement than mirroring as it indicates awillingness not only to mimic others, but also toplay with other people’s ideas. For example,Lindon and Curtis went on to a situationillustrated in figure 6 in which Lindon wasmaking quite complex variations of curves – aplay on the original work by Curtis throughtransformation.

Figure 6: Transformation

Time and transitionTo indicate what an activity looked like, andwhat we believe the intentions behind it were wepaired the graphical and motivational codes. Forexample, drawing a straight line experimentallywould be coded as LE.

We are interested in the points at whichintentional activities showed some level ofattunement with others as we argue that this isan indicator of mutual engagement. In thesesituations we suggest that participants are tyingto construct musical pieces (rather thanexploring individual sounds) and so try to linknotes together to make a tune. This isdistinguished from experimentation as itinvolves some editing of the notes – un-settingas well as setting.

Being able to identify instances of attunement isa useful activity in itself for spotting points atwhich mutual engagement may occur. However,it does not help us to compare designs or tounderstand the overall nature of thecollaboration. To this end we also analyse the

amount of time indicators of mutual engagementoccur for, and the transitions between differentindicators. We found that that most time wasspent in DE (Dot Experimenting – 926s) whichmay have been due to the novelty of the userinterface meaning that many people wereengaged in learning to play music on it –essentially playing with the interface rather thanplaying with each other through the interface aswe had hoped. We then found that the mostfrequent transition (9 times) from DE was tochat (writing on the graphic area), and lessfrequently (8 times) a transition to LE (299s) asthe participants learnt how to create lines, notjust dots. The most striking aspect of theanalysis was the lack of transformations ormirroring. These were never reached from anyother state more than once, and the total timespent was relatively small: LM 9s, CM 17s, LT26s, CT 30s, DT 35s.

DiscussionSo, what does our analysis tell us about thedesign and use of Daisyphone and groupcreativity systems in general? First it indicatesthat the level of expertise was low withparticipants tending to move betweenexperimenting and chatting. This was to beexpected given the novel nature of the interface,a potential development could be to try todevelop a more intuitive, or easier to playinterface, for example, by only allowingproviding notes from certain musical scale sothat they were always in tune with each other.Second, it illustrates the typical development ofskill from experimenting with dots (DE) toexperimenting with lines (LE) and on toexperimenting with curves (CE). This probablyreflects development of musical understandingthrough the use of graphical patterns. A usefuldevelopment might be to exploit people’sfamiliarity with graphical shapes in order todevelop musical skills e.g. by using free formmusic to create gesture rather than a circularinterface. Third, it illustrates a learning pathfrom experimenting with lines (LE) to makinglines with reference to one’s own compositions(LS) and on to skilled composition with respectto one’s own contributions (DS). We couldbetter exploit this path by providing more

Page 7: Mobile Group Music Improvisation - EECSnickbk/papers/nickbk_eandd_full.pdf · Mobile Group Music Improvisation N. Bryan-Kinns IMC Research Group, Dept. of Computer Science, Queen

Presented at 1st International Conference on Engagability & Design,Birmingham Institute of Art and Design, Birmingham, July 2004

coherent building blocks such as rhythmicpatterns which could be joined together to makelonger pieces and experimented with. Indeed,the feedback we received from the pupils wasthat only 50% felt that they could develop goodtunes with Daisyphone, so there is clearly stillsome development to be done on the productionmechanism, though it should be noted that it wasa very unusual interface for the pupils and theyonly had a short time to interact with it. Finally,it illustrates that acknowledgement was thehighest level of attunement reached which hadmore than one transition to any other state – thisis not what we had expected when we designedthe system and is discussed in more detailbelow.

Considering the two design criteria ofproduction and co-ordination, it seems to us thatwhilst Daisyphone’s production mechanisms aresomewhat effective, the co-ordinationmechanisms appear to not be providing enoughsupport for engagement between participants.Figure 7 illustrates this situation – we seeindicators of mutual engagement (attunement inthis case) as a function of the production and co-ordination mechanisms as illustrated by thebands of italic text. As production and co-ordination mechanisms improve fromconcentrating on logistics of the interaction tofocussing more on engaging interaction we positthat there will be more indicators of higherlevels of attunement indicating more mutualengagement is happening. From ourobservations we believe that there is typically acombinatorial effect of production and co-ordination mechanisms, hence the curved shapesof the attunement indicators. For example, toolswith very engaging production mechanisms butpoor co-ordination mechanisms may be betterfor mutual engagement (e.g. reaching theacknowledge band) than a tool with bothmediocre production and co-ordinationmechanisms.

Our current design of Daisyphone is illustratedby the ‘Current’ dot within the Acknowledgeband of attunement – we need to improve boththe production and co-ordination mechanisms toget to our ‘Aim’ in the Transform band ofattunement. In its current state, we believe thatdevelopment of Daisyphone’s co-ordination

mechanisms would make the greatest difference.For instance, making other people’scontributions glow more prominently, orproviding more feedback about who is in thespace at the same time, what they are doing, andtheir focus of attention.

Figure 7: Indicators of Attunement inRelation to Production and Co-ordination

Mechanisms

SummaryWe are still a long way from understanding whatmakes that creative ‘spark’ happen in groupcreative situations. In this paper we haveproposed a framework for analysing theinteractions within a group in order to identifypoints of mutual engagement. We used this toanalyse group play with a novel group musicinstrument and to form proposals for redesign inorder to hopefully make the experience moremutually engaging. We believe that suchframeworks can provide insights into the designand evaluation of a wide range of playfulinterfaces as they are based on fundamentalnotions of interaction rather than modal specificanalyses.

AcknowledgementsSupported by EPSRC grant GR/S81414/01 andQueen Mary, University of London.

Page 8: Mobile Group Music Improvisation - EECSnickbk/papers/nickbk_eandd_full.pdf · Mobile Group Music Improvisation N. Bryan-Kinns IMC Research Group, Dept. of Computer Science, Queen

Presented at 1st International Conference on Engagability & Design,Birmingham Institute of Art and Design, Birmingham, July 2004

ReferencesBerry, R., & Tadenuma M. (2002). “AugmentedReality for Music.” Proceedings of the 2002International Computer Music Conference, 110-113.

Bryan-Kinns, N., Healey, P., Thirlwell, M., &Leach, J. (2003a). “Designing for GroupCreativity.” Supplementary Proceedings of HCIInternational 2003.

Bryan-Kinns, N., Healey, P., & Thirlwell, M.,(2003b). “Gudar – A Novel Group MusicInstrument.” Supplementary Proceedings ofINTERACT 2003.

Bryan-Kinns, N., Healey, P., & Thirlwell, M.,(2003c). “Graphical Representation for MusicalImprovisation.” Proceedings of 2ndInternational Workshop on InteractiveGraphical Communication 2003.

Bryan-Kinns, N., & Healey, P. G. T. (2004).“Daisyphone: The Design and Impact of a NovelEnvironment for Remote Group MusicImprovisation.” Proceedings of DIS 2004,Boston, USA.

BT. (2004). BT Music On-Line.http://music.york.ac.uk/musiconline/ (last visited19 May 2004)

Burk, P. (2000). “Jammin' on the Web - a newClient/Server Architecture for Multi-UserMusical Performance.” Presented at ICMC2000.

Burr, V. (1995). An Introduction to SocialConstructivism. Routledge.

Coady, N. (2002). Interaction design for real-time music creation environments – The ElectricCircus. MA thesis, University of Westminster.

Csikszentmihalyi, M. (1991). Flow: ThePsychology of Optimal Experience. HarperCollins.

Douglas, Y., & Hargadon, A. (2000). “ThePleasure Principle: Immersion, Engagement,Flow.” Proceedings of Hypertext 2000, SanAntonio, USA, pp. 153-160.

Fass, A., Forlizzi, J., and Pausch, R. (2002).“MessyDesk and MessyBoard: Two DesignsInspired by the Goal of Improving HumanMemory.” Proceedings of DIS 2002, London,UK, 303-311.

Hall, G. (2002). “Bands Without Borders.”Electronic Musician Magazine, October 2002,72-86.

Healey, P., & Bryan-Kinns, N. (2000).“Analysing Asynchronous Collaboration.”Proceedings of HCI 2000: People andComputers XIV – Usability or Else!, 239-254.Springer-Verlag, London.

Jordà, S. (2001). “Improvising with Computers:A Personal Survey (1989-2001).” Proceedingsof 2001 International Computer MusicConference. La Habana.

Leach, J. (2001). MetaTone: SharedEnvironment for musical collaboration. MSc ITThesis, Queen Mary, University of London.

Nabavian, S. (2002). Supporting the Emergenceof Composition from Improvisation. AMScThesis, Department of Computer Science, QueenMary, University of London.

Sawyer, R. K. (2003). Group Creativity: Music,Theater, Collaboration. Lawrence Erlbaum Ass.

Stern, D. N. (1985). The Interpersonal World ofthe Infant: A View from Psychoanalysis andDevelopmental Psychology. Karnac Books,London.

Tabor, P. (2002). “A Space for Half-FormedThoughts.” Speech given at Doors of Perception7 Conference, Amsterdam, 2002.

Tabensky, A. (2001). “Gesture and SpeechRephrasings in Conversation.” Gesture. 1:2,213-235.

Titon, J. T. (1996). World of music. SchirmerBooks, NY.

Wright, K. (2000). “To Make Experience Sing.”In Caldwell, L. (ed.) Art, Creativity, Living,Chapter 5, pp. 75-96.