Weakly augmented reality

Weakly Augmented Reality: observing and designing thework-place of creative designers

Giorgio De Michelis*^, Flavio De Paoli*, Costanza Pluchinotta^, Marco Susani°1

1 *DISCO, University of Milano – Bicocca, Via Bicocca degli Arcimboldi 8, 20126 Milano, Italy [[email protected]]

^IRSO, Piazza Giovine Italia 3, 20123 Milano, Italy [demichelis, [email protected]]

°DARC, Domus Academy, Via Savona 97, 20144 Milano, Italy [[email protected]]

ABSTRACTIn this paper we distinguish between two spatiallyoriented system design paradigms: weak and strongaugmented reality. The weak augmented realityparadigm is then applied in the design of a systemsupporting co-operation and knowledge creation withina design centre. The system has been designed on thehints we got from two subsequent ethnographies of thework practice of creative designers. Since the designcentre moved to a new location between our two sets ofobservations, we clearly focused our attention on theimpact that space arrangements had on the practice ofits members. The comparison between the two settingsin terms of layout, ICT equipment, changes occurred inthe practices and attitudes of the designers throughoutthese years, lead us to design a collaborativeenvironment weakly augmenting the place wheredesigners are working.

KeywordsAugmented reality, knowledge management systems,creative design, spatial arrangements.

INTRODUCTIONThe space metaphor not only provides a means foranalysing and designing the practices of human beingsbut also creates the language through which they speakabout their lives. Proximity and distance, being or notbeing there, drawing near and moving away are factorswhich allow us to characterise the relations bindingdifferent people while they are involved in a commonpractice, for they are used by those people whileoperating together. In human life a space «is investedwith understandings of behavioural appropriateness,cultural expectations, and so forth» [10], characterisingthe practice of the people living in it, transforming itselfinto a place. In short, the space metaphor originatesfrom the basic fact that we live in a space; but it is keptalive by our capability to use it as the framework forgiving sense to that space on the basis of our practice.

The growing presence of Information andCommunication Technology (ICT) in any type of placehas changed the way people understand and operatewithin them [7] and therefore the way people live andexperience the spatial metaphor. Adopting Harrison andDourish’ conceptual framework, we can say thatinformation and communication technology transformsthe places where we spend our lives, extending thembeyond the limits of the portions of physical space,which host them. Let us recall some examples that offerhints of the directions that this concept may follow.

When people initiate an email message writing «Hi… ,»instead of «Dear…,» they say, within an asynchronouscommunication medium, what they would say whenmeeting face to face, as if the velocity of email couldcreate a ‘virtual’ common space where the sender isright there with the recipient while communicating.

Dourish et al. [8], reporting on experiments carried onat Xerox Research Labs., observe that if the physicalspace is reconfigured taking into account the presenceof an audio and video channel, then it happens thatpeople change their behaviour taking into account thetransformation of the place. First of all , visitors enteringan office begin their interaction greeting not only itsinhabitant but also her remote partner, who is «present»through the audio and video link. Moreover, since theoffice has been rearranged in such a way that the remotepartner could see its screen, she too is behaving as ifshe and the office inhabitant lived in the same place.

The Olivetti Research Laboratory at CambridgeUniversity in UK (currently continuing its activity asATT Research Laboratory) has created a good exampleof an environment that is enriched by ICT in theComputing Department of the same University [21].Both members and visitors are equipped with activebadges that send signals to sensors distributed inoffices, meeting rooms and corridors activating a largevariety of services for its users. Some examples shouldbe sufficient: doors open in front of an authorisedincoming person; the computer screen is obscured whenits owner is out of her room; lights are automaticallyswitched off when the room is empty; personal phonenumbers make the nearest telephone ringing when theperson is called; a list containing the places where

81

people are located is updated continuously, saying e.g.registering that the person who left the Lab were lastseen in front of the elevator.

These can be considered as examples of different waysof how ICT changes the space perception andunderstanding of human beings. While the emailexample shows how a communication medium inducesthe users to behave (to speak) as if they were in acommon virtual place, thus enriching the physicallocations where they are situated, the second one showshow video and audio links can draw two differentphysical locations so close to each other, that theyappear as (parts of) a unique place. The lastly thirdexample shows how enriching a physical location withICT can transform it into a place whose features changein order to react to the behaviour of the human beingsoperating in it.

All these examples can be considered as examples ofaugmented reality [15, 22, 23], since they are notaiming to create a virtual counterpart of the physicalspace [2, 17] where people act and interact, but enrichthe physical locations where they live so that their placegoes far beyond the limits of the physical reality [16].While Wendy Mackay describes three strategies foraugmenting reality focusing, respectively, on users,physical objects and the environment surrounding bothusers and objects [15], we adopt an orthogonalviewpoint, characteris ing within it two paradigms: letus call them strong and weak augmented reality.

The strong augmented reality paradigm aims to changethe space perception of its inhabitants so that theycannot distinguish the physical base of their place fromits virtual extension. When a person working with adangerous material is allowed to handle it withoutdirectly touching it, the quality of the system dependson its capability to allow the user to perceive the remotedangerous material as if it were directly in her hands.

The weak augmented reality paradigm aims to inducepeople to behave as if their place were transformedeven if the physical space where they are located doesnot seem changed. The email system does not changethe physical space in any sense; audio and video linksthat you can hear and see through a window on aworkstation do not significantly change the space wherethe workstation is situated; even the ubiquitouscomputing extensions of the Olivetti ResearchLaboratory at Cambridge University do not change thespace where they are located, they only enrich theirbehavioural properties.

We must underline at this point that there is no clear-cutdistinction between weak and strong augmented reality:strong augmented reality applications frequentlyenhance their users feeling of being in a different place.As a matter of fact, the three examples we have groupedwithin the weak paradigm can be distinguished from thestrong paradigm viewpoint: from email that does not inany way affect user perception, to the audio and videolinks and active badges and sensors which, on the

contrary, create more realistic representations of theaugmented place.

Weak augmented reality paradigm is important in that itoffers a perspective for evaluating and designing ICTapplications from the spatial metaphor viewpointwhether they are aiming to change the space perceptionof their users or not. While weakly augmented realitycan create new places, even without inventing newdevices or changing existing objects, it offers aviewpoint for evaluating strong augmented realitysystems , allowing to understand their usability and theireffectiveness above and beyond technical feasibility.

In this paper we present an application of the weakaugmented reality paradigm. We were lucky enough tohave had the possibility of observing a design centretwice (in 1996 and 1999), before and after it underwentits major changes: a new location, an increase in thenumber of the employed designers and of the ongoingprojects, an improvement of the technology in use.

Giorgio De Michelis carried on the first observationtogether with Edmundo Leiva Lobos and ElianaCovarrubias [5, 14] within the Desarte project [20],while the second he carried on together with CostanzaPluchinotta as part of the Klee&Co Project [6]. MarcoSusani, director of the design centre under observation,participated in both the observations as an active user.Flavio De Paoli is responsible for the system to bedesigned and developed for the design centre within theKlee&Co project.

Evaluation of the two work settings allows us to designthe system we are developing as a means of augmentingthe design centre workplace so that it combines thegood points of both previous and more recent worksettings, while avoiding the weaknesses of each. Inanalysing and designing the system, therefore, the spacedimension plays a major role, even if we are neithercreating new devices nor deeply transforming existingobjects. This will become clear in the followingsections of our paper.

The next section reports on the findings of theobservations of the two workplaces. Later, we describehow the current work-space of the observed designcentre can be augmented without making use ofsophisticated technologies, but by trying to deeplymodify the behaviour of its inhabitants and visitors.Then, the reader will find a section outlining the systemwe are designing and developing to fulfil therequirements established in the previous section. Theconclusion comes back to the augmented realityparadigms to indicate some future directions for ourresearch.

OBSERVING DOMUS ACADEMY TWICEDomus Academy (DA) was established in 1983 as aprivate Italian Education Institution dealing withinnovation in Industrial Design. Some years later aResearch Center (DARC) was created in order tosupport educational activities and to explore the most

82

innovative design themes from a practical viewpoint.Currently DARC is involved in both consultancyprojects and research projects. We had the occasion tomake ‘ethnographic’ observations of DARC workpractice twice in the last four years, once in 1996,within the EC funded Desarte Project, and later in 1999within the EC funded Klee&Co Project. WithinDesarte, one of the authors conducted a feasibility studytogether with Edmundo Leiva Lobos and ElianaCovarrubias Gatica [14] . The aim was to analyse theneed for a co-operation support system on the part ofcreative people (within the same project other groupsmade similar observations on architects, landscapearchitects and software designers [20]). As a follow upto Desarte, a new project named Klee&Co was set upand funded by the EC with the aim of designing anddeveloping the co-operation support system forindustrial designers outlined in Desarte. There are twouser organisations in Klee&Co: Domus Academy andPhilips Design. In order to update the knowledge on thework practice at DARC and to better define therequirements for the system to be designed, it wasnecessary to repeat the ethnographic observation atDARC. The intention was to focus the system onsupporting its informal learning process mediated by"knowledge of experience".

Our ‘ethnographic’ observation was part of a designprocess [11, 18]. It focused on the spatial dimension ofthe work setting and on its influence on the practice ofthe designers’ community [3, 13, 24] as well as on theway people interact within and across project teams andwith external visitors and customers [5, 14].

The fact that between 1996 and 1999 the situation atDARC experienced major changes (Domus Academymoved to a new larger location; DARC had aextraordinary growth becoming deeply involved ininteraction design projects; ICT became widely diffusedwithin DARC so that each designer had a networkedworkstation) gave us the opportunity of using theknowledge about the transformation of the workspacein analysing and evaluating its current performances.Let us outline our findings.

DA’s organisation of space in 1996In 1996 DA operated on the top floor of a buildingoutside Milan. The space where all educationalactivities were performed was physically adjacent to theDARC open space. The two places were decorated withpictures, sketches and artefacts produced and chosen,respectively, during education working sessions andstudent thesis work on the education space side and 2 We put the word between single quotation marks,

since we do not use a standard ethnographicobservation style, since on the one hand ourobservation sought to contribute to the process ofdesigning a cooperation support system for theobserved organization respecting its strict timing andon the other we involved the observed community ofpractice in an action learning process so that the lattercould fully participate in the design process

during design and/or research projects on the DARCside.

The contiguity of the two spaces enhancedcommunication among students and researchers aboutwhat was going on in both environments. As a matter offact, there was a close exchange between the two parts:students felt free to visit the DARC, asking about thepictures and sketches hanging up all around and aboutthe projects they were related to, similarly, researcherspassing through the educational space to reach theirworkplace were used to poking their nose into theartefacts produced by the students, offering suggestionsand asking for more information.

The DARC open space was divided into sub-areas foreach ongoing project (see figure 1) so that all thedesigners co-operating on a project had a common workplace. This space organisation had two main features.

Figure 1 - The DA Research Centre in 1996: a “perproject” organisation of space.

First, its walls (and cabinet doors) were fully decoratedwith photographs and drawings contributing to inspirethe creative design process of the project (see figure 2).These highly evocative pictures captured the attentionof both occasional visitors – designers, students, etc. –and the DARC designers involved in other projects,thus provoking conversations with members of thedesign teams about what they were doing.

Conversing about the work they were engaged in, theyboth improved their ability to describe and explain theirwork and had the occasion to get new ideas and hearoutside opinions. Conversations triggered, moreover,cross-fertilisation, exchange of viewpoints, andenhanced awareness about what was going on in theResearch Centre.

Secondly, graphic workstations, scanners, office tools,etc. were placed on the hot desk of each project (socalled because that’s where the work is really going on)and in the graphic design section of the DARC, since allproject teams shared the technological support systemsand no designer had a personal workstation. Computersdid not play the major role in the creative process;rather they were used mostly for creating the project

83

presentations, writing documents of any type, searchingin the web, interacting via e-mail. For the rest,designers were used to draw on paper and to buildphysical mock-ups and models of the artefacts theywere designing.

Figure 2 - The DARC space was fully decorated.

As mentioned above, the project workplace was full ofimagery and knowledge supporting co-operation amongteam members. So even those who were away from theDARC for some days could, upon returning, be updatedof what happened during their absence by simply takinga look at the new images and objects in the projectworkplace.

This work setting presented some weaknesses too.Technological equipment was insufficient. This couldlead to a bottleneck when various project teams needed,for example, to prepare a presentation for theircustomers. Thus conflicts over the use of computingresources were frequent. Using workstations ascommunication media and/or as a means of creating acommon knowledge space within a project was alsoquite difficult due to the fact that they were not on thedesk of the designers.

Since project leaders and senior designers werefrequently away for business and research purposes,their absence provoked breakdowns in the projectswhen their contribution was needed. In fact, they bothrepresented the «communication bridge» between theproject team and the client(s) and had the final word inany decision concerning the project. In each case, inorder to solve or avoid the breakdowns, an effective andbroad communication network connecting all DARCmembers, at the workplace or not, should have beenintroduced.Finally, with regards to communication betweenclient(s) and project team, there was another issue to betaken into account. Client-performer interactions werelimited to the periodical review meetings and

milestones. A temporary space was set up for thesemeetings, but the communication between a projectteam and clients was episodical and badly supported.This led, sometimes, to reciprocal misunderstandings.

The space organisation at DA in 1999In 1998 DA moved to the second floor of an oldindustrial building, always in Milan. The new layout isdivided in three main parts: the educational space, theresearch centre and, in between, a management andinternal services third area where business, corporatecommunication, and administrative activities arelocated (see figure 3).

When you enter Domus Academy there are a smokingarea and a wide hall where you can find all kinds ofpeople –students , designers and visitors–.From this hall you can access both the educationalspace (turning right) and the staff offices and theResearch Centre workplace (turning left).In the educational space there are various classroomsand a computer room, a library plus an open spacewhere students work (alone or in small groups), discussand brainstorm during project sessions.To the left of the entrance hall there is a long corridorwhere all the internal staff offices and a meeting roomare located.At the end of the corridor there is the DARC workplace,consisting of an open space (the “laboratory”) wheredesigners work side by side and concurrently ondifferent projects as well as the offices of the DARCManager and of the supervisors of the different researchand design areas (namely: Interaction, Communicationand Material Design).

The new Domus Academy layout has significantlychanged the way designers and students communicateand interact. In fact, on the one hand, designers do notpass through the education area in order to reach theirworkspace and, on the other, students do not go to theDARC very often since they would have to passthrough the business administration area. Moreover, aswe will see later, the DARC workspace has lost itscharacter as a place of conversation and exchange ofviewpoints.

From 1996 to 1999 DARC has greatly increased itssize: the number of projects has more than doubled and,consequently, so has the number of designers. As thegrowth is mainly due to the Interaction Design area,DARC’s technological equipment has expanded too.The DARC today has a “per person” organisation of thespace. The laboratory is divided in two parts by a set ofshelving. The first hosts the Communication andInteraction Design working areas, while the other hasno specific destination.Some Interaction Design senior designers sit there whenco-ordinating activities within the projects while theMaterial Design staff use the other two desk areas.From an ICT equipment viewpoint, each designer hasher own workstation with a suite of productivity anddesign tools.

84

Figure 3 - DA’s layout in 1999.

Today computers represent the main working tool fordesigners: people use computers to create artefacts(with the support of graphical applications), to writepapers, to navigate the Web looking for all kinds ofinteresting news and to communicate via email.A key point: email is used for communicating both withdistant people and with other DARC designers, evenwhen they are in the laboratory.

Figure 4 - The DA Research Centre today: a “perperson” organisation of space.

As a matter of fact, the workspace is organised nolonger by projects but according to roles –designer,senior designer, supervisor– and to design areas (seefigure 4). This space organisation is more likely toinduce individual rather than team activities, even if theDARC workspace is now often crowded and noisybecause of the increase, over recent years, in thenumber of people working there.

In sum, on the one hand there is a sort of «loss of sightof project boundaries» and on the other it is difficult tograsp all the benefits related to a «per-person-organisation of space».

Due to the lack of a shared project context,communication among team members is sometimesdifficult and knowledge sharing is no longer a naturalfallout of their social behaviour. DARC designers have

learned to use the email to overcome these limits,frequently exchanging messages even with colleagueswho are also in the laboratory. However, theeffectiveness of their communication is not the same asbefore. Moreover, the individualised space organisationand the absence of decoration characterising a projectneither facilitate the exchange of viewpoints andexperiences both among designers and with visitors norsupport their mutual awareness [1]. The DARCmanager tries to overcome this problem by inviting toany project’s brainstorming and survey meetings alsosome senior designers who are not directly involved init. Meetings in the entrance hall for a coffee are alwaysa good occasion for increasing awareness of what isgoing on at Domus Academy. Nevertheless, theperipheral participation of all DARC designers andeven occasional visitors in all DARC’s activities is nomore possible.

Another consequence of the current space organisationis a neat separation between individual and team workactivities. In the laboratory designers performindividual activities or conduct only brief exchanges ofviews. Meanwhile group activities are generallyperformed in the meeting room, where no trace of theirwork is visible except for the documents and drawingsthey bring there when a meeting is arranged. As ourobservation at Philips Design has also shown [6],communication is more effective when it occurs in aspace whose decoration brings forth the context of theactivity to which it makes reference.

As far as interaction with customers is concerned, todaycommunication is still episodic and sometimesineffective. The project leader remains the«communication bridge» between the project team andthe client(s) (with all potential breakdowns andbottlenecks that this solution can generate) lettingdesigners develop the creative phase without any directinterference by the client(s) so that their imaginationcan run free. However, the DARC ICT infrastructurecould support the creation of a virtual space wherecommunication and interaction between the projectteam and the client(s) could take place withoutinterruptions, preventing breakdowns and bottlenecks.Remote participation in project activities on the part of

85

senior designers and project leaders –often away onbusiness– has been also improved and can be furtherimproved by the ICT infrastructure.

In conclusion, in its new location Domus Academy hasmoved to a more powerful technological support systemallowing its designers to enrich their communicationand design equipment, but its space organisation (and inthe DARC particularly) does not lead to a naturalsupport to communication, co-operation and knowledgecreation and sharing. The latter are very important increative organisations since they represent not only theway through which a design studio improves its skillsand refines its style, but also main components of thedesign outcome itself. The ICT infrastructure of DomusAcademy and the way it is already widely used by itsmembers for both designing and communicatingsuggests that through it the DARC workplace can beaugmented so that co-operation and knowledge sharingwithin and across project teams are enhanced.

AUGMENTING THE DOMUS ACADEMY SPACEThe comparison of the 1996 and 1999 observations atDomus Academy allows us to characterise our analysisin terms of gains and losses in the passage between thetwo settings. Moreover it allows us to underline howthe practice of the DARC members immediately tried toovercome the problems that could arise from the newworkplace organisation, how they are learning to liveand work as effectively as possible in it. Diachronicanalysis seems quite useful in orienting design choicessince it characterises the observed situation indynamical terms. Thus, the designed solutions areevaluated with respect to their impact on the changingpractice of the DARC designers and on their learning.

The story we recalled in the previous section shows thatspace arrangements play a very important role in thepractice of human beings. The place we live in shapesour behaviour as well as the language we use to speakabout it. Whenever we move to a new place, we needtime to feel at home in it, to connect with it. Wetherefore took a straightforward approach inconsidering even the ICT infrastructure we had todesign a means of transforming the DARC workplaceby augmenting it.

Our analysis shows that the new DARC workplaceintroduces new limits with respect to the previouslocation of Domus Academy . The physical layout (theseparation between educational space and the DARC),the enrichment of the technological infrastructure (aworkstation and a desk for each designer), and theincreased size of Domus Academy and in particular ofDARC, prevent the workplace to support per se theawareness of its inhabitants [1]. Moreover, it sharesother limits with the previous location: the absence of aplace devoted to the interaction with clients and the useof the open space for both individual and team workactivity impact on performance.Therefore augmenting the DARC workplace meansdesigning its technological infrastructure so that it canserve the different needs of the users in different

situations.

The new system should make the physical space plastic,so that it can be tailored to meet different needs in anysituation. This new augmented space should involve notonly the DARC open space and its offices, but also anyother used location, like meeting rooms , remotelocations hosting designers when travelling, offices ofclients when communicating with designers. The designof the new system is centred on the projects carried onat Domus Academy, since the internal organisation isbased on projects. The same situation has been detectedat Philips Design and other design centres, so that it canbe considered general enough to design the wholesystem around projects.

The system will be based on a knowledge managementsystem that collects and organises every informationabout projects. This includes formal and informaldocuments, participant profiles, conversations anddiscussions, comments and sketches , and everythingelse is useful to build up a comprehensive description ofthe projects ' development from its definition to itscompletion. The knowledge will be made available tousers in different ways according to their profile andtheir current situation. The quality of the presentation isfundamental to achieve the satisfaction of the users, andconsequently to make the system successful. A shared,comprehensive view of a project will be made availableto members of the project team to provide them with abackup repository of the project knowledge. Moreover,partial and tailored views of it will be used to addressdifferent types of activities: individual work, face-to-face meetings, project meetings, distant co-operation,etc.. A partial view of the project will be created in avirtual space where the customers can interact with thedesigners reaching an adequate awareness of projectevolution.

Meeting rooms will be equipped with wall projectors orlarge screens to decorate it with inspirational imagesand drawings to let participants feel as if they were inthe project workplace. The same vein social spaces, likethe entrance hall, will be equipped with wall projectorsto show a selection of evocative images presenting theongoing projects and soliciting the curiosity of who isthere to trigger conversations about the projectthemselves. Any social space should have windows onwhat is going at Domus Academy , in order to increasemutual awareness.

These are typical examples of the weak augmentedreality perspective. Without changing the physicallayout of the Domus Academy location, withoutcreating new devices nor transforming existing objects,only using typical interfaces like PC’s and largescreens, the space becomes flexible and plastic.The entrance hall becomes the place in which peopleare updated on activities carried on in DomusAcademy ; the meeting room becomes part of a projectworkplace during members' meetings; the DARCworkspace is at once an open space where designersperform their individual activities and the home of the

86

different workplaces of the ongoing projects; a newvirtual place plays the role of a permanent dedicatedspace for interacting with the customers.

The place of the DARC will be changed in a way that isnot possible in the physical space by the new ICTinfrastructure and services, since they create newinteraction possibilities. Weakly augmented reality isnot looking for virtual extensions of the space that areinvisible to human eyes; rather, it aims to create placeswhere people can live and practice in ways extendingand modifying their previous habits.

A goal of the system we are developing is to minimisethe human effort needed for using and maintaining it.Otherwise, it would increase the articulation of work[19], distracting users from their focus on design. Theinformation and the knowledge are captured from theevery-day activities of designers, and stored in theknowledge bases associated with projects. The projectknowledge bases are, in fact, the source for any of thedescribed views, like populating the (virtual) place forthe interaction with the customers or decorating thewalls of the entrance hall.

The presentation of the knowledge must be specificallydesigned to be effective in specific situations. Thesystem must be able to provide a tailored view of aproject when the designer is working at her desk, whenshe discusses a detail with a colleague, or when she ispresenting intermediate results during a meeting.Moreover, the team members should access a view ofthe project knowledge according to the role they playand the situation they are in : for example, the projectleader, when working alone, can access managementinformation on the project that should not be displayedto designers. The system is requested to automaticallydetect the current situation and behave accordingly.

The system we are designing requires the devise of anew kind of interface that can host several, relatedinformation. The basic idea is to make the focus of eachview the centre of the presentation surrounded by itscontext, i.e. by the collection of knowledge items ,information, and data related to it. For example, thecontext of a document includes related documents ofdifferent kinds –sources, inspirational, references,historical–, people somehow involved –authors, clients,developers, managers, experts–, related knowledgeareas, and so forth. Figure 5 illustrates a visionaryinterface in which the evolution of the central documentis browsable (going along the stream that collects theseries of document's versions), and the displayedversion is surrounded by its context (the ring around thedocument). A goal of the Klee&Co project is to deviseand implement an interface that comes close to thisvisionary interface.

In conclusion, the system we are developing musttherefore be very flexible , capable of managing anddistributing the knowledge created within a processwherever it is needed and presenting it in any situationin a suitable way.

wherever it is needed and presenting it in any situationin a suitable way.

Figure 5 - A "visionary" interface.

THE KLEE&CO SYSTEMAs discussed in the previous sections, our aim is toaugment the place of DARC, in such a way thatdesigners can practice without the constraints createdby the physical arrangements of their location. Themain problem is therefore to offer them the access tothe knowledge, supporting their awareness. Sincedocuments are a relevant part of the explicit knowledgecreated within design, the system we aim to develop hasto provide multi-view access to project documentation.The goal is to give the user comprehensive informationabout technical contents and related documents fromboth Domus Academy repositories and external sourcessuch as the Web. Moreover, non-technical issues likethe process that delivered the current version of adocument or the debate that underwent the decisionshave to be properly addressed. This information formsthe context of a document, to give to the userknowledge about documents throughout their entire life.

A first step toward an integrated environment is thecapability of managing documents of different naturealong with associated information that forms thedocument context. This requires a semi-structureddefinition of documents and a system that fills andmanages the structures according to the nature of eachdocument. Next a system to support the activity withina company must be devised and developed. This systemrequires an open architecture since it has to include thetools already in use, filters and adapters to treat specificfile formats, and new components that implement theinnovative functionality.

87

In what follows, we discuss the structure of a documentand outline the architecture of the Klee&Co system.

The document structureA comprehensive knowledge management systemneeds to handle different kinds of items. Examples areprojects, user profiles and documents. A project is theusual context in which DARC members perform theiractivities. A project can be modelled as a collection ofdocuments and people. People are described by userprofiles that capture roles, interests and activities of thepersons who interact with the system. Profiles are usedto customise the system behaviour according to userexpectations. A document is anything produced withina project.

The typical document at DARC is a text or a drawingfile created by an application like MS Word orAutoCAD or any graphic design system. In our contexta document becomes a complex entity in which thetypical text or drawing file, i.e. the content, is just oneof the composing elements. Moreover, associated with adocument there are its versions to capture the evolutionof its content.

Each (version) of a document is composed of thefollowing parts:§§ General information: author(s), format, location, ...§§ The content: text, drawing, pictures, ... in specific

format;§§ Content representations: summary, abstract, HTML

version, ASCII version, thumbnail, preview, ...;§§ Knowledge information: kind of content, keywords,

timestamps and other indicators' values;§§ Context information: notes, conversations, related

documents:

The first two parts cover the typical informationassociated with any document file. The others arespecific to the Klee&Co system. Each contentrepresentation supports a different view of thedocument. For example, the HTML format of a MSWord document allows the visualisation of the contentin a Web browser, the summary could be used to give agood understanding of the content without displayingthe whole document, and the ASCII version is betterexploited for automatic processing.

Beyond supporting typical personal use, representationsaddress the use of documents in a social context. Forexample, a MS Power Point version makes a documentsuitable for presentations during meetings; a selectionof pictures together with the indication of the name ofthe project and the name of a reference person couldform a document representation to be displayed in apublic space like the entrance hall. The Klee&Cosystem will be able to build up and select the rightrepresentations of documents, either automatically orwith user help.

Content representations are also important since theyallow for the inclusion of non-electronic documents inthe knowledge base. Assume that in a project a wooden

mock-up of an artefact has been made. Of course, themodel cannot be included as is, but it is possible torepresent it electronically. Representations are usefuleven for electronic documents since the original files,e.g. MS Word or AutoCAD files, are often difficult tomanage because of their proprietary formats and oftheir being too detailed to be presented as they are.Representations overcome these two limits byproviding the best view of a document in every context.

The information associated with a document supportsthe definition of its context and therefore of itsrelationships with other documents. For example, thekind of content specifies whether an MS Word file is atechnical report or a marketing presentation, whilekeywords capture the knowledge embedded in thedocument to allow for document classification andcomparison.

More specific relational indicators can be introduced todefine the relationships among documents. Thetimestamp indicator, for example, defines the temporalrelation among the documents created within a project.Assume that there is a project plan that defines theexpected time dependence among a set of documents.Two possible representations of this relation can begiven: a Gantt chart where documents are associatedwith timestamps or a file of documents labelled by thevalues of the timestamp. In the former case, a referenceto the Gantt would be included in the contextinformation of the documents, in the latter thetimestamp values would be included in the knowledgeinformation section.

Beyond references to related documents, the contextinformation associated with documents allowsmanagement of (short) documents that are exchanged asattachments to another document. Examples of theseare notes and conversations. A note is a short textassociated with a document or part of it. Typically, anote is a comment or a sketch on a picture or a chunk oftext. A conversation is similar to a news thread andidentifies a sequence of (mail) messages to discuss orcomment on (part of) a document.

The Klee&Co architectureThe objective of the Klee&Co system is to supportusers in every activity they perform within DARC. Thismeans that we need to augment a traditional DocumentManagement System, in our case DocuShare by Xerox,with new functionality. The system should be able tounderstand the form of use and behave consequently; itshould be possible to save and restore the currentworking session; a seamless switch between commontools, like editors and word processors, and our systemshould be supported; uploading of documents andinformation should be assisted by proper tools to collectand organise the knowledge; accordingly, output toolsshould address knowledge retrieval and presentation.

The Web technology supports remote access from anykind of platform and connection to many commontools, it is the best choice for any successful knowledge

88

management system that needs to be easily accessiblefrom any computer and location. Web access allows forremote connection and interaction to both internal andexternal users. Internal users, like members of a projectteam, can work from home or any other location almostas if they were at their workplace in DARC. Externalusers, like clients, can use the same environment toaccess the workspace created for interacting with them.

The Klee&Co system is a Web-based system. Themulti-tier architecture of the current prototype issketched in Figure 6. A key issue in designing thesystem was the inclusion of existing tools into theoriginal structure of DocuShare. The solution was theintroduction of a new tier that acts as a bridge betweenthe web server and DoscuShare, and interfaces COTSand tools. The Klee&Co tier traps and processes therequests coming from the user and prepares the output.Moreover, the Klee&Co tier includes a set of agentsthat collect and process data to build up the knowledgerelated to projects, people and documents to meet therequirements described in the previous sections.

Figure 6 – The Klee&Co system.

The client side is a common Web browser augmentedby software components like plug-ins, JavaScriptscripts and Java applets to enrich the user interface.Moreover, various applications will be embedded toaddress specific activities, like PlaceWare to supportconferences and AskOnce to enhance searching (bothby Xerox).

The Klee&Co system offers different interactionmodes, according to the profile of the connected user, tosatisfy the different needs. For example, the project co-ordinator can discuss a detail of a product with the teamwhile sitting in a hotel room. The client can follow thedevelopment of a product by browsing and discussingthe documentation made accessible to her at any time.The system interface resembles the visionary interface

presented in Figure 5. Figure 7 is a snapshot of theinterface of the current prototype. The centre of thescreen is occupied by (a representation of) a document.The windows around it define the context for thatdocument. Both the document and the context areactive. For example, when going on a note icon with themouse pointer on the document, a window opens overthe document to show the content of the note and theicon of the author is highlighted; moreover, by clickingthe note it is possible to add a comment in differentways through a pop-up menus. In the same way, goingon a related document icon the system opens a windowwith further information, and clicking makes thatdocument the centre of the presentation and triggersscreen reorganisation.

Figure 7 - A snapshot of the Klee&Co prototype.

The document management system is the back-end thatacts as a knowledge repository. Tools like mailingsystems, search agents, and awareness agents –beyondthe information supplied by the users– contribute tobuilding up the knowledge. Filtering agents process theinput to extract and store the knowledge. Examples offiltering agents that will be embedded into the Klee&Cosystem are information filters, conversation managers,document categorises (WebSOM; [12]) and keywordextractors (KEA; [9]). Smart filters will play anessential role since they have the task of automaticallybuilding the representations and extracting knowledgefrom the uploaded documents. Output agents willprocess knowledge before being sent to the user. Theseagents have the task of formatting what will bedisplayed by the user interface, that is, the browser andthe software mentioned before.

CONCLUSIONBy observing and analysing the practice of designers atDomus Academy and designing a system to augmenttheir workplace, we have discovered that within a socialcontext the sense of place is not directly related to theperception of its spatial dimension, but rather to itscapacity of bringing forth its main features from thepractice point of view. This has more to do with thelanguage it offers its inhabitants in speaking about theircommon operations and about the things they createthan with its creating realistic simulations of the

89

physical space.

Augmented places are not virtual copies of a physicalplace: their spatial arrangements are not possible in thephysical reality, they are the genuine outcome of thedesign of a new form of space, with qualities likeplasticity and flexibility quite absent from the physicalspace where we live.

Only after we have conceived the new qualities of anaugmented place we can shift our attention toimproving its perception by its inhabitants. But hereanother story begins starts , one we are not yet ready totell.

ACKNOWLEDGEMENTSThis paper reports on research done within the ECfunded Esprit Project, Klee&Co. The authors gratefullyacknowledge the other participants in the Klee&Coproject and in particular Alessandro Rancati of DomusAcademy, Lucia Massarutto and Thomas Schael ofIRSO and Luca Bernardinello and Domenico Perrottaof DISCO at University of Milano – Bicocca for themany discussions on the issues presented here.Particular thanks go to the members of the DomusAcademy Design Center for the patience and spirit ofcollaboration they exhibited while observed by ourteam.Finally, the authors wish to thank the anonymousreviewers who offered comments helping them to focusthe Klee&Co project toward its goals.

REFERENCES

1. Agostini A., De Michelis G., Grasso M. A., PrinzW., and Syri A. Contexts, Work Processes andWorkspaces. Computer Supported CooperativeWork. The Journal of Collaborative Computing ,5.2-3 (1996), pp. 223-250.

2. Aukstaklnis, S., and Blatner D. Silicon Mirage –The Art and Science of Virtual Reality. PeachpitPress, Berkeley, 1992.

3. Brown J. S., and Duguid. P. OrganizationalLearning and Communities of Practice: a unifiedView of Working, Learning and Innovation.Organization Science 2. 1, (1991), pp. 40-56.

4. De Michelis G. Aperto, molteplice, continuo.Dunod Italia, Milano, 1998.

5. De Michelis G. Cooperation and knowledgecreation. In: Nonaka I. and Nishiguchi T., Eds.Knowledge emergence: social, technical andevolutionary dimensions of knowledge creation.Oxford University Press, New York, (to appear).

6. De Michelis G., Massarutto L., Pluchinotta C., andScahel T. Users profile analysis and requirements(Esprit Project N° 28842 - Klee&Co – DeliverableD1.1), Milan, 1999.

7. Dertouzos M. L. What will be: how the new worldof information will change our lives. HarperEdge,New York, 1997.

8. Dourish, P., Adler, A., Bellotti, V., and Henderson,A. Your Place or Mine? Learning from Long-Term

Use of Video Communication. ComputerSupported Cooperative Work. The Journal ofCollaborative Computing, 5.1 (1996), pp. 33-62.

9. Frank, E., Paynter, G. W., Witten, I. H., Gutwin C.,and Nevill-Manning, C. G., Domain-Specific Key-phrase Extraction. In: Proceedings of the SixteenthInternational Joint Conference on ArtificialIntelligence. Morgan Kaufmann Publisher, SanFrancisco, 1999, pp. 668-673.

10. Harrison, S., and Dourish, P. Re-Place the Space:the Roles of Place and Space in CollaborativeSystems. In: Proceedings of the ACM 1996Conference on Computer Supported CooperativeWork. ACM Press, New York, 1996, pp. 67-76.

11. Hughes, J. A., King, V., Rodden, T., and Andersen,H. Moving Out of Control Rooms: Ethnography ubSystem Design. In: Proceedings of the ACM 1994Conference on Computer Supported CooperativeWork, ACM Press, New York, 1994, pp. 429-440.

12. Kaski, S., Honkela, T., Lagus, K., and Kohonen,T., WEBSOM - self-organizing maps of documentcollections. Neurocomputing 21 (1998), pp. 101-117.

13. Lave, J., and Wenger, E. Situated learning.Legitimate peripheral participation. CambridgeUniversity Press, Cambridge, 1991.

14. Leiva-Lobos, E., De Michelis, G., and CovarrubiasGatica, E. Augmenting and Multiplying spaces forCreative Design, In: Proceedings of Group’97,ACM Press, New York, 1997, pp. 177-186.

15. Mackay, W.E. Augmented Reality: linking real andvirtual worlds. In: Proceedings of ACM AVI '98,Conference on Advanced Visual Interfaces.L'Aquila, Italy: ACM Press, 1998.

16. Milgram, P., and Kishino, F. A Taxonomy ofMixed Reality Visual Displays. IEICETransactions on Information Systems E77-D.12(1994), pp. 1321-1329.

17. Rheingold, H. Virtual Reality . Summit, New York,1991.

18. Shapiro, D. The limits of Ethnography: CombiningSocial Sciences for CSCW (eds) Proceedings ofthe ACM 1994 Conference on computer SupportedCooperative Work , ACM Press, New York, 1994,pp. 417-428.

19. Schmidt, K., and Bannon, L. Taking CSCWseriously. Supporting articulation work. ComputerSupported Cooperative Work. An InternationalJournal , 1.1 (1992), pp. 7-40.

20. Wagner, I. The Computer-Supported Design ofArtefact and Spaces (DESARTE) in Architecture,Landscape Architecture, Industrial Design, ITDesign (Esprit LTR Project N° 21870 – Desarte -First Phase Final Report: part 1), Vienna, 1997.

21. Want, R., Hopper, A., Falcao, V., and Gibbons, J..The Active Badge Location System. ACMTransactions on Information Systems, 10.1 (1992),pp. 91-102.

90

22. Weiser, M. The computer for the twenty-firstcentury. Scientific American, (September 1991),94-104.

23. Wellner, P., Mackay, W., and Gold, R. (eds.)Special Issue on Computer-AugmentedEnvironments: Back to the Real World.

Communications of the ACM, 36.7 (1993), pp. 24-97.

24. Wenger, E. Communities of Practice. Learning,Meaning and Identity (Learning in Doing: Social,Cognitive and Computational Perspectives).Cambridge University Press, Cambridge, 1998.

91