Collaborating with writing tools

Collaborating with writing tools

Interacting with Computers 15 (2003) 737–757 Collaborating with writing tools An instrumental perspective on the probl...

167KB Sizes 1 Downloads 52 Views

Interacting with Computers 15 (2003) 737–757

Collaborating with writing tools An instrumental perspective on the problem of computer-supported collaborative activities Teresa Cerratto Pargman Department of Computer and Systems Sciences, DSV, Stockholm Universitet/KTH, Forum 100, 16440 Kista, Sweden

Abstract This paper presents an analysis of the modifications that a synchronous computer support for collaborative writing introduces into the organization of co-authors’ writing. The analysis is grounded in case studies of different groups of co-authors writing a report together face to face and at a distance through a collaborative writing computer system. Drawing from these studies I suggest that the problems with using a collaborative writing computer system to provide a fully collaborative writing environment derive from underlying assumptions concerning collaboration within the coauthoring activity. I point out that a more thorough understanding of how co-authors organize their writing can provide resources to envisage more radical solutions to the problem of computer support for collaboration. I conclude by considering ways that might be adequate to reconfigure collaborative writing systems in order to provide more satisfactory support for collaboration in writing environments. q 2003 Elsevier B.V. All rights reserved.

1. Introduction Despite the central role that written text and collaboration play in many different contexts and the substantial corpus of research concerned with the design of collaborative writing prototypes, there still has been no successful development of actual use of such kinds of support in real settings. Still, after more than a decade of selling commercial software to support collaborative writing, little software is able to maintain and facilitate human collaboration. The more recent use of the web as an alternative infrastructure for the development of collaborative writing applications is still far from providing appropriate support for the complex activity of writing documents together. These are E-mail address: [email protected] (T. C. Pargman). 0953-5438/$ - see front matter q 2003 Elsevier B.V. All rights reserved. doi:10.1016/j.intcom.2003.09.003


T. C. Pargman / Interacting with Computers 15 (2003) 737–757

just a few of the facts raising serious issues in the research on computer supported cooperative work: Can computer systems support writing and collaboration at a distance? What do we know about human collaboration when we write documents together? What are the assumptions about the collaborative activity that informs the design of collaborative technologies? This paper attempts to discuss these issues based on case studies. Focusing on the use of computer systems in real collaborative writing activities, I analyze the modifications that a synchronous computer-supported collaborative writing system introduced into the collaborative activity of different groups of graduate students preparing an academic document together. The study is grounded in an instrumental approach (Be´guin, 1994; Rabardel, 1995; Ve´rillon and Rabardel, 1995; Cerratto, 1999; Be´guin and Rabardel, 2000; Cerratto, 2000; Cerratto and Rodriguez, 2002) and also draws on concepts taken from activity theory (Engestro¨m, 1987, 1990, 1999; Bannon and Bo¨dker, 1991; Kuutti, 1996; Kaptelinin, 1996). Results of the studies show that the use of collaborative technologies has profound consequences on co-authors’ activity and end product. The analysis of the impact of the computer system on coauthors’ communication and final documents indicates that they have communicated very little with each other; that their interest in the evaluation of others’ ideas and texts was low; and that they spent much more energy and time trying to coordinate their actions than co-elaborating the contents of the text. The characteristics of the communication observed are consistent with the type of document produced. According to the teacher, the shared documents written with the computer-supported collaborative writing system “lack critical reflection on the subject analyzed and also on the methods. They have written, but that is not enough. The analysis is superficial. They should have taken care of the style and the orthography”. Even more important, co-authors indicate that the document written with computersupported collaborative writing system support is not really the product of a true collaboration between thinkers. They said they had a hard time recognizing themselves as the authors of the document produced. They agreed on the idea of having produced a document following the instructions provided by the teacher but they felt that the use of the collaborative writing system did not provide them with adequate conditions for the development of an interesting and critical discussion of the content. These findings revealed that the use of collaborative writing systems encompassed a rather restricted view of collaboration in the context of creative activities such as writing. Many of the difficulties that arised when using collaborative computer systems seem to derive from the assumptions about the collaborative activity that informed the design of the collaborative writing technology. It is thus necessary to understand the specificities of computer-mediated collaborative writing activities in order to better inform the design of collaborative systems. In this context, the concept of instrumented activity (Vygotsky, 1934/1997; Rabardel, 1995) is promising. It opens new ways of thinking about the design of computer-supported collaborative writing systems and allows us to consider ways to reconfigure collaborative writing systems in order to provide more satisfactory support for collaboration in writing environments.

T. C. Pargman / Interacting with Computers 15 (2003) 737–757


2. Background Collective aspects of human activities aroused interest and were studied already in the late 60’s by French language ergonomists (Leplat, 1993), although they were rarely taken as a central object of research. The emergence of computer systems designed to facilitate collaborative activities on the one hand, and the need to have accurate descriptions of collaborative activities on the other, have triggered a general interest in the study of collaborative situations. This interest took more concrete form with the foundation of a specific field of research called computer support for collaborative work (CSCW), emerging in the mid 80’s (Grudin, 1994). CSCW covers, among others, the study of tools and techniques of computer-supported collaborative writing systems as well as their psychological, social and organizational impacts. The study of computer-supported collaborative writing, particularly in the form of a shared editing environment, emerged some years later emphasizing a more task-focused approach to CSCW applications (Miles et al., 1993). Early interest in the field of CSCWriting focused on the design of computer systems. Many designers were tempted to extend existing single-user tools, such as word processors or outliners, to support multiple users. Others concentrated their efforts on finding solutions to constraints of time and distance and only a few showed interest in the study of collaborative writing activities in real contexts. This had important consequences for the understanding of the problem of computer support for collaborative writing. According to Neuwirth et al. (1990), early software that was designed to support collaboration and coordination was inappropriate. Software for collaborative writing has also been perceived as poorly integrated with other writing tools and as likely to force users into premature commitment (Sharples, 1993). The current generation of editors focus on a very detailed set of activities required for the production of typewritten documents instead of considering the overall strategies and activities involved in the production and development of documents (Kirby and Rodden, 1995). Appropriate design of systems for collaborative writing needs to take into account an array of social factors and practices that are commonly taken for granted and that are invisible in normal face-to-face collaboration (Newman and Newman, 1993). As Heath et al. (1995) emphasize, collaborative activities rely upon individuals subtly and continuously adjusting their access to each other’s actions. An individual’s ability to contribute to the activities of others and fulfill his or her own responsibilities relies upon peripheral awareness of the current situation and monitoring of others’ activities. Questions about supporting activities that are collaborative have often been understood from a perspective that distinguishes individual from collaborative aspects of human activities. In this paper, collaboration is neither understood from a purely individual nor a purely social perspective. Collaboration involves the ongoing and seamless transition between individual and collaborative tasks, where persons are simultaneously participating in multiple, interrelated activities.

3. Computer-mediated collaborative writing as an instrumented activity The question addressed in this paper concerns the appropriation “or mastery” of a synchronous computer-supported collaborative writing system by groups of co-authors.


T. C. Pargman / Interacting with Computers 15 (2003) 737–757

This question opens up important issues that implicate the mediation of writing in particular and the mediation of collaboration in general. The issue of mediated writing is brilliantly treated by Haas (1996). Writing is in her words made material through the use of technologies, and writing is technological in the sense and to the extent that it is material. In her studies Haas demonstrates the inextricable and profound relationship between writing and the material world. “Writing has the potential for even more profound transformations of humans because it operates on both these levels-both its psychological (semiotic) aspects and its material-technological aspects have the potential to transform” (Haas, 1996 p.16). For Haas, Vygotsky’s notion of mediation is important in addressing contemporary issues of computer technology and writing because it provides a theoretical basis for the prediction that technology can have real and important, if complex, effects on writing. In Haas’s work, technology matters, for it can bring about important changes individually and culturally. “The technology is transparent myth sees writing as writing as writing, its essential nature unaffected by the mode of production and presentation. The most serious drawback to the transparent technology assumption is that it encourages an overly positive, whole-hearted acceptance of computer technology without any consideration of possible negative effects of that technology” (Haas, 1996 p. 22). Haas underlines that the question of technology for the most part remains latent. Researchers do not address technology as technology; rather, technology is posited as important without being examined in itself in any systematic way. Most studies approach technology as merely a tool, a neutral and transparent means to produce written language, which is somehow imagined to exist independent of those means. Therefore, it is quite logical that the mediated writing activity is considered as unchanged in any substantial way by the ‘transparent medium’ through which it passes; which would also then be true for the transition from single-user to multi-user computer supported writing tools. Haas clearly points to the failure of such thinking and of notions of writing as existing independently of and uninfluenced by technological means. Thus, technologies do matter; they are shaped through our thinking and talking about them, and through our use of them. “Technologies are not static-certainly not static enough to be unequivocally transparent or all-powerful. Rather they are modified subtly and constantly by the uses to which they are put, and by the discourse that accompanies those uses. Technologies continue to evolve, not just because technological breakthroughs but because their contexts of use, and their users, continue to shape them” (Haas, 1996 p.36). The perspective on mediated writing developed by Haas joins Rabardel’s discussions about the status of the instrument in human activities. Rabardel (1995) regards instruments as the result of a developmental process in which the user accommodates and assimilates his/her activities—personal schemes of use—to the properties of the artifact while using and attributing new functions to them (Be´guin and Rabardel, 2000). Special emphasis is thus placed on the idea that an artifact cannot be confounded with an instrument. An artifact only becomes an instrument through the association of personal schemes to the properties of the artifact. The distinction between artifact and instrument is crucial to understanding the aspects involved in the integration of a new computer system into human practices. A computer system does not immediately constitute an instrument for the user. Even explicitly constructed as an instrument, it is not, as such, instantaneously an instrument for the subject. It becomes an instrument only when the subject is able to

T. C. Pargman / Interacting with Computers 15 (2003) 737–757


elaborate it, appropriate it for him/herself or, in other words, when she/he is able to subordinate it as a means to accomplish his/her ends. The issue of mediated collaboration is developed by Suchman (1987), Engestro¨m (1990, 1999), Schmidt and Bannon (1992), Hutchins (1995), Schrage (1995), Rabardel (1995) and Schmidt and Simone (1996) who from different perspectives have contributed to the understanding of fundamental aspects of collaborative activities and the collective. Rabardel (1995) in particular focuses on the status of artifacts in collective activities. In his model for the analysis of collaborative activities with instruments, he conceptualizes the instrumental mediations that an artifact can or cannot support between actors, in relation to their final product and in the relations between the actors themselves. In his analysis of mediated collaborative activities, Rabardel’s point is that it is the actors of a collective who through their interactions ‘with others, with the object of the activity and with the artifacts’ elaborate an artifact into an instrument. Rabardel regards the collaborative writing computer system as an artifact and studies it through its mediational role in the coauthors’ collaboration. According to Be´guin (1994), interpersonal or collaborative mediations involve a fine interplay of coordination and integration processes within a collective. It is through these processes that the actors of a collaborative activity can elaborate a shared product. In order to better understand the evolution of collaborative mediations in the context of computer-mediated collaborative writing I have studied the activity of different groups of co-authors. I focused on their activity through the analysis of their text-based communication and their writings. The initial research question concerned the process of appropriating a collaborative writing system into the co-authors’ activity and it was operationalized into the following questions: What is the impact of the computer-supported collaborative writing system on the coauthors’ communication when they are elaborating their common text? What is the impact of computer-supported collaborative writing system on the coauthors’ collaboration?

4. Case studies: design and procedure I have studied the activity of four groups writing a report for the examination of a course they attended at a Business School. Two of the groups worked face-to-face using a word processor, floppy disks and material from the course; the others two groups had the same facilities but used a synchronous collaborative writing system and worked at a distance. Co-authors performed the writing activities as normal except that groups using computer-supported collaborative writing system wrote the report in the place where the system and the videotape equipment were installed. The following table gives some details about the groups, the time employed in writing the common report and the environment in which they worked (Table 1). The fact that I studied the activity of groups collaborating face-to-face when my interest was in the activity of groups interacting at a distance through the use of computer supported collaborative writing systems has a natural explanation. The activity of the groups interacting face-to-face served as a frame of reference for the interpretation of


T. C. Pargman / Interacting with Computers 15 (2003) 737–757

Table 1 Characteristics of the groups observed Case studies



Documents /marks

Group F Group Y Group M Group X

5 h 20 min 4 h 13 min 6 h 30 min 11 h 20 min

Face-to-face with a word processor

12 pages/15 points 15 pages/13 points 12 pages/10 points 21 pages/14, 5 points

Remote with a synchronous collaborative system

the findings about the use of the computer-supported collaborative writing system. Also, according to the instrumental approach, a new tool does not carry inside itself a new activity that will ‘bloom’ and be developed when the subject(s) starts to use it. On the contrary, the tool will be assimilated and accommodated into an organization predating the tool. In this sense, the new tool will modify the activity at the same time as the user will shape the tool itself. 4.1. The task The task consisted of writing an argumentative report (Cf. Scardmalia and Bereiter, 1987). More specifically, it consisted of a study concerning decision-making processes in French companies. The report comprised an analysis, an evaluation and a comparison between computer-supported decision-making processes (an expert system) and human decision-making processes. The teacher asked the groups to reflect on the limits and benefits offered by expert systems in decision-making processes and write about it. In particular, the students had to think critically about the limited rationality of artificial intelligence tools. The report was part of the examination of the course. The groups had 20 days to write the document. Before they started to write the report, the groups attended a half-day training session with the computer-supported collaborative writing system. 4.2. The authors I observed sixteen subjects who worked distributed in four small groups with four persons in each group. Co-authors were graduate students attending a course of Management and Information Technology at a Business School. They were familiar with working in groups and in particular with writing shared documents; they were furthermore familiar with word processors, electronic mail, spreadsheets, and with using the Internet. They were also interested in new technologies of communication and work. 4.3. The computer program The computer program for collaborative writing consisted of a commercial synchronous collaborative editor, a word processor running on Macintosh computers. The program allows a group of users to work on a single shared document using separate computers. It provides a common writing space through which the group can communicate and interact. The text that appears on each user’s screen is the same text that every other

T. C. Pargman / Interacting with Computers 15 (2003) 737–757


Fig. 1. Two spaces in the collaborative writing interface: the working and the communication space.

user sees and has access to. Each person can look at different parts of the document. The overall document is always the same. Remote users’ selections show up as hollow rectangles, whereas the local user’s selections appear in the usual manner, as a reversed rectangle. The computer-supported collaborative writing system provides a chat-box for communication, electronic pointers that appear on the other users’ screens. Because of the technical problems of distributing a text document across a computer network and still maintain consistency of the document, a number of control mechanisms are built into the system. For example, all users are able to enter text at the same time but only one user may edit a given paragraph at a given time. A bar to the left in the text window indicates that someone has the paragraph locked (see Fig. 1). A black bar represents the local user whereas a gray bar represents a remote user. 4.4. Data collected and method The groups were observed, videotaped, and interviewed. The communication within the groups using the computer-supported collaborative writing system was logged while the face-to-face communication was transcribed from a video recording. Drafts and the final version of the common text were collected as well as the comments of the teacher who evaluated the quality of the group reports. As communication and writing are highly related in collaborative writing (cf. the interfunctionality of writing and communication pointed out by Plowman, 1993, 1995) two units of analysis were used: exchanges (verbal and written) and paragraphs (written). The exchange was the unit of analysis for communication. An exchange equals a turntaking unit. I have coded exchanges according to the number, form and the type of the exchanges produced during the first 30 min of the activity (Kerbrat-Orecchioni, 1992; Cerratto, 1999, 2000; Cerratto and Rodriguez, 2002). The term form refers to minimal and complete exchanges. A minimal exchange is the production of a set of initiative-response utterances. A complete exchange is the production of a set of initiative-response-evaluative utterances. The choice of minimal or complete exchange is important since evaluative utterances, i.e. complete exchanges, imply that the interlocutors are evaluating the ongoing exchange, which in its turn implies a more significant engagement with the ongoing activity (cf. Kerbrat-Orecchioni, 1992; Severinson-Eklundh, 1986). By type, I mean the object of attention that co-authors refer to in their exchanges during the collaborative activity. I have distinguished between three categories: co-authors’


T. C. Pargman / Interacting with Computers 15 (2003) 737–757

actions, the content of the shared document and the artifact (Cerratto, 1999, 2000; Cerratto and Rodriguez, 2002). The paragraph was the unit of analysis of the written text. Paragraphs were distinguished according to the following writing operations: repetition, modification, addition, deletion and transferal (Besse, 1993). The first, intermediate and final versions of the documents were analyzed.

5. Findings 5.1. Overview of the activity observed To contextualize the collaborative writing activity studied, I here provide an overview of the different phases observed in the collaborative writing situations. As we are still far from a complete model of collaborative writing activities, even though Ede and Lundsford, (1990), Beck (1993), Posner and Baecker (1992) and Sharples (1993) have provided rich accounts of co-authoring, I will refer to collaborative writing in relation to solo writing (Flower and Hayes, 1981; Sharples and Pemberton, 1988; Sharples, 1996; Haas, 1996). The reason for doing so is to refer to the specificities of writing when coelaborating a shared document and to emphasize differences observed in the activities of the different groups. From the study of different collaborative writing situations, two main phases of the activity can be distinguished: the planning and the production of the shared document. Planning collaboration is central when writing a document together although the participants do not always respect the plan or agreement produced. Without discussing the modalities, times, places and tools with the help of which they will collaborate, the group has few chances to carry on the work of writing. In a collaborative situation, planning results in the establishment of a frame for the collaborative writing project. The aim of this phase is to produce a plan for the group of co-authors involved in the collaboration, not to produce a plan for the common document although co-authors may start to discuss and brainstorm ideas about their writing. In particular, this phase involves actions such as (a) defining the main ideas to treat in the document, (b) the resources that each co-author has in order to support interesting ideas, (c) defining the form in which the group will collaborate and (d) allocating tasks within the group (cf. Kraut et al., 1988). Planning collaboration is a phase that involves a great deal of discussion. Co-authors define how they will work together and what type of text they are going to write. Questions taken up during this phase are: what are the main ideas we would like to write about? What arguments do we have for developing idea X? How are we going to work as a group? Who is going to do what? Producing the shared document involves the concrete work of crafting it. In this phase the following actions are developed: sketching, composing and reviewing the document. These actions are the same when writing a non-shared document. The difference when writing a shared document is that co-authors have to explicate their ideas in order to decide and agree with the other writers. Negotiation of ideas and text become a must in a collaborative writing situation. An important difference compared to writing a non-shared

T. C. Pargman / Interacting with Computers 15 (2003) 737–757


document is that the actions of composition and reviewing, which are extremely intertwined in solo writing, and even more so in computer-mediated writing (Haas, 1996), can be more easily distinguished in a collaborative situation. For example a co-author composing a document may stop her production until the other co-author is able to review her writing. Writing collaboratively thus carries with it a double development of the actions of sketching, composing and reviewing, at a subjective level of writing and at an interpersonal level of writing a common document. Another critical characteristic of writing a common document is that the main phases of planning and producing develop in a sequential manner. In relation with solo writing collaborative writing seems to be a much less flexible activity. For example, when writing a solo or non-shared document, it is possible to plan and plan again the document while producing it. It is even common to change the main ideas of the initial document and end up with a quite different text. In a collaborative situation, changing the frame of the document that the group had agreed on while producing it is rarely done. It would mean questioning the way in which the group is collaborating. Modifying prior agreements while writing local pieces of text would demand a lot of effort and would increase the risk of writing an inconsistent document. And what is central in the collaborative activity is that people are not only managing a piece of text production but also engaging in interpersonal relations through their text; responsibilities for the production, intellectual ownership and need for consensus etc. (cf. Dillon, 1993). The phase studied in detail in this study was the production of the document. In particular, the actions of composing and revising common documents were analyzed. 5.2. Impact of the computer-supported collaborative writing system on the co-authors’ communication The analysis of the exchanges of the groups collaborating face-to-face and with the computer supported collaborative writing system indicates that there are important differences between the activities observed. First, distributed groups communicated considerably less compared to the groups working face-to-face; second, the content of the exchanges mainly referred to the organization of their actions while face-to-face groups mainly produced exchanges about the content of the shared document; third, the forms of the exchanges showed that there is an absence of evaluative utterances in the dialogues of distributed groups as compared to the communication of face-to-face groups. The following table shows absolute and relative numbers of the exchanges observed during composition and revision phases. The impact of the computer-supported collaborative writing system, and more precisely of the text-based communication on the production of exchanges, was very large. Distributed groups produced only a fraction of the number of exchanges produced by the face-to-face groups: 14%. While the communication of the face-to-face groups increased from composition to revision, the communication of the distributed groups decreased at the same time. It is a fact that text-based communication constrains collaboration (cf. Kraut et al., 1992). However, the constraints observed do not only have to do with the differences between the processes of spoken and written language (Severinson-Eklundh, 1986; Goody, 1994; Ong, 1982). The difficulties observed are primarily concerned with the problem of relating commenting the text with the text. The extra twist is that in order to


T. C. Pargman / Interacting with Computers 15 (2003) 737–757

be able to relate a comment to a paragraph, it is first necessary to be able to distinguish what is a comment. As pointed out by Churchill et al. (2000), “although chat tools are providing very useful support in the workplace for maintaining contacts and exchanging information, they do not support the close integration of documents with ongoing conversations… Text conversations tend to occur in windows that are separate from, and have no relationship to work related resources”. (Churchill et al., 2000 p. 454). The following example illustrates co-authors’ problems in understanding each other’s comments on the text. Example 1 Paul is writing his sections in the shared file when he suddenly discovers a paragraph, written by his co-author (Fernand). The paragraph that does not correspond to what Paul thinks should have been written. Paul addresses his question to the author of the paragraph, who is sitting at a distance. The following excerpt shows the text-based dialogue between Paul and Fernand. Paul: why did you say that there is a diversification of products due to the use of the machine making samples? Fernand: wait, I have to finish my part and we start a new one Paul: yes but don’t forget, No as Kent did it Fernand: which is the part that you would like we start with? Paul: Check my question about the diversification of products due to the use of the machine Fernand: I don’t get it, what you want to say? Paul: and now! Do you see it otherwise follow the face [he refers to the electronic pointer] Fernand: Enter your message and let me read it Paul: Go up in the chat box and look for my question. Fernand: OK. Paul: finally! Fernand: after “it will allow” what you’d like to say? Paul: we are not talking about that, we are talking about the diversification of products, that is below. Paul says to himself “we seem to be dummy people, I have to meet him… after it “would allow”… what do you mean? oh no! he does not understand!” Paul asks a question referring to the content of a paragraph written by Fernand. In fact Paul does not understand why Fernand had written “there is diversification of products”. The question about this phrase is repeated four times. It took a total of 25 min until the other party finally could identify and relate comments with the paragraph in question, that is, 11% of the time employed in writing the document. The micro-sequence of activity illustrates the problem of trying to understand each other’s ideas and references to the text through two writing spaces (chat-box and writing file) that are dissociated from each other. For Paul and Fernand it was hard to collaborate on the paragraph; they did not have the possibility to develop actions that are highly related with writing. Face-to-face groups deployed commenting on, pointing, manipulating and navigating shared text while writing

T. C. Pargman / Interacting with Computers 15 (2003) 737–757


and communicating. These activities to a great extent facilitated co-authors’ focus and understanding. Commenting-on is an activity that organize face-to-face communication of groups elaborating a common text. Although the content of the comment may help co-authors to make sense of others’ ideas, it is how, when and where the comment is written in the writing space (page, file, etc.) that facilitates the understanding of others’ ideas and suggestions. Writing a comment in a margin or between lines or even making an asterisk in a phrase conveys much information about the change to be made. Face-to-face commenting-on made collaboration between authors possible. The space between lines, the margins, and the asterisks associated with commenting, functioned as instruments of collaborative writing. Moreover, it was through the use of the margins, spaces between lines, and asterisks, that co-authors from a single writing space created a common place for co-authoring. It was quite a difficult task for distributed groups to comment on texts. Authors had to look for the comment in the recorded dialogue of the chat box, then search for the paragraph in the shared file and relate to it. When comments are no longer oral, they lose the capacity to work integrationally, to coordinate text and ideas, and instead become a source of misunderstandings. Pointing is an activity that in combination with commenting organized much face-toface collaboration when writing. Pointing essentially worked as a bridge between coauthors’ ideas and texts. Co-authors collaborating face-to-face sometimes used fingers to point at a paragraph being discussed; at other times they drew a circle, an arrow or a cross behind the phrase, underlined the phrase or wrote a question mark behind it. The use of crosses, arrows, circles, lines, and fingers was associated to pointing and worked as an instrument mediating authors’ understanding and collaboration. The computer-supported collaborative writing system studied offers different electronic pointers that co-authors tried to use in the example above to help others find the paragraph under discussion. Although the idea of providing tools for pointing has been implemented in the system, the activity of supporting pointing was not. Co-authors had problems understanding what was pointed at when they were using the pointers. The shapes of the pointers hooked much of the co-authors’ attention. They therefore looked more at the pointer than at the paragraph being pointed at. Also, when using the electronic pointer, it is impossible to see the text being pointed at; the electronic pointer is not transparent and it hides parts of the sentence that is pointed at. Tactile manipulationof text occurred when face-to-face groups were using pen and paper. They held the text in their hands, spread sheets out, pointed when rereading text, and moved toward and away from the text. According to Haas (1996), sometimes tactile manipulation has an obvious purpose, as when one of the writers picks up the document in order to sit back and read it, or when one of the writers shuffles through several pages to find a section she/he wants to revise. At other times, however, such manipulation of the text has no clear purpose. Writers might spread pages out or stack them up in a rather distracted manner. “In these cases, the physical manipulation of text may have provided a way for the writers to take time out—a few seconds in which to collect their thoughts but still be engaged with the text” (Haas, 1996 p.131). The only comparable activity with the computer would include touching the screen, an activity in which neither of the distributed


T. C. Pargman / Interacting with Computers 15 (2003) 737–757

groups engaged. However, one of the distributed groups printed hard copies of the common documents and manipulated when revising the document. One of the authors made notes and comments on the whole document on paper. He afterwards distributed the corresponding pages to each one of the co-authors. Navigation of shared text resulted in a rather disoriented and demanding cognitive activity for distributed groups. As one cannot always be aware of what others are writing, where and when, it was quite difficult for distributed group to construct a mental representation of their text. In particular, reading the shared text online was hard for coauthors. According to Dillon (1994) and Haas (1996), reading is an important part of the writing activity. Distributed groups had problems moving back and forth between writing and reading when producing text online; they could reread to evaluate and to revise locally but not globally; they definitely could not reread to detect coherence problems and proofread. Having the possibility to evaluate others’ texts has important consequences for co-authors’ understanding of the common text and consequently for learning from the text (cf. Haas, 1996). Flower and Hayes (1981) have mentioned that the text produced so far is a powerful guiding force in the composing process for the writer; an existing text may serve a heuristic function in the inventions of ideas. For distributed groups the existing text written by others could not serve this purpose. The example presented above shows that the problem of scarce communication observed between distributed co-authors is not only an effect of the difficulty to communicate via written text. It is also an effect of trying to communicate through an environment that does not support essential and related actions of collaborative writing. Not understanding the sense of a phrase written by another co-author may become a serious concern in one’s relation with the other’s text, in the co-authors’ relationships and finally in one’s relation as a co-author, as showed in the example above. 5.3. Impact of computer-supported collaborative writing system on co-authors’ collaboration One of the specificities of face-to-face collaborative writing is that co-authors’ communication drives the writing activity, that is, the text is produced from the negotiation of ideas and text. Being aware of this characteristic, I analyzed the forms of exchanges produced by the different groups of co-authors while they were writing. The analysis indicated that distributed groups presented a much larger proportion of minimal exchanges, while the exchanges of the face-to-face groups instead were characterized by complete exchanges (cf. Table 2 above). These findings were maintained in both the composition and revision phases of collaborative writing. Face-to-face groups tended to enclose the exchanges with evaluative utterances, which was interpreted as a sign of interest and commitment to the writing activity. Evaluative utterances have a function: to specify what was not understood and to push the discussion further (Kerbrat-Orecchioni, 1992). The following examples of excerpts show the differences between the activities developed by the groups. Example 2 Face-to-face group: Yves, Oscar and Marc are discussing face-to-face. Oscar is the scribe writing on a laptop.

T. C. Pargman / Interacting with Computers 15 (2003) 737–757


Table 2 Differences between groups in terms of number, type and form of the exchanges during the collaborative revision phase Production phase Composition

Face-to-face groups

Distributed groups

Amount Type Action Content Artifact Form Minimal Complete

87 exchanges

17 exchanges

Revision Amount Type Action Content Artifact Form Minimal Complete

17 70 0

20 80 0

13 0 4

76% 0% 24%

22 65

25 75

11 6

65% 35%

117 exchanges

12 exchanges

30 82 5

26 70 4

6 6 0

50 50 0

43 74

37 63

9 3

75 25

Yves: and we ask the concerned persons directly [INITIATIVE] Oscar: you mean the group of employees… [RESPONSE] Yves: it is not the group of employees but the group of concerned persons [EVALUATIVE] Marc: the group of concerned persons [EVALUATIVE] Oscar: (writes) ok, and after that? [INITIATIVE] Distributed group: Martin and Paul are discussing with Fernand and Kent M&P: why did you say that there is a diversification of products due to the use of the machine making samples? [INITIATIVE] F&K wait, I have to finish my part and we start a new one [RESPONSE] These excerpts exemplify differences between the forms of communication in the two types of groups. The small number of complete exchanges produced by the distributed groups pointed to a contradiction. Collaborative writing is an activity where much of the collaboration is based on the evaluation of others’ ideas and others’ texts. The fact that distributed groups did not evaluate what the others suggested or wrote anticipated an important impact on the form in which collaboration is organized with the computersupported collaborative writing system. Collaboration was reduced to an interaction shaped by “questions and responses” which then had an important impact on the writing


T. C. Pargman / Interacting with Computers 15 (2003) 737–757

activity and on the quality of text. One could ask if an activity in which one has difficulties to evaluate the contributions of others can truly be called a collaborative activity. The forms of exchanges observed in the distributed groups told us a great deal about the writing developed by the groups. The lack of engagement observed in their dialogues is a consequence of the autonomous and rather independent writing that co-authors were forced to develop. Distributed groups adopted the strategy of identifying and allocating different parts of the text for different authors to write separately. Curiously the solution to their collaboration was to write independently. Writing different parts of the text at the same time made evaluating others’ parts difficult. In fact, one cannot review what is still under composition. Not being engaged in what others were writing was a consequence of having different objects of activity driving their writing. Face-to-face groups also wrote parts of the text separately, but with the difference that they reviewed and changed those parts together. Together for the groups working face-to-face meant to discuss an idea and finish by writing or rewriting it into a paragraph. It comes as no surprise that face-to-face groups were engaged in their collaboration when they had the possibility to actually affect the others’ ideas and text. What the face-to-face groups wrote was the result of the ideas discussed. The analysis of the activity of face-to-face groups revealed that they composed and reviewed the same versions of the common document together. The co-elaboration of the document was systematically coordinated with the concurrent actions of the other authors and it consisted of stating, arguing, accepting, refusing and writing down ideas. The distributed situation presented a different organization of the collaborative activity. Distributed groups did not compose and review local parts of the shared document together. Instead, they oriented their activities towards individual objects of writing. One of the distributed groups drew a table to represent the sections of the text to be written and the name of the person responsible for each part. Forty-nine sections were distributed to four members. Before separating them the authors took a copy of the table that they used in order to remember the allocation of tasks agreed upon. The forty-nine sections in the table testify of the complexity and inflexibility of the collaboration in such computer mediated situation. How much can one review the other writers’ sections of the document? How much can one integrate and coordinate sections of the texts written by others? How much can one improvise? The analysis of the type of exchanges confirmed the difficulty that the groups had in elaborating a common object of their activity. To a considerably larger extent, distributed groups oriented their activity towards the understanding of others’ actions rather than towards others’ ideas or paragraphs. The following example illustrates the type of collaboration developed by the face-to-face and distributed groups respectively: Example 3 Face-to-face group. Yves and Oscar are discussing and composing a part of the shared document face-to-face 18 , Yves: ok, let’s recapitulate the main ideas that we have, I mean the strong ideas for each part [CONTENT] Oscar: yes! That’s it, ok, the strong ideas… we have them in our heads… so let’s go… and then we close with the conclusion [CONTENT]

T. C. Pargman / Interacting with Computers 15 (2003) 737–757


Yves: ok . 19 , Yves: Well, eh… which are the strong ideas? [CONTENT] Oscar: concerning the method we have said that the committee consists of 14 people from each division and services that were directly related to them [CONTENT] Yves: Yes . [writes] Distributed group. Martin and Paul and Fernand and Kent are composing the conclusion through the computer-supported collaborative writing system , Martin and Paul: advice for the amateurs just watching, we start the conclusion. Check the flow of my words in the paragraphs and see what is missing (like the desert misses the rain) [ACTION] Fernand and Kent: ok but we are still far from having finished our part . [ACTION] , Martin and Paul: this is boring [ACTION] Fernand and Kent: we are going to look at your conclusion if that does not bother you . [ACTION] , Martin and Paul: the end, title the conclusion [ACTION] Fernand and Kent: ok we arrive . [ACTION] There is a considerable difference between the writing activities developed by the groups. Face-to-face groups organized their collaboration through the elaboration of the content of their text. Face-to-face groups do not have to check the others’ actions when working together; they can see what the others are doing and can thus adjust their actions instantly and only concentrate on the content of the document. According to Luff et al. (1992) the ecological flexibility of working face-to-face can allow participants to view collaboratively a document and discuss the various details it contains. It facilitates a variety of cooperative orientations which both support and are engendered through the particular actions and activities at hand. In face-to-face situations it is common to see ways in which the activity of writing and re-writing becomes public and enables colleagues who are present and yet engaged in potentially unrelated activities to become aware of the reasoning behind making changes to a document. Distributed groups instead had to ask, check and question others’ actions in order to regulate, coordinate and adjust activities. They needed to concentrate on the re-organization of their collaboration. Although some information about others’ actions was presented by the computer-supported collaborative writing system-such as with what paragraph the coauthor is working-co-authors felt the need to confirm this information, asking others.

6. Discussion The analysis of the impact of the computer-supported collaborative system on the activity of different groups showed that co-authors lacked support for collaboration.


T. C. Pargman / Interacting with Computers 15 (2003) 737–757

They produced their documents without really co-elaborating its content. Furthermore, the authors claimed to have experienced writing as an individual production of set of pieces of text putting it together at the end of the activity. Although activities such as writing and communication, for example, were supported by the computer system, the relation between them was not taken into account; neither was the relation that writing and communication maintain with other activities. The design of the computer program used by the groups revealed an understanding of writing as independent of communication, commenting, making notes and pointing. Even more striking, in the design of the software studied, writing is dissociated from reading, navigation, searching and manipulation of shared texts. First, software for collaborative writing are not transparent and they matter. Depending of how they are configured and used, collaborative writing technologies can substantially influence writing, collaboration and authors. Looking at how collaborative technologies impact relationships-between authors, activities and artifacts-should be a criterion for the design and evaluation of collaborative writing software. Second, software for collaborative writing should not be designed to support individuals or groups; they should be designed to support and to optimize relationships between individuals. Tools for collaboration should allow a seamless transition between individual work and collaborative efforts. Design should be aware of the inherent tension between shared space and personal space. The two are not, however, mutually exclusive, and different people have different perceptions of these spaces. Third, writing a document together does not equal solving a fixed problem and it is not always the case that people write a common document because they are looking for efficiency and rapidity. Far from being only a task to be performed, collaborative writing is a developmental, creative and social activity, mediated twice: by writing technology and by the experience of interacting with others. Understanding collaborative writing in these terms as an instrumented activity would also mean to understand the inadequacy of rational models practicing structured decomposition and logical progression of products in software design. Collaborative writing technologies should be designed to support various forms and levels of interactions ‘conversations, sketches, statements, agreements, and conflicts’ and not to predestine any specific set of results. According to Schrage (1995) the trade-off here is balancing the demand for flexibility and adaptability with the need for some sort of structure that can hold the contexts and contents that these collaborations generate (Schrage, 1995 p.167). The results obtained in the study showed important failures in the computer-mediated collaboration of remote groups. But it is important to point out that the specifics problems that were found in the activities of the groups might be due to the specific collaborative writing program used in the study rather than being intrinsic problems of the computermediated collaboration.

7. Reconfiguring collaborative writing systems With the results in mind, design understanding collaborative writing as an instrumented activity would move:

T. C. Pargman / Interacting with Computers 15 (2003) 737–757


From the design of a ‘space for communication’ to the design of a ‘place for mutual understanding’ Supporting communication in the context of collaborative writing is much more complex than just supporting a space for co-authors’ communication. For instance, context plays a significant role in situating text and its comments. When the context of a text is unclear there is a potential for that text to be misunderstood. A milieu of mutual immersion in a fully shared design letting co-authors work on the common text while going through others’ comments should be designed in collaborative writing systems. The design of such a milieu should support the relationship between a piece of text and its content. The system Colzlaboracio´ (Rodrı´guez, 2001) for example, creates a shared space to which co-authors can append comments for every section of a document. The commenting feature offered by the system is the core of the communication. Co-authors can make comments in connection with each section, referring to the content of the section and/or to previous comments. Once a comment is appended, it is available to all co-authors. This dialogue is the basis for discussions of plans, coordination, negotiation, and revision focused on that particular section. The contents of the section and the comments made so far on it are visible when the user selects this section. As every section and the comments on it can be explored individually from the whole document, the system helps users to focus on the content of the section in an easy way. Miles et al. (1993) have mentioned the potential of embedding annotations within the shared document, that is, having annotations that point to specific sections of the text. They suggest that instead of having one conversation space per section of the document, it would be useful for co-writers to generate a number of annotation spaces. They could be embedded within a segment’s text and assisting with accurate referential identity. In terms of structure, this type of annotations, which are called topic-specific conversation, could be structured as conversation spaces, allowing sequential, reviewable contributions. The anchored conversations tool presented by Churchill et al. (2000) addresses this issue of coupling conversations and work artifacts when working remotely. The anchored conversations tool allows text-based conversations to be coupled with the document that provides the context for work discussions (e.g. word documents, presentation files or spreadsheets). The model implemented is that of ‘sticky chats’—adhesive chat windows that can be stuck to documents for conversations in context, much as a post-it note can be affixed to a printed document. From the design of features for ‘manipulating others’ paragraphs’ to the design of features for ‘elaborating content together’ As we have seen, it is very hard for distributed groups using a synchronous computersupported collaborative writing system to elaborate content in the way that face-to-face groups do. Unlike face-to-face collaboration, computer mediated collaboration is much less flexible in the sense that for example, phases such as planning and composition, traditionally intertwined in solo writing, have to be distinguished and performed one after the other. We should consider the combination between synchronous and asynchronous modes of interaction for the support of content-elaboration. The prototype SEPIA (Haake and Wilson, 1992) is an example of support for both asynchronous and asynchronous modes of interaction.


T. C. Pargman / Interacting with Computers 15 (2003) 737–757

From the ‘design of a public writing space’ to the ‘design of both private and public place’ for drafting different versions of the same document Distributed groups observed used the synchronous collaborative file as an asynchronous and individual writing space. Contrary to what was expected from a collaborative writing system this transformation facilitated much independent writing and thus a loss of interest in the content of the other’s part. For the distributed groups the common document was the result of a juxtaposition of different pieces of text that the authors were composing and reviewing individually. As indicated by the teacher, final documents produced by the distributed groups presented serious inconsistencies and repetitions, and they included much more restituted than critically treated information. Ellis et al. (1990) have considered the issue of private work in a public space of interaction and suggested a means by which private-user activity information can be signaled. They suggest that private-user editing activity should be visible immediately to the active person, but should be indicated on other users’ screens by the appearance of clouds over the original text. The position and size of a cloud would indicate the approximate location and extent of the modification. According to Miles et al. (1993), collaborative writing tools should allow users to alternate freely between private and public work. In this way users would be able to choose how they communicate through the artifact. As pointed out by Haake and Wilson (1992), designing for cooperation support in a writing tool should not affect an author’s ability to write an individual document. This implies that the cooperation support should not be intrusive when it is not used, and furthermore that the transition from individual authoring to cooperative authoring should be supported in a comfortable and natural way. The issue of private-public spaces is related to the structuring of the document being edited. As Miles et al. (1993) have already pointed out, one might consider allowing users to determine their own document structure instead of imposing one by default. “Support could be given in the form of a graphical representation of the structure they choose. Users would be able to ‘chunk’ a document into segments, which they considered appropriate, as the interaction progressed. Each user-defined segment could be given ‘object’ status. Attributes of such objects could include details of when the segment was edited and by whom” (Miles et al., 1993 p.153). However, as it is comparatively easy to segment and generate tasks, we should be aware of the possibility of providing tools for coordinating and integrating private with public segments. It is the relationship between these aspects that should be anticipated from the beginnings of a collaborative application. As suggested by Cerratto and Rodriguez (2002) collaborative writing tools should allow users to get different views of the document in relation to their interest on the document. During the writing process one author may be more closely tied to certain parts of the document. The document supported by Colzlaboracio´, for example, is a set of sections that can be handled as modules. Users can select the modules (sections) and form a view of the document. While the author uses the system, the index list presented in the left frame always presents the structure of the document. Co-authors can include a new section, modify its content or the structure of the document, and remove any other section. A section and comments on the section is saved in a file of its own. The system handles the sections as modules, which can be used as a filter mechanism both for getting an overview

T. C. Pargman / Interacting with Computers 15 (2003) 737–757


of the document and for creating a version of the document. Co-authors may select a subset of sections, with or without comments, to provide an overview of the common document. This subset does not necessarily have to consist of several contiguous sections. For example, one co-author can be interested only in two sections separated in the structure by other sections, say, the introduction and the conclusion of a document. The system allows the user to have a view of the document that includes only those sections. Finally, the computer program alone never defines an activity, and the function of an artifact is not an external attribute but comes from an activity of attribution of the subject. The design of technologies for collaboration should seriously consider interdependences between individual activities and artifacts, as well as dependencies and independencies of collaborative activities. Supporting collaborative activities through computer systems is thus not only a problem of supporting shared spaces for common activities but rather of supporting and facilitating relationships between co-authors. In other words, supporting relationships between co-authors’ ideas and texts as well as facilitating the relation with oneself as a coauthor.

Acknowledgements The study has been supported by Ex-IRPEACS/CNRS—Ecully, France and University of Paris 8—St Denis. Many thanks to Claire Be´lisle and Romain Zeiliger who provided useful and insightful comments during my PhD studies; to Kerstin Severinson-Eklundh for reading earlier versions of this document and to Daniel Pargman for his valuable ˚ kerstedt who proofread this paper. comments on this paper. Thanks also to Rachel A

References Bannon, L., Bo¨dker, S., 1991. Beyond the interface: encountering artefacts in use. In: Caroll, (Ed.), Designing Interaction: Psychology as the Human Computer Interface, Cambridge University Press, Cambridge, pp. 227–253. Beck, E., 1993. A Survey of Experiences of Collaborative Writing. In: Sharples, M., (Ed.), Computer Supported Collaborative Writing, Springer-Verlag, London, pp. 9–28. Be´guin, P., 1994. Travailler avec la CAO en inge´nierie industrielle: de l’individuel au collectif dans les activite´s avec instruments. Dissertation. CNAM, Paris. Be´guin, P., Rabardel, P., 2000. Designing for instrument-mediated activity. Scandinavian Journal of Information Systems 12, 173–190. Besse, J.M., 1993. L’activite´ conceptualisatrice de l’enfant face a` l’e´crit. Les actes de la Villette. E.d. Nathan, pp. 231–238. Cerratto, T., 1999. Activite´ collaborative sur re´seau. Une approche instrumentale de l’e´criture en collaboration. Dissertation, University of Paris 8-St Denis, Paris. Cerratto, T., 2000. Analyse instrumentale des transformations dans l’e´criture collaborative, suite a` l’utilisation d’un collecticiel. Proceedings of the IC’2000 Conference Inge´nierie des connaissances, Mai 10–12. Toulouse, pp. 299 –310. Cerratto, T., Rodriguez, H., 2002. Studies of Computer Supported Collaborative Writing. Implications for System Design. In M. Blay-Fornarino, A. Pinna-Dery, K. Schmidt, P. Zarate´ (editor), Proceedings of COOP’02Cooperative Systems Design, pages 139–154, IOS Press, Amsterdam, 2002.


T. C. Pargman / Interacting with Computers 15 (2003) 737–757

Churchill, E., Trevor, J., Bly, S., Nelson, L., Cubranic, D., 2000. Anchored Conversations. Chatting in the Context of a Document. CHI 2000 Conference Proceedings, ACM Press, pp. 454–461. Dillon, A., 1993. How collaborative is collaborative writing? An analysis of the production of two technical reports. In: Sharples, M., (Ed.), Computer Supported Collaborative Writing, Springer-Verlag, London, pp. 69 –85. Dillon, A., 1994. How Collaborative Is Collaborative Writing? Ergonomic Aspects of Human Information Usage, Taylor and Francis, London. Ede, l., Lundsford, A., 1990. Singular Texts/Plural Authors: Perspectives on Collaborative Writing. Southern Illinois University Press, Carbondale and Edwardsville. Ellis, C.A., Gibbs, S.J., Rein, G., 1990. Design and Use of a Group Editor. In: Cockton, G., (Ed.), Proceedings of the IFIP Engineering for Human-Computer Interaction Conference, North Holland, Amsterdam, pp. 13–25. Engestro¨m, Y., 1987. Learning by Expanding: An Activity—Theoretical Approach to Developmental Research. Orienta-Konsultit oy, Helsinki. Engestro¨m, Y., 1990. Learning, Working and Imagining. Twelve Studies in Activity Theory. Orienta-Konsultit Oy Helsinki. Engestro¨m, Y., 1999. Activity theory and individual and social transformation. In: Engestro¨m, Y., Miettinen, R., Punama¨ki, R.-l. (Eds.), Perspectives on Activity Theory, Cambridge University Press, Cambridge, pp. 19–38. Flower, L., Hayes, J., 1981. A cognitive process theory of writing. College Composition and Communication 32 (1), 365–387. Goody, J., 1994. Entre l’oralite´ et l’e´criture. PUF, Paris. Grudin, J., 1994. CSCW: history and focus. IEEE Computer 27 (5), 19–26. Haake, J., Wilson, B., 1992. Supporting collaborative writing of hyper documents in SEPIA. In: Turner, J., Kraut, R. (Eds.), Proceedings CSCW’92. Sharing Perspectives, ACM press, Toronto, pp. 138 –146. Haas, C., 1996. Writing Technology: Studies on the Materiality of Literacy, Lawrence Erlbaum Associates, Mahwah, New Jersey. Heath, C., Luff, P., Sellen, A., 1995. Reconsidering the Virtual Workplace: Flexible Support for Collaborative Activity. In: Marmolin, H., Sundbland, Y., Schmidt, K. (Eds.), Proceedings of the Fourth European Conference on CSCW, Kluwer, Stockholm, pp. 83 –99. Hutchins, E., 1995. Cognition in the Wild, MIT Press, Cambridge, MA. Kaptelinin, V., 1996. Activity theory: implications for human-computer interaction. In: Nardi, B., (Ed.), Activity Theory and Human– Computer Interaction, MIT Press, Boston, pp. 103–116. Kerbrat-Orecchioni, C., 1992. Les interactions verbales, Tome II, Armand Colin. Paris. Kirby, A., Rodden, T., 1995. Contact: Support for Distributed Cooperative Writing. In: Marmolin, H., Sundbland, Y., Schmidt, K. (Eds.), Proceedings of the Fourth European Conference on CSCW, Kluwer, Stockholm, pp. 101–116. Kraut, R., Galegher, J., Egido, C., 1988. Relationships and Tasks in Scientific Research Collaboration. Human Computer Interaction, Vol. 3, Number 1. LEA, Hillsdale, New Jersey, pp. 31 –58. Kraut, R., Galegher, J. Fish, R., Chalfonte, B., 1992. Task Requirements and Media Choice in Collaborative Writing. Human Computer Interaction, Vol. 7 number 4, LEA, Hillsdale, New Jersey, pp. 375 –408. Kuutti, K., 1996. Activity theory as a potential framework for human–computer interaction. In: Nardi, B., (Ed.), Activity Theory and Human–Computer Interaction, MIT Press, Boston, pp. 17 –44. Leplat, J., 1993. L’analyse psychologique du travail. Quelques jalons historiques. Le travail Humain 56 (2–3), 115–131. Luff, P., Health, C., Greatbatch, D., 1992. Tasks-in-interaction: Paper and Screen Based Documentation in Collaborative Activity. Proceedings of the Conference on Computer Supported Cooperative Work, Toronto, pp. 163 –170. Miles, V., McCarthy, J., Dix, A., Harrison, M., Monk, A., 1993. Reviewing designs for a synchronous– asynchronous group editing environment. In: Sharples, M., (Ed.), Computer Supported Collaborative Writing, Springer-Verlag, London, pp. 137–160. Neuwirth, C., Kaufer, D., Chandhok, R., Morris, J., 1990. Issues in the Design of Computer Support for Coauthoring and Commenting, Proceedings of the Third Conference on CSCW’90, ACM Press, Baltimore, MD, pp. 183 –195. Newman, R., Newman, J., 1993. Social writing: premises and practices in computerized contexts. In: Sharples, M., (Ed.), Computer Supported Collaborative Writing, Springer-Verlag, London, pp. 29–40.

T. C. Pargman / Interacting with Computers 15 (2003) 737–757


Ong, M., 1982. Orality and Literacy, The Technologizing of the Word, Routledge, London. Plowman, L., 1993. Tracing the evolution of a co-authored text. Language and Communication 13 (3), 149–161. Plowman, L., 1995. The interfunctionality of talk and text. In: Bruce, B., et Sharples, M. (Eds.), Computer Supported Cooperative Work, Special Issue on Computer Supported Collaborative Writing, pp. 229–246. Posner, I., Baecker, R., 1992. How people write together. Proceedings 25th Hawaii International Conference on System Sciences IV, 127 –138. Rabardel P., 1995. Les hommes et les technologies: Approche cognitive des instruments contemporains. Paris, Colin. Rodrı´guez, H., 2001. Using the WWW as Infrastructure for Collaborative Production of Documents. Licentiate thesis. The Royal Institute of Technology, Stockholm. Scardmalia, M., Bereiter, C., 1987. Knowledge telling and knowledge transforming in written composition. In: Rosenberg, (Ed.), Reading, Writing and Language Learning, Advanced in Applied Psycho-linguistics, vol. 2. Cambridge University Press, Cambridge, pp. 142–175. Schmidt, K., Bannon, L., 1992. Taking CSCW seriously: supporting articulation work. Computer Supported Cooperative Work 1 (1/2), 7– 40. Schmidt, K., Simone, C., 1996. Coordination mechanisms: towards a conceptual foundation of CSCW systems design, Computer Supported Cooperative Work: The Journal of Collaborating Computing number 5, Kluwer Academic Publishers, Netherlands, pp. 155–200. Schrage, M., 1995. No More Teams! Mastering the Dynamics of Creative Collaboration. Currency Doubleday, New York. Severinson-Eklundh, K., 1986. Dialogue Processes in Computer-Mediated Communication. A Study of Letters in the COM System. Doctoral dissertation. University of Linko¨ping, Linko¨ping. Sharples, M., 1993. Introduction. In: Sharples, M., (Ed.), Computer Supported Collaborative Writing, SpringerVerlag, London, pp. 1–7. Sharples, M., 1996. An account of writing as creative design. In: Levy, C.M., Ransdell, S. (Eds.), The Science of Writing, Lawrence Erlbaum, Hillsdale, NJ. Sharples, M., Pemberton, L., 1988. Representing Writing: An Account of the Writing Process in Regard to the Writer’s External Representations. In Cognitive Science Research paper CSRP nro. 119. School of Cognitive and Computing Sciences University of Sussex, UK. Suchman, L., 1987. Plans and Situated Actions: The Problem of Human–Machine Interaction, Cambridge University Press, Cambridge. Ve´rillon, P., Rabardel, P., 1995. Cognition and artifacts: a contribution to the study of thought in relation to instrumented activity. European Journal of Psychology of Education X (1), 77–81. Vygotsky, L., 1934/1997. Pensee´ et Langage. La dispute. Translated by F. Se`ve.