Collaborative Writing as a Process of Formalizing Group

Proceedings of the 28th Annual Hawaii International Conference on System Sciences - 1995
Collaborative Writing as a Process of Formalizing Group Memory’
Munir Mandviwalla
Stanley D. Clark II
Kent Sandoe
Temple University
Computer and Information Sciences
Claremont Graduate School
Programs in Information Science
Abstract
that emphasizesmedia effects and representation; and a
A common form of collective memory is the groupauthored document.
Examining collaborative writing
from the group memory perspective may yield important
design and theoretical insights. This paper views
collaborative writing as a cognitive process in which a
portion of group memory is formalized. This view is the
basis for expanding an existing model of individual
writing to include colla.borative writing and group
memory. A collaborative writing tool is analyzed with the
design implications from the expanded model. Zbe
analysis suggests that a group memory view of
collaborative writing contributes useful insights that can
guide the design of computer-based ‘colkborative writing
tools.
group focus that emphasizesroles, interdependence,and
managementof conflict.
This paper will explore an alternative perspective in
which collaborative writing is viewed as a cognitive
process in which a portion of group memory is
formalized. It is our contention that collaborative
documents emerge from the process of formalizing
abstract and concrete memories to a concrete
representation -- the document. A collaborative
documentat the most basic level is created by some type
of collective thinking process. Examining the formation
of a collective memory can provide useful insights into
this thinking process.
2. Previous work on collaborative writing
1. Introduction
There are several empirical studies of the
collaborative writing process [lo, 18, 3, 21. These
studies are useful for understanding the activities that
occur in the collaborative writing act. However, a
cognitive accountof the underlying processesinvolved in
collaborative writing is missing.
A number of experimental and commercial systems
support collaborative writing (e.g., SASSE [l], PREP
[ 171, GROVE [41, Quilt [ 111, ForComment, Markup,
and Aspects). Some are full featured multi-user writing
systems, while others focus on specific aspects of
collaborative writing such as annotation. All are selfcontained systems that concentrate on supporting
interaction among collaborators. Sharpleset al. [20] in a
summary of collaborative writing research issues,
identify partitioning strategies, interleaving of tasks,
interdependenceamong group members, roles, conflict
management, communication and media effects, and
constraints, as areasof previous and future research.
In today’s increasingly team-based organizations,
team work often culminates in the production of reports
that reflect the contributions of their various members.
Such contributions may take the form of independently
written sections or may stem from direct collaboration
on all or portions of these documents. Attempting to
reconcile the words and ideas of individuals with the
intent of the group is a difficult and daunting task.
Computer-basedcollaborative writing tools may serve an
important role in helping teams successfully complete
their collaborative writing tasks.
Over the past decade a rich body of empirical and
developmental work has emerged on computer-based
systems that support collaborative writing [ 10,18,20].
Previous work has typically viewed collaborative writing
from one or a combination of the following perspectives:
a task focus in which the production of a joint report is
the root of analysis and typically addressesissues of
partitioning and coordination; a communication focus
’Author contact information: Munir Mandviwalla: [email protected].
Stan Clark: clarksQcgs.edu.Kent Sandoe:[email protected]
342
1060-3425/95 $4.00 0 1995 IEEE
Proceedings of the 28th Hawaii International Conference on System Sciences (HICSS '95)
1060-3425/95 $10.00 © 1995 IEEE
Proceedings
of the 28th Annual Hawaii
International
THE
r.RITER’
LONG-TERM
S
MEMORY
Knawladge
0,Topic,
Conference
on System Sciences -
V.FtITING PROCESS
1995
I
Audience.
at-d wimp
Plm!J
11
Figure 1: Individual Writing Model ([5])
In the following sectionsof the paper we develop a
cognitive description of collaborative writing as a
process in which group memory is formalized. We
introduce a model of individual writing from the
collaborative perspective. An expanded model of
collaborative writing is proposed where writing is
viewed as the processof formalizing group memory. The
model and requirements from previous studies of
collaborative writing are used to describe and then
analyze the design of a prototype tool called
GroupWord.
writing, individuals accesstheir long term memory for
topic, audience, and writing plan knowledge. The
componentsof the model represent elementary mental
processesthat may be performed at any time; they do not
representstagesof linear activity. The processeshave a
hierarchicalstructurein which any given processmay be
embeddedin anotherprocess. A “monitor” governs the
writer’s movementfrom one processto another. Major
aspects of the planning process include goal-setting,
generatingideas, and organizing. The writing is directed
by a network of hierarchically related goals.
3. Importance of individual writing
5. Understanding collaborative writing
Studies show that individual work is integral to
collaborative writing. Kraut et al., [lo] report that
authorsdivide responsibilityfor tasksamong themselves;
sequencingtheir work so that a partner remains passive
until they receive a completed task. In their study,
individuals work on assignedportions of a project, meet
frequently to discussideas and share feedback,and rely
on simple mechanismsto facilitate collaboration. The
proportion of joint work to independent work is
moderatedby the type of task, phaseof the project, and
channelsavailable to the group [6]. Posnerand Baecker
[18] define single writer, scribe, separate writers, joint
writing, and consulted (i.e., outside consultant) as
collaborativewriting strategies. The first three strategies
emphasizeindividual work.
5.1. The underlying forces
According to Flower and Hayes [5], there are three
forces that serve as critical inputs to the individual
writing process. These forces also play a role in
collaborativewriting:
l
the text producedso far,
l
individual long term memories and stored
knowledge, and
l
the task.
Text-.
Each additional word and
sentencein a growing document constrainsthe text that
comesnext. In other words, the documentis a unfolding
concrete-representationof memory and the initial
scaffolding in this memory constrainsthe structuresthat
come next. For example, a sentencecreated by one
writer establishesa concretememory that influences the
sentenceconstruction of another writer becauseof the
need to maintain conceptual and stylistic consistency.
The text produced in the document also representsthe
formalization of the various memories involved in the
collaborativewriting task.
4. Individual writing model
Flower and Hayes [5] present a model of the
cognitive processunderlying the individual writing task
(see Figure 1). The model is based on a detailed
protocol analysis of the writing processand consistsof
During the
planning, translating, and reviewing.
343
Proceedings of the 28th Hawaii International Conference on System Sciences (HICSS '95)
1060-3425/95 $10.00 © 1995 IEEE
Proceedings of the 28th Annual Hawaii International Conference on System Sciences - 1995
into a document. Where formalization is the key process
of creating a concrete group memory. Examples include
establishmentof shared goals on procedures (“lets start
with a discussionof problem”) and content (“we should
put in somethingabout our methodology”), grouping and
organizing together ideas from different writers, and the
actual generationof ideas and text.
Translatinp. This is the process of creating text to
represent ideas. In the individual writing process,
writers are sometimes overwhelmed by the amount of
information that they need to juggle in short term
memory. The collaborative situation may place greater
demandson short term memory given that writers will
be faced with having to translate and understand ideas
from other writers.
Reviewing. This processoccurs as a planned activity
or is triggered by an unplanned evaluation of written
text. In collaborative writing, the reviewing process is
where coordination problems are most likely to surface.
Revising a group memory is a fertile ground for conflict
such as when one writer’s revisions of another’s work
are viewed as a rejection of the original writer’s
interpretation of the collective memory. One way to
understandthis problem is to consider the act of revising
as an exercise in interpreting the memories embeddedin
the document.
&IQ&X. The monitor strategizesthe above processes
in terms of sequencingand time spent. Such activities
may need to be negotiatedin the collaborative case. For
example, one writer triggered by their internal monitor
may want to move on to formal review while another
may desire to keep writing. One indication of successful
writing is returning to high level goals. In collaborative
writing this process may be aided by the collective
memory of the group that remembersgoals and reminds
other members of those goals. However, the collective
memory may be so large and fragmented that nothing of
value can be contributed. Thus, providing for a useful
group memory is an important issue in designing a
collaborative writing tool.
Each writer will bring a different set of individual
memories to the collaborative writing task. These
memories will come from printed materials such as a
reference book and memories stored inside a writer’s
brain cells.
One problem in the individual writing task is
accessingthis memory to serve the current context. In
the collaborative writing situation fragments of one
writer’s memories have the potential of triggering and
mixing with the memories of other writers. For
example, a sentenceor spoken utterance by one writer
may trigger: concrete memories in a secondwriter which
may involve either factual information or inferences
about the first writer’s orientation; or abstract memories
such as conceptualizingthe point being made or thoughts
about the cultural backgroundof the issue.
Thetask. Some groups may see documentcreation as
a primary task, others may see it as a secondaryresult of
their main mission. In terms of timing, a group may be
engagedin collaborative writing on a continuous or ongoing basis from its inception forward, or it may attempt
to document its activities in a post-hoc or after-the-fact
fashion. Arrayed in their simplest terms, the dimensions
of role (primary and secondary) and timing (continuous
and post-hoc) provide four basic modes of interaction.
These modes of interaction place different constraintson
the writing process. For example, a secondaryrole and
post-hoc scenario may require fewer revisions and a
lower intensity of interaction during the writing process.
How the task is stored in memory may also influence
the writing. For example, one writer may interpret the
task based on the language or wording of the problem
statement. Others may use the social structure
surrounding the problem statement to understand what
the “real task is.”
5.2. The formalization
process
Writing also involves four basic processes:planning,
reviewing, and monitoring [5]. The
translating,
processes are embedded in each other in a recursive
iterative manner. Each process is described next in the
context of collaborative writing.
planninp. The term planning refers to the formation
of an internal representation of the writing related
knowledge. This internal planning knowledge includes
abstract images that will eventually lead to the prose in
the document.
Planning in a collaborative situation may also be
understood as the process of creating a group memory
where abstract and concrete knowledge is formalized
6. Model of collaborative writing
In the previous section, the individual writing model
of Flower and Hayes [5] was expanded to describe
collaborative writing issues. This section integrates these
issues into a group memory-oriented model of
collaborative writing; an integrated model is useful for
generating research hypothesesand design implications.
It should be emphasizedthat this is a first-cut working
model that requires refinement through empirical testing.
344
Proceedings of the 28th Hawaii International Conference on System Sciences (HICSS '95)
1060-3425/95 $10.00 © 1995 IEEE
Proceedings of the 28th Annual Hawaii International Conference on System Sciences - 1995
INDIVIDUAL
ACTIVITY
GROUP
MEMORY
planning,
coordinating
rules, procedures
norms, roles
norming, defining
I
I
ORGANIZATIONAL
/ EXTRA-ORGANIZATIONAL
Figure 2: A group memory
model of collaborative
CONTEXT
writing
document segments, and still others to review and edit
the material.
. .
&m&kng.
This is the act of collaborative writing
in which the forming and framing activities are
crystallized into text. Formalizing also includes revising
and editing. From the group memory perspective, these
activities also involve formulating abstract memories to
concretetext2.
Each of these processesare continuously active and
may interrupt another process. They are organized in a
hierarchical nestedmanner. For example, a formalizing
process may contain several embedded framing
processes.
The form of collective memory has been describedas
ranging from the abstract, informal and intangible to the
concrete, formal, and tangible [19]. To the extent that a
collaborative document reflects the memory of its
writers, it is likely to be located on the latter part of this
After all, documents themselves are
continuum.
concrete, tangible, and formed through the articulation of
highly formal symbols. However, the abstract memories
and interpretations of the concrete memories play a
critical part in the collaborative writing process.
Therefore, a group memory view of collaborative
writing implies the need to attend to all the various forms
of memory; not just the concrete representation. The
model in Figure 2 shows three processesthat reflect this
need. The processesare continuously active during the
collaborative writing task:
Forming. All groups begin making memory
immediately and automatically as soon as their members
start to interact. Principally, this occurs through the
establishment of norms and the definition of roles.
Forming is the construct in which some of the constraints
on the writing process(long term memories and the task)
combine together to establish a contextualizednon-verbal
working area in memory for framing and formalizing
activities.
Framinp. Refers to activities that include the creation
of document preparation rules and procedures. One
important activity is the role definition. For example,
one individual is often assigned to coordinate and
monitor the writing process, others to produce specific
7. Design implications
The group memory model of collaborative writing
presents interesting design implications for developing
computer-basedsupport for collaborative writing:
Entanplement. When writers interact through
sentencesor other media, they transform and absorb
memoriesof other writers. The fidelity with which these
memories are transferred may cause problems.
Collaborative writing may create so many conflicting
memories that writers find it hard to filter and forget
material and goals that should be discarded. Writers
* We also use the term “formalizing” in this paper to refer to the
overall process of creating a group document because the actual
writing is the critical step in collaborative writing.
345
Proceedings of the 28th Hawaii International Conference on System Sciences (HICSS '95)
1060-3425/95 $10.00 © 1995 IEEE
Proceedings
of the 28th Annual Hawaii
International
may end up spending all their time with the framing
process to reconcile different backgrounds and not
allocate enough time on the important process of
formalizing.
. .
u.
Conference
on System Sciences -
1995
Windows such as redlining, annotation, and document
comparison. GroupWord was designedto work with the
CGSE meta environment [14]. The CGSE meta
environment is a groupware shell through which users
select the features and applications they need from a
menu of options. GroupWord may be used with other
environmentssuch as Lotus Notes. The overall strategy
is to use the meta environment for general collaboration
and GroupWord for document-basedcollaboration.
It will be difficult to synchronize
the forming, framing, and formalizing processesamong
writers. Writers fluidly move from thinking about some
memories such as high level goals (forming -- framing)
to other memories such as plans for revision (framing -formalizing).
. .
m.
The act of producing a collective
memory can produce different interpretations of the
collective memory. As noted earlier this can be a fertile
ground for conflict and negotiation.
In the remainder of the paper we apply the above
insights to analyze the design of computer-based
collaborative writing tools. Such an analysis can elicit
design options that have been neglectedor overlooked in
the past. The context of our analysis is the critical
evaluation of an existing collaborative writing tool -GroupWord.
Requirements
ndividual Writing
8. GroupWord
GroupWord (I[151, [ 131)is basedon the premisethat a
major component of any writing is the individual
working in front of a word-processor. Creating a
collaborative writing tool that is separatedfrom the tool
used for individual writing can lead to problems in
training becauseusers will have to learn two packages;
usage becausethe interface of the collaborative writing
packagewill probably be different from and less familiar
than a word processor; and features because the
collaborative writing tool will probably lack the features
of a commercialword processingpackage.
Recent efforts such as SEPIA [S] explicitly
incorporate individual work support. However, SEPIA
only supports the process of doing work individually; it
does not provide the large feature set available in
commercialword processors. GroupWriter [ 121is based
on a commercialword processingengine and is interface
compatible with Microsoft Word. This approach while
addressing some of the problems listed above, still
resultsin a separateself-containedproduct.
GroupWord
Basicword-processing
Seamlessness
with other media
Collaborative
Writing
++
Preserveidentities
Enhancecommunication
Enhancecollaborator awareness
Focusedcollaboration
Peripheralawareness
Annotations
Undo
Sessioncontrol
loles
Explicit roles
Lctivities
Variety of activities
Brainstorming
Researching
Planning (outline)
Planning (process)
Writing
Editing
Reviewing
Transitions betweenactivities.
documentControl Methods
Severalaccessmethods
Separatedocumentsegments
Version and changecontrol
Lkiting Strategies
One or several writers
Synchronouswriting
Asynchronouswriting
+
+*
Legena:
+ + gooa,
++
++
++
*
*
++
*
++
++
++
+ -t
++
++
++
++
+
+
+ some, - none, - meta enwronment
Table 1: Design requirements
and GroupWord
The three main distinguishing characteristics of
GroupWord, as compared to other collaborative writing
tools, are:
l
seamlessness
with individual writing,
l
comprehensiveindividual writing features and,
l
the assumptionthat a collaborative writing tool will
only support document-specific activities leaving
other areas of group support such as guided
brainstorming to other groupware applications
available from the meta environment.
GroupWord is built “on top of” Microsoft Word for
Windows and Microsoft Access3. It is programmedwith
the WordBasic macro language and takes advantageof
built-in collaborative writing features in Word for
3 GroupWord was developed in Word for Windows l.lA and has now
been upgraded to version 6.OA. This newest version of GroupWord
has been enhanced to use Access 2.0 for recording historical
information.
346
Proceedings of the 28th Hawaii International Conference on System Sciences (HICSS '95)
1060-3425/95 $10.00 © 1995 IEEE
Proceedings
of the 28th Annual Hawaii
International
Most of the other characteristics of GroupWord
follow a list of empirically derived collaborativewriting
requirements provided by Posner and Baecker and
associates[18, 11. Table 1 subjectivelyrates GroupWord
on each requirement. See Appendix I for a detailed
descriptionof GroupWord.
Conference
on System Sciences -
1995
in historical sequencewhere the writer can see what the
paragraphlooked like two weeks ago, or what it looked
like when it was modified by one writer two weeks ago
without any of the changesmade by another writer. In a
future version, it may also be useful to provide statistical
information about changes to the document. For
example, Hill et al. [9] use a scroll bar to visually depict
the number of times a sectionhas been read and edited.
Another problem of entanglementis differentiating a
personal contribution from the group document.
Individuals may need to focus solely on the text that they
have produced and go through a whole sequence of
forming, framing, and formalizing processesbefore they
fully formalize a single thought. If these thoughts are
representedtoo early in the group document then they
may entangle with other writers thoughts. GroupWord
addressesthis issue by emphasizingindividual work so
that it is possible for users to work on their own on a
section.Individual work support is also available in other
systems[e.g., 8.11.
At a higher level, outlining features in GroupWord
can delineate different sections of the document. This
feature is useful for addressingproblems of entanglement
at the topic, as opposed to the paragraph and sentence
constructionlevel.
. .
Svnchronu;atlon. Most current collaborative writing
tools synchronize access to the formalized document
through various strategies. For example, in GroupWord
write accessto the document is coordinatedto an event:
the releaseof a section. A potential problem in this or
any other scheme is that the type of accessmay be
incompatiblewith a particular writer’s style of working.
In a computer-basedenvironment it is important to
establishsometype of structuredmethodologyfor shared
file access;whether it is implemented at the sentence,
paragraph, section, or document level is dependenton
the flexibility and power of the underlying computing
environment.
Our model of collaborative writing suggeststhat the
processof formalizing a group document is an iterative
recursiveprocess. A writer may formalize one sentence
in one part of the document, but this may trigger some
framing activities that influence another part of the
document, that may in turn trigger the revision of a
paragraph in another part of the document. Computerbasedsynchronizationof collaborative writing will have
to balance the need to provide a flexible mode of
working and at the same time guarantee structured and
safe accessto the main group document.
.
Annotation and revision marking
features in GroupWord and other tools can support the
recording of a negotiation process. The negotiation
information must be easily accessiblein the collaborative
9. GroupWord and group memory
Any computer-basedwriting tool can supportthe final
creation of a concrete representationof group memory.
However, collaborative writing and group memory
creation is not a stage-basedprocessthat ends with a
comprehensive“memory dump” to paper. Research
indicates that writing with others is a complex process
where formalization iteratively occurs in concert with
other activities. There are several obstacles to the
formalization process that may benefit from computer
support. We revisit the design implicationspresentedin
section7 to outline potential obstaclesand corresponding
solutions. The completeness,accessibility,and relevance
criteria identified by Sandoe et al. [19] for evaluating
group memory are used as part of the analysis.
Entanglement. Identifying annotationsby author can
ensurethat there is no confusion about who addedwhat.
However, annotations may be deleted and it may be
difficult to disentangle the underlying meaning of
annotations scattered all over the document. For
example, to accessannotationsa writer has to follow the
standard linear format of a document. There may be
other more useful views of this information.
From the perspectiveof entanglementit would also be
useful to keep track of revision marks including the
original text. Current collaborative writing tools
including GroupWord address this aspect of
entanglementby color coding the new text. However,
the color coding approach may be incomplete because
there are few ways to easily accessthis information. In
addition, a large number of revision and annotation
marksjumbled together in one paragraphis confusing. In
other words, it is not enough to have a complete
recording of all the relevant memories; the memories
also have to be accessible[ 191.
Therefore, the latest implementation of GroupWord
also stores a permanent, easily retrievable record of all
annotations and revised paragraphs in an database
stamped with author, date, and time information. The
databaseallows different views of this information. For
example, view changes by historical sequence, by
author, or grouped by sub-document. Seeing all the
changesmade by a particular author in a single screen
will help other authors learn (or revisit) the thinking
underlying the changes. Paragraphscan also be viewed
347
Proceedings of the 28th Hawaii International Conference on System Sciences (HICSS '95)
1060-3425/95 $10.00 © 1995 IEEE
Proceedings of the 28th Annual Hawaii International Conference on System Sciences - 1995
6. Galegher, J. & Kraut, R. Computer-Mediated
Communication and Collaborative Writing: Media
writing tool (or environment). Decision making tools are
also needed becausea simple record of the negotiating
processdoes not help with the actual negotiation. Since
negotiation in writing centers around text, these tools
will have to operate within the medium of a document.
For example, it may be useful to have a voting tool popup over the document, automatically collect the text
under contention, and then present all the proposed
revisions to that text to other writers as voting choices.
At present, GroupWord does not support this feature.
There may be a short window of opportunity in the
document creation process where such negotiation
features will be used; unless the issue is very important,
writers will tend to forget the actual context of the
disagreement.
Influenceand Adaptationto CommunicativeConstraints.
Proceedings of CSCW’92, 155-162.
7. Greenberg,S. Personalizable
Groupware:Accommodating
Individual Roles and Group Differences. Proceedings of
European CSCW’91,
17-31.
8. Haake, J. M. & Wilson, B. Supporting Collaborative
Writing of Hyperdocurnentsin SEPIA. Proceedings of
CSCW’92, 138-146.
9. Hill, W., Hollan, J., Wroblewski, D., & McCandless, T.
Edit Wear and Read Wear. Proceedings of CHI’92, 3-9.
10.Kraut, R., GalegherI., & Egido, C. Relationshipsand
Tasks in Scientific Research Collaborations. Greif, I.
(Ed.) Computer Supported Cooperative Work: A book of
Readings, California: Morgan Kaufmann, 1988.
11. Leland, M., Fish, R., & Kraut, R. Collaborative
Document Production using Quilt. Proceedings of
CSCW’88, 206-215.
12. Malcolm, N. & Gaines,B. R. A Minimalist Approachto
the Development of a Word Processor Supporting Group
Writing Activities. Proceedings of COCS’91, 147-152.
13. Mandviwalla, M. The Design of Group Support astems.
Unpublished Ph.D. dissertation, Programs in Information
Science, Claremont GraduateSchool, January, 1994.
14. Mandviwalla, M., Gray, P., Olfman, I.., & Satzinger, J.
The Claremont GDSS Support Environment. Proceedings
of HICSS’91, III, 600-607.
15. Mandviwalla, M. & Clark, S. The Importance of
Individual Work in Collaborative Writing. Working Paper.
Temple University, September1993.
16. Miles, V., McCarthy, J., Dix, A., Harrison, M., & Monk,
10. Conclusion
There is still a great deal of researchneededon what
group memory is, the process of collaborative writing,
and how the two interact. In this paper, we have
expandedan existing and relatively well-supportedmodel
of individual writing to collaborative writing using group
memory as the main construct. A recurring theme in our
analysis is that collaborative writing involves important
memories that support the process of creating a final
document. These processmeta-memorieswhich involve
negotiations, synchronization steps, and strategies to
avoid entanglementare typically forgotten and lost when
the group has finished the task of documentconstruction.
However, capturing met&memories and providing
computer-based tools to act upon them during the
collaborative writing process may assist groups in the
goal of creating better documents.
A. ReviewingDesignfor a Synchronous-- Asynchronous
Group Editing Environment. In Sharples, M. (Ed.)
Computer Supported Collaborative
Writing. London:
Springer-Verlag, 1993, 137-160.
17. Neuwirth, C., Kaufer, D., Chandhok, R., & Morris, J.
Issues in the Design of Computer Support for Co-writing
and Commenting. Proceedings of CSCW’90, 183-195.
18. Posner, I. R., & Baecker, R. M. How People Write
Together. Proceedings of HZCSS’92, IV, 127-138.
19. Sandoe, K., Olfinan, L., & Mandviwalla, M. Meeting in
Time:
Recording the Workgroup Conversation.
Proceedings of ICIS’91, 26 1-271.
20. Sharples, M., Goodlet, J. S., Beck, E. E., Wood, C. C.,
Easterbrook, S. M., & Plowman, L. ResearchIssues in the
Study of Computer Supported Collaborative Writing. In
Sharples, M. (Ed.) Computer Supported Collaborative
Writing. London: Springer-Verlag. 1993, 9-28.
References
1. Baecker, R. M., Nastos, D., Posner, I. R., & Mawby, K.
L. The User-centered Iterative Design of Collaborative
Writing Software. Proceedings of IhTERCHZ’ 93, 399-405.
2. Beck, E. E., A Survey of Experiences of Collaborative
In Sharples, M. (ed.) Computer Supported
Writing.
Collaborative Writing. Springer-Verlag, London, 1993, pp
87-112.
3. Dillon, A. 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, 1993, pp 69-85.
4. Ellis, C., Gibbs, S., & Rein, G. Groupware: Some Issues
and Experiences. Communications of the ACM, 34 (1),
1991, 39-58.
5. Flower, L. S., & Hayes, F. R. A cognitive processtheory
of writing. College Composition and Communication, 32,
1981, 365-387.
348
Proceedings of the 28th Hawaii International Conference on System Sciences (HICSS '95)
1060-3425/95 $10.00 © 1995 IEEE
Proceedings of the 28th Annual Hawaii International Conference on System Sciences -
1995
different requirementsthan a person that contributes text.
GroupWord supports different roles such as writing,
editing, commenting, or simply viewing various sections.
Unlike systems such as Quilt [ll], roles are not
explicitly assignedin GroupWord. The user may change
roles by picking the appropriate command from the
menu.
.
Multiple users access the
compositedocumentin read-only mode while working in
individual sections. Sectionsmay be as large or small as
necessaryfor the functioning of the group. Users may
only edit a section of the document at a time. To edit a
section of the group document a user places the cursor in
the area of the document to be edited and then pulls
down the Edit menu shown in Figure 2. By clicking on
Edit section.. . users can edit the section in a separate
sub-section window if they have the appropriate
authority. The menus in this window allow accessto a
selectedsubsetof Word for Window’s standardfeatures.
In this menu, users can at any point select Save for
Group (see Figure 1B) to save their changesand update
the group document. GroupWord supports an “explicit
release’floor control schemein which the person editing
a section must explicitly release it (by closing the subdocument) before it is available to another user for
editing [7]. Other members can see additions and
changesto the main group document by updating their
view.
Appendix I -- Description of GroupWord
GroupWord will operate on any Windows 3.X
compatible personal computer and network. It requires
Microsoft Word for Windows 6.OA and, optionally for
revision and annotation history, Microsoft Access 2.0.
A copy of the application resides on each workstation
while the documents are stored on the central server.
The GroupWord architecture uses a multi-document
design. The group document is a composite document
that consistsof sub-documents(sections). The composite
document has pointers (or links) which reference the
sub-documents. The system automatically creates
individual views of documents based on user and
terminal information. The documents and sections are
maintained in a manner that is transparent to everyone
except the documentcoordinator.
S&~QJDMenu Structure. GroupWord consists of a
set of custom menus along with associatedroutines that
complement the standard Word for Windows interface.
Through this interface a team member has an
independentview of the composite group document and
can edit individual sections. Figures 1A and 1B show
the composite and section views. The user can go back
and forth to the normal Word for Windows menus by
selecting Restoreor GroupWord from the Utilities menu.
Edit section...
Bedline section...
Get Updated View.
Figure IB: File Menu in Document
Figure 2: Menus in Composite
. .
Section
Document
m.
When a section of the
group document is locked for annotation, only the
assignedowner can edit the section, while others have
annotation capability. Users select the portion of the
group document that they wish to annotateand then click
on Annotate section... from the Edit menu (see Figure
2). Within the sub-documentusers follow the normal
annotation procedures of Word for Windows. The
systemautomatically inserts a user’s initials beside every
annotation. To remain anonymous, users can select
Annotate Anonymously to insert ‘ANON’ instead of
d asynchronous access. Team
members can work on the group document at any time
assuming they have the appropriate access.
Simultaneousaccess is available whether it is from the
same location -- such as a conference room where the
team is gathered for a meeting, or from distributed
locations -- such as individuals in their offices conversing
on the phone.
&les. Team members have different roles in the
collaboration process. For example, a reviewer has
349
Proceedings of the 28th Hawaii International Conference on System Sciences (HICSS '95)
1060-3425/95 $10.00 © 1995 IEEE
Proceedings of the 28th Annual Hawaii international Conference on System Sciences - 1995
their initials. Redlining is accessedin a manner similar
to annotating. Additions and revisions by individual
team members are marked so that deleted text is not
erased but merely crossed out, new text is underlined,
and change indicators in the margin show where the text
has been changed.
Views of the w.
As with reports
run against a database,a user’s view of the composite
document is current as of the moment it was obtained,
but the view does not automatically change as the
underlying document information changes.
Users get a new view every time they enter the
composite document or update a sub-document. At any
time users can go to the View menu (see Figure 2) and
select Get Updated View.
. .
.
e
The collaborative writing process is
started with a designated document coordinator who
generatesthe group document by importing documents,
outlining the document, or starting with a blank slate.
An underlying assumption is that collaborative writing
takes place in a cooperative environment where free flow
of information between members of the group is the
expected norm. Therefore, all members of the group
can view all parts of the document.
350
Proceedings of the 28th Hawaii International Conference on System Sciences (HICSS '95)
1060-3425/95 $10.00 © 1995 IEEE