A NARRATIVE APPROACH TO COLLABORATIVE WRITING

A Business Process Model

Peter Henderson, Nishadi De Silva

School of Electronics and Computer Science, University of Southampton, Southampton, UK

Keywords: Collaborative writing, Business Processes, Narratives, Rhetorical Structure Theory (RST), CSCW.

Abstract: Narratives have been used in the past to enhance technical documents such as research proposals by

implementing a single-user writing tool called CANS (Computer-Aided Narrative Support). This study has

now been extended to collaborative writing (CW); another area that can greatly benefit from a narrative-

based writing tool. Before implementing such an asynchronous, multi-user system, however, it was

imperative to do a concrete design for it. Therefore, after studying existing CW tools and strategies, a

concise business process (BP) model was designed to describe the process of narrative-based CW. This

paper introduces narrative-based CW for technical authors, the BP model for it and discusses the benefits of

such an implementation on particular areas of research, such as the development of Grid applications.

1 INTRODUCTION

Collaborative writing (CW) is becoming

increasingly common; often compulsory in academic

and corporate work. There are many software tools

that support CW and address the complications

inherent in most multi-user applications. However, a

common complaint with CW is the lack of

coherence between the independently-authored

sections in a document. The current practice of CW

groups writing to an agreed outline

(Alred et al., 2003) is successful but does not

address this problem. Developing a narrative instead

of an outline and agreeing to use it as a formal

structure to the document will resolve the lack of

coherence in collaboratively-authored documents.

It was shown previously (De-Silva and Henderson,

2005, De-Silva, 2005) how narratives enhance

single-author technical documents. A tool called

CANS was built based on Rhetorical Structure

Theory (RST) (Mann and Thompson, 1988) that

guides an author through a narrative-driven writing

process. This research was then extended to build a

similar tool for CW.

It is an established fact that successful applications

rely on well understood business processes

(Henderson, 2000). Therefore, after studying

existing CW strategies and tools, a business process

model was designed to describe the collaborative

processes of reading, writing and reviewing, from a

narrative perspective. This model is currently being

implemented using Web Services.

There are diagrammatic ways of presenting business

processes in UML (Maciel et al., 2005) which are

ideal for complex distributed systems. However,

since the focus is mainly on introducing narratives to

CW, the business processes remain simple and do

not warrant the use of such diagrams. Pseudocode is

used instead.

This paper has brief introductions to narratives, RST

and CW. Following this, the business process model

for narrative-based CW is described and its

implementation and uses are discussed, particularly

as an application suitable for deployment on the

Grid.

2 OVERVIEW OF NARRATIVES

A narrative is a representation of events

meaningfully connected in a temporal and causal

way (Onega and Landa, 1996, Abbott, 2002). For

the scope of this paper, it is sufficient to think of a

narrative as being analogous to a story.

A document narrative is the implicit ‘story’ a

document conveys to the reader; a fundamental

aspect of a successful document. A document

narrative clarifies the authors’ intentions and

provides a coherent structure. Narratologists and

166

Henderson P. and De Silva N. (2006).

A NARRATIVE APPROACH TO COLLABORATIVE WRITING - A Business Process Model.

In Proceedings of the Eighth International Conference on Enterprise Information Systems - SAIC, pages 166-173

DOI: 10.5220/0002462901660173

 SciTePress

linguists have developed several theories to analyse

and synthesise coherent narratives. Rhetorical

Structure Theory (RST) was chosen to produce

coherent document narratives.

2.1 Rhetorical Structure Theory

When a narrative is analysed using RST, it is

divided into text segments and relationships are

defined between them. A segment is of arbitrary

size; often a clause or sentence. The relationships

can be illustrated using diagrams (Figure 1). A text

segment assumes one of two roles in a relationship:

the nucleus (N) or the satellite (S). Nuclei express

what is more essential to the understanding of the

narrative. Satellites provide supporting information.

However, in this paper, there is no great distinction

between nuclei and satellites.

N S

Motivation

Figure 1: A motivation relationship in RST (N=Nucleus,

S=Satellite).

There is an overall effect associated with each

relation. For instance, if the nucleus in a

MOTIVATION relation presents an action, the

satellite should increase the reader’s motivation to

perform it. There are 23 relations defined in (Mann

and Thompson, 1988). The RST analyses in this

paper use just six of these relations (Table 1).

RST identifies a hierarchical structure in text.

Therefore, a coherent text is expected to produce a

tree of relations. It is possible to have multiple valid

RST trees for one narrative.

Table 1: Subset of RST relations used in this paper.

Relation Description

Background Provides background information

Elaboration Provides extra information

Justify Justifies a certain decision

Motivation Provides motivation to perform an

action

Sequence Links events that happen in

sequence

Solutionhood Provides a solution to a problem

2.2 Narrative-based Writing

To familiarize the reader with the use of RST on

document narratives, an example of writing a fable

is presented. Later, these techniques are applied to

technical documents. The document narrative below

is the structure for a set of fables that an author

wishes to write.

I want to write a short story that will contain

an implicit moral lesson.

I will use animal

characters with human features.

I believe this

will convey the wisdom in an enjoyable and

memorable way.

I will introduce two or three

characters with opposite human characteristics

(one righteous, one immoral).

These

characteristics will be revealed through brief

conversations at the start of the story.

Then

there will be a series of events that will be

tailored to demonstrate that the characters

with the moral attitude always win and that the

others suffer consequences for their unwise

actions.

Thus the reader will be gently

persuaded to take on the characteristics of the

successful characters.

Figure 2: RST analysis of the document narrative for a set of fables.

I want to write a

short story that will

contain an implicit

moral lesson.(1)

Thus the reader will

be gently persuaded

to take on the

characteristics of the

successful

characters.(7)

Motivation

2-6

Elaboration

Then there will be a

series of events that

will be tailored to

demonstrate that the

characters with the

moral attitude always

win and that the

others suffer

consequences for

their unwise

actions.

(

)

Sequence

2-3

Sequence

I will use animal

characters with

human features.(2

)

I believe this will

convey the wisdom

in an enjoyable and

memorable way.(3)

Justif

4-5

Sequence

I will introduce two or

three characters with

opposite human

characteristics (one

righteous, one

immo ra l). (4)

These characteristics

will be revealed

through brief

conversations at the

start of the story.(5)

Elaboration

A NARRATIVE APPROACH TO COLLABORATIVE WRITING - A Business Process Model

167

RST is used as a tool to verify coherence. If the

document narrative can be placed in a RST tree, it is

assumed to be coherent. The document narrative for

the fables was analysed using RST (see Figure 2

A fable structured according to this narrative should

satisfy all the RST relationships. For instance, the

famous story of the Ant and the Grasshopper (see

Appendix) fits this document narrative. This fable

can be divided into segments (see Appendix where

the fable is divided into three segments,

corresponding to sections 4, 5 and 6 of the narrative

above) and each segment creates the expected effect

on the reader. Section 4 introduces the Ant and the

Grasshopper and section 5 elaborates their

characteristics using dialogue. Section 6 presents the

onset of winter when the hardworking Ant emerges

as the winner. These sections are placed in sequence

and convince the reader that the Ant is the better role

model. The ‘Motivation’ and ‘Justify’ relations are

satisfied too.

This process can be applied to technical documents.

See (De-Silva and Henderson, 2005) for a document

narrative for a research proposal and section 3.3 of

this paper for a generic narrative for a scientific

conference presentation.

3 COLLABORATIVE WRITING

Collaborative writing (CW) is the process of

multiple authors producing one document, by

writing together and soliciting one another’s

opinions about their writing.

Since the early 70’s there has been great interest to

provide computer support for this process (Noël and

Robert, 2004); a move made more concrete by the

formation of a specific field of research called

Computer-Supported Collaborative Working

(CSCW). The resulting groupware and the World

Wide Web have revolutionized the art of writing

together. In preparation for this paper, CW software

such as PREP (Neuwirth et al., 1990) were studied

and newer technologies such as Wikis were

experimented with (JotSpot, 2004).

All RST trees in this paper have been drawn using the

free software tool, RST Tool.

O’DONELL, M. (2000) RSTTool 2.4 – A markup tool for

Rhetorical Structure Theory. Proceedings, International

Natural Language Generation Conference (INLG’2000).

Mitzpe Ramon, Israel.

3.1 Collaborative Writing Strategies

A CW group usually has an agreed-upon strategy for

producing the document which defines how the co-

authors will coordinate. There are several strategies

to choose from. Two popular methods are discussed

below.

3.1.1 Sequential Writing Model

In this model, only one person writes at a given time

and once his/her task is complete, passes the

document along to the author next in line. This

model is easy to organise and improves coordination

between the authors.

However, there are several disadvantages (Lowry et

al., 2004) such as the lack of group consensus and

the difficulty in ensuring that all document sections

are addressed adequately. Also, the order of authors

greatly influences the final document. One author

can change previous contributions or significantly

bias subsequent authors.

3.1.2 Parallel Writing Model

With this strategy, a team divides the CW task into

discrete units and works in parallel. This model has

several variants. In one, each team member is

assigned roles such as ‘writer’, ‘reviewer’, ‘editor’

and so on, depending on their expertise. Members

then work on the document according to their roles.

In another variation, the document is divided into

sections and each author is assigned a section that

he/she is responsible for. The completed sections are

submitted to the team leader who assembles them

together to form the final document. This approach

is sometimes called horizontal-division writing

(Lowry et al., 2004) and is the model that this paper

concentrates on. In (Alred et al., 2003), this process

is described in more detail (reproduced below).

1. Designate one person as the team coordinator.

2. Collectively identify the audience, purpose and project

scope.

3. Create a working outline of the document.

4. Assign segments or tasks to each team member.

5. Establish a schedule: due dates for drafts, revisions,

and final documents.

6. Agree on a standard reference guide for style and

format.

7. Research and write drafts of document segments.

8. Exchange segments for team member reviews.

9. Revise segments as needed.

10. Meet your established goals.

(Source: Alred et al., 2003)

ICEIS 2006 - SOFTWARE AGENTS AND INTERNET COMPUTING

168

3.2 Problems with CW

There are several known problems with CW and

three of them are discussed here.

Access control: CW groups have varying access

requirements. Most often, only authorized authors

are allowed to edit the document. Sometimes fine-

grained protection is required; for example, an

author may have edit privileges over only one

section in the document (Shen and Dewan, 1992).

Version control: If two authors check out version X

of the document and, after some period of time,

submit edited documents back to the system, there is

a chance one version could overwrite the other. Even

if both versions are saved, which of the two versions

would be the latest?

Lack of coherence: Another drawback of parallel

writing is the lack of coherence between the sections

that have been independently authored (Lowry et al.,

2004). The team leader is often burdened with the

task of collating these sections to produce a

consistent, coherent document. A narrative approach

to CW will help solve this problem.

3.3 Narrative-based CW

Instead of an outline, a CW group can decide on a

document narrative. They can also determine

narratives for each section of the document. Then

authors can craft their respective sections with full

understanding of the implicit and explicit narrative

goals that the document is expected to fulfil. This

will result in improved coherence and reduced work

for the team leader.

Authors may also be allowed to change the narrative

or RST relations during the writing process. If the

CW team agrees on the changes, they will modify

the sections to satisfy the new narrative.

As an example of a collaboratively produced

technical document, a generic document narrative

for a presentation of scientific results at a conference

is given below. Through this narrative, the scientists

involved in this presentation share and clarify their

intentions and agree on one ‘story’ that they wish to

convey.

This narrative can be divided among the team and

each scientist can construct a part of the presentation

such that it satisfies the narrative. For instance, the

Previous Work section should convince the audience

that it is absolutely necessary to solve this problem

and that no one else has been successful so far. It

should conclude with a lead to the next section by

stating that previous work has, however, helped this

team develop their experiments.

There was an unsolved problem in this scientific field and we have solved it.

Our research into previous work

revealed that there was no complete solution to this particular problem

and this lack was affecting specific

groups of people.

We gathered some useful ideas from these previous researchers

and set about designing our

own experiments to overcome the hurdles that they faced.

Here is the design of the experiments we conducted

and a list of our results.

These results are much better than those of our predecessors but we hope to improve

them further by conducting more experiments.

Thereby, we conclude that our results are currently the best in

this field and greatly help the people who were most affected by this problem.

2-9There was an

unsolved problem in

this scientific field

and we have solved

it.

Solutionhood

4-7

We gathered some

useful ideas from

these previous

researchers

Sequence

and set about

designing our own

experiments to

overcome the

hurdles that they

faced.

Sequence

Here is the design of

the experiments we

conducted

Sequence

and a list of our

results.

Sequence

These results are

much better than

those of our

predecessors but we

hope to improve the

further by conducting

more experiments.

Elaboration

Thereby, we

conclude that our

results are currently

the best in this field

and greatly help the

people who were

most affected by this

problem.

Motivation

2-3

Background

Our researc h into

previous work

revealed that there

was no complete

solution to this

particular problem

and this lack was

affecting specific

groups of people.

Elaboration

Figure 3: Document narrative and RST tree for a presentation of scientific results at a conference.

A NARRATIVE APPROACH TO COLLABORATIVE WRITING - A Business Process Model

169

4 BUSINESS PROCESS MODEL

Narratives are a powerful contribution to CW.

Therefore, a narrative-based tool is being developed

to support asynchronous CW. Such an application is

not easy to design and requires a way of precisely

articulating how the authors can interact with the

system. Therefore, the business process (BP) model

presented here was used to design this tool. In this

context, a BP model is an enumeration of all the

actions in which the participants can engage.

Although it may seem unusual to describe CW as a

business, it has all the characteristics of coordinated

actions inherent in any business.

To be formal about a BP, these actions, the order in

which they can be performed and the effect they

have on the shared global state must be stated. This

is non-trivial when there is more than one author

working asynchronously. Unusually, there is no

constraint on the order in which these actions can be

performed.

In the model, there is a repository which contains,

for now, one document and the corresponding

narrative. Multiple authors have access to this

repository and can perform a set of actions. These

actions are described below using pseudocode.

Every changed document or narrative submitted to

the repository is saved as a new version and assigned

the next biggest version number. Any of these

versions can be retrieved by providing its unique

version number. The version with the highest

number is assumed to be the most recent. Each

version also holds the version number of the

document it was derived from (called

parent_ver).

In the following pseudocode, ‘p’, ‘n’ and ‘d’ refer to

version numbers.

4.1 Document

Properties of a document

– ver: The version number of this document.

When a document is first created, it gets version

number 0.

– parent_ver: The version number of the

document that this document is based on.

– checked_all: This property is TRUE when

this document satisfies all the relations in the

LATEST narrative. Each version of the

document starts with checked_all=FALSE

because it is assumed that one or more of the

sections fail to satisfy the latest narrative. After

several cycles of read, write and review, the

document should reach the ideal state where

checked_all=TRUE. Documents that do

not satisfy the latest narrative (but may have

satisfied older narratives) are not acceptable.

Therefore, when the narrative is modified, the

checked_all property of all document

versions are set to FALSE based on the

assumption that all documents no longer satisfy

the new narrative.

4.2 Narrative

Properties of a narrative

– ver: The version number of this narrative.

When a narrative is first created, it gets version

number 0.

– parent_ver: The version number of the

narrative that this narrative is based on.

4.3 Repository

Properties and actions relevant to the documents in

the repository are prefixed with the letter ‘d’ and

those pertaining to the narratives with ‘n’. ‘Rep’

denotes the repository.

Properties of the repository

– dlatest: The version number of the latest

document

– nlatest: The version number of the latest

narrative

Actions that can be performed on the repository

– dget(v): Returns the document with version

number v (0 <= v <= rep.dlatest)

– nget(v): Returns the narrative with version

number v (0 <= v <= rep.nlatest)

– dput(doc,p): Saves the document doc

in the repository and sets the following

properties:

doc.parent_ver = p

doc.ver = rep.dlatest + 1

doc.checked_all = FALSE

rep.dlatest = doc.ver

– nput(nar,p): Saves the narrative nar in

the repository and sets the following properties:

nar.parent_ver = p

nar.ver = rep.nlatest + 1

rep.nlatest = nar.ver

Set checked_all property of all

document versions to FALSE

–

check(doc,nar): This is a process

performed by authors/reviewers to check if the

ICEIS 2006 - SOFTWARE AGENTS AND INTERNET COMPUTING

170

sections in document doc satisfy the

relationships in narrative nar (see section 2.2 for

an example of a document that satisfies its

narrative). ‘Nar’ is often the latest narrative. If

each section satisfies all the relationships

pertaining to it, then the property

checked_all of document doc is set to

TRUE.

4.4 Specification of an Author’s

Tasks

Typically, a co-author’s tasks in this model are

reading, writing and reviewing the document and,

reading and editing the narrative. These tasks are

described below.

‘Do forever’ loops mean that

an action can be repeated as many times as the

author wishes until the document is complete.

1. Read latest version of document

do forever {

Document doc;

doc = rep.dget(rep.dlatest);

~~~~ read ~~~~

}

2. Edit version v of the document and submit new

version to the repository.

do forever {

Document doc;

doc = rep.dget(v);

~~~~ edit ~~~~

doc’ = modified doc;

rep.dput(doc’, v);

}

3. Review a version of the document

An author or reviewer retrieves a version of the

document and checks it against the latest narrative.

If the narrative is satisfied,

checked_all of the

document is set to

TRUE. If any of the narrative goals

are not fulfilled, the authors need to edit the

document or the narrative so that they once again

become fulfilled.

do forever {

Document doc = rep.dget (v);

Narrative nar= rep.nget (rep.nlatest);

~~~~ review content of doc ~~~~

rep.check (doc,nar);

if all relationships are satisfied {

doc.checked_all = TRUE;

}

4. Read the latest version of the narrative

do forever {

Narrative nar;

nar = rep.nget(rep.nlatest);

~~~~ read ~~~~

}

5. Edit version v of the narrative and submit new

version to the repository.

do forever {

Narrative nar;

nar = rep.nget(v);

~~~~ edit ~~~~

nar’ = modified nar;

rep.nput(nar’, v);

}

4.5 Authorized Access and Version

Control

The repository can maintain a table with the author

IDs and their access rights. Before each request to

read or edit a document/narrative, this table can be

queried and only authorized authors can be allowed

to perform these operations.

Version control needs to be addressed in more detail

but is not the focus of this research. In this model,

any previous version can be used to produce a new

one. Each new version is stored with a link to the

document it was derived from (

parent_ver), thus

producing a tree of versions as shown below.

Figure 4: Tree of document versions in the repository.

According to this model, version 5 of this document

will be considered the latest. However, versions 3, 4

and 5 are at the same level in the tree and have equal

chances of being superior or the most appropriate.

Either the authors will be given the authority to

choose the latest version among themselves or the

3 5

A NARRATIVE APPROACH TO COLLABORATIVE WRITING - A Business Process Model

171

model will be enhanced to reconcile the differences

between the leaves of the tree. This approach with

maximum freedom has been selected over more

restrictive methods such as ‘check-in, check-out’

version control because, in designing this creative

writing model, the aim has been to support rather

than constrain the authors.

5 CONCLUSION AND FUTURE

WORK

Collaborating with other authors to write is

increasingly common. A document thus produced is

often lacking in coherence because of the

independently-authored sections. Research presented

in this and previous papers (De-Silva and

Henderson, 2005, De-Silva, 2005) has shown that

applying narratives and RST can significantly

improve coherence in single-author and

collaboratively-produced documents.

A critical factor for an effective document is a

coherent document narrative. Noticing a lack of

support for document narratives in existing writing

tools, we developed CANS (De-Silva and

Henderson, 2005): a tool that guides an author

through a narrative-based writing process (Figure 5).

Ideas from this single-user tool are now being

developed into a multi-user application that supports

asynchronous CW.

Before embarking on this, it was critical to articulate

how multiple authors will interact with the system.

This was achieved by designing a business process

model for the ‘business’ of narrative-based CW.

This model is implemented using a Web Service.

The document and all updates to it are stored in an

XML database maintained by Xindice (Xindice,

2004).

Finer details of this model are constantly being

refined. A more elaborate model will include a

record of which documents satisfied previous

versions of the narrative. Also, it is possible that

some sections in a document will remain unaffected

even after the narrative is updated. So, a document

can be tracked more closely by assigning each of its

sections a ‘

checked’ property which is true only if

the section satisfies the RST relations in the latest

narrative. After a new narrative has been submitted,

only sections with

checked=FALSE will need to be

changed.

Owing to the authors’ involvement with the Open

Middleware Infrastructure Institute (OMII, 2005)

collaborative working over the Grid is of particular

interest. Collaborative writing is an ideal application

for the Grid. After sufficient development and

testing, the narrative-based CW tool will be

deployed on the OMII middleware to serve as an

example.

During this process, several important lessons can be

learnt. Simple and generic ways to convert single-

author tools such as CANS to multi-user

applications deployable on the Grid will be

invaluable to the Grid community. It will also be

useful to enumerate the differences between Web-

based and Grid-based applications. One such

difference is the continuous change in resource

locations in the Grid that require location-

independent design and implementation. So,

documents/narratives in a CW application could

move or be split across multiple databases, but

authors should still have an interface which gives

them reliable access to their work.

This work builds an island amidst several strands of

parallel research and brings together many

technologies. The current CW practice of working to

an outline is adequate, but does not solve the lack of

Figure 5: Screen shot of single-author writing tool CANS (First step: Entering document narrative).

ICEIS 2006 - SOFTWARE AGENTS AND INTERNET COMPUTING

172

coherence. Therefore, narratives are a novel and

better approach, providing more support for CW.

Document production from a narrative aspect will

revolutionize the way people read, write and

evaluate documents.

REFERENCES

ABBOTT, H. P. (2002) The cambridge introduction to

narrative, Cambridge, UK, Cambridge University

Press.

ALRED, G. J., BRUSAW, C. T. & OLIU, W. E. (2003)

Handbook of technical writing, Boston, MA,

Bedford/St. Martin's.

DE-SILVA, N. (2005) A narrative approach to technical

document construction. PREP 2005, awarded best

oral presentation in Computer Science. Lancaster

University, UK.

DE-SILVA, N. & HENDERSON, P. (2005) Narrative

Support for Technical Documents: Formalising

Rhetorical Structure Theory. 7th International

Conference on Enterprise Information Systems

(ICEIS). Miami, FL, USA.

HENDERSON, P. (2000) Business Processes, Legacy

Systems and a Fully Flexible Future. IN

HENDERSON, P. (Ed.) Systems Engineering for

Business Process Change. New York, USA, Springer-

Verlag.

JOTSPOT (2004) JotSpot - The application Wiki. Found

at http://www.jot.com/index.php (Last accessed June,

2005).

LONG, J. R. (1997) Aesop's fables: Online collection.

Found at

http://www.pacificnet.net/~johnr/cgi/aesop1.cgi?sel&

TheAntandtheGrasshopper&&antgrass.ram (Last

accessed: 6.10.2005).

LOWRY, P. B., CURTIS, A. & LOWRY, M. R. (2004)

Building a taxonomy and nomenclature of

collaborative writing to improve interdisciplinary

research and practice. Journal of Business

Communication, 41, 66-99.

MACIEL, R. S. P., SILVA, B. C. D., FERRAZ, C. A. G.

& ROSA, N. S. (2005) An MDA-EDOC based

development process for distributed applications. 7th

International Conference on Enterprise Information

Systems (ICEIS). Miami, USA.

MANN, W. & THOMPSON, S. (1988) Rhetorical

Structure Theory: Toward a functional theory of text

organisation. Text, 8, 243-281.

NEUWIRTH, C. M., KAUFER, D. S., CHANDHOK, R.

& MORRIS, J. H. (1990) Issues in the design of

computer support for co-authoring and commenting.

3rd conference on computer-supported cooperative

work. LA, California, USA, ACM press.

NOËL, S. & ROBERT, J.-M. (2004) Empirical Study on

Collaborative Writing: What do co-authors do, use,

and like? Computer Supported Cooperative Work, 13.

OMII (2005) http://www.omii.ac.uk/ (Last accessed:

11.10.2005).

ONEGA, S. & LANDA, J. (1996) Introduction. IN

ONEGA, S. & LANDA, J. (Eds.) Narratology. New

York and London, Longman Group Ltd.

SHEN, H. & DEWAN, P. (1992) Access control for

Collaborative Environments. ACM conference on

Computer Supported Cooperative Work. Toronto,

Canada, ACM press.

XINDICE (2004) Online at http://xml.apache.org/xindice/

(Last accessed on 25.6.2004).

APPENDIX

The Ant and the Grasshopper

In a field one summer's day a Grasshopper

was hopping about, chirping and singing

to its heart's content. An Ant passed by,

bearing along with great toil an ear of

corn he was taking to the nest.

“Why not come and chat with me,” said

the Grasshopper, “instead of toiling and

moiling in that way?”

“I am helping to lay up food for the

winter,” said the Ant, “and recommend

you to do the same.”

“Why bother about winter?” said the

Grasshopper; “we have got plenty of food

at present.”

But the Ant went on its way and

continued its toil. When the winter came

the Grasshopper had no food and found

itself dying of hunger, while it saw the

ants distributing every day corn and grain

from the stores they had collected in the

summer.

Then the Grasshopper knew: It is best to

prepare for the days of necessity.

(Long, 1997)

*Segments numbered according to document

narrative in section 2.2

A NARRATIVE APPROACH TO COLLABORATIVE WRITING - A Business Process Model

173