WEB BASED COLLABORATIVE DOCUMENT CREATION AND

REVIEW SYSTEM

Marius Ioan Podean, Raluca Arba

Faculty of Economics and Business Administration

Babes Bolyai University of Cluj Napoca, Teodor Mihali 58-60, Cluj Napoca, Romania

Loredana Muresan

Faculty of Economics and Business Administration

Babes Bolyai University of Cluj Napoca, Teodor Mihali 58-60, Cluj Napoca, Romania

Keywords: Collaborative document editing, document management, workflow, virtual teams, XML.

Abstract: An important aspect in distributed teams and organization is their ability to manage documents. A

collaborative document editing system that integrates functionalities from document management systems,

workflow, collaborative editing with support for virtual teams can increase team efficiency and allow users

to concentrate their efforts on content development. This paper reports a case study implementing this

approach in collaboratively creating scientific papers. The use of XML when treating documents proves to

be the appropriate solution to develop user and document centered systems.

1 INTRODUCTION

Creating scientific papers is most of the time a very

complex and elaborate task. This type of documents

usually is written by more then just one person, and

often the authors are located in different parts of the

world. This collaborative process is sometimes

slowed down by the fact that dealing with the

technology used to support it is very much time

consuming. Taking in consideration the authors need

for location and time independence, the usage of

different operating systems and applications, at this

moment technology offers solutions that satisfy just

partially the aforementioned requirements.

Document management systems (DMS) focus on

tracking and storing documents created and

exchanged by there authors (Aversano et al. 2001).

They provide components for defining metadata for

the documents (i.e. date of creation, authors, version

etc.), indexing (usually based on metadata), storage

and retrieval (based on the unique document

identifier).

Workflow management systems (WMS) allow

the automation of processes within an organization,

enabling greater coordination and control among

geographically distributed teams (Nallaparaju et al.

2005). Using WMS, the organization can integrate

different software technologies, leading to the

improvement of the collaborative activities

(Aversano et al. 2001). This technology leaves

authors with the consumption of a great deal of time.

Collaborative editing allows multiple persons to

edit simultaneously the same document, see who is

working on the document and watch in real time the

changes that they have made (Raikundalia and

Zhang 2004). In order to have sufficient knowledge

about the changes that others perform upon the

document, group awareness mechanisms have been

created, such as: telepointers (multiple cursors of

users appear within the document), radar views,

multi-user scrollbars and, as shown by Raikundalia

and Zhang (2004), structure-based multi-page view,

point jumping mechanism and user info list. The

main downside of this solution is that the

implementations available are platform-specific and

can generate conflicts between members when

someone changes often content created by others.

The collaboratively edited file is stored on the

document owner’s computer, leading towards

versioning problems when participating members

make local copies of the document.

437

Ioan Podean M., Arba R. and Muresan L. (2008).

WEB BASED COLLABORATIVE DOCUMENT CREATION AND REVIEW SYSTEM.

In Proceedings of the International Conference on e-Business, pages 437-442

DOI: 10.5220/0001907404370442

 SciTePress

Wikis represent “a piece of software that allows

users to add, modify, and/or delete information from

a knowledge base via the web” according to Spek

(2008). As the author of this definition emphasis as a

main characteristic, wikis are anarchistic systems,

many implementations allowing anonymous users to

modify the content. On wiki systems “conflicts can

quickly result in ‘edit wars’ when multiple users

keep on reverting each others changes because they

don’t agree” as shown by Spek (2008).

Creating research papers requires a system that is

simple to use and allows users to focus their efforts

on the content rather that on the technology used to

create it. The system has to support the space

independence of the authors and integrate their

collaborative effort in a common workplace in order

to obtain greater team efficiency (Guerrero et al.

2004). Collaborative work involves information

exchange in order to support negotiation and

communication between group members and

different mechanisms through which the team can

regulate and manage itself in order to be goal

directed (Millward and Kyriakidou 2004). A more

supportive system would have document

management facilities and support for task

automation. To achieve greater efficiency the system

will have to be user centered and non–restrictive

regarding de operating system.

In this paper, we start by presenting current

technologies used for document management and

editing highlighting their main characteristics and

downsides. We will continue in Section 2 with the

theoretical approach of our system and then discuses

in Section 3 the implementation and the technical

details concerning it. Based on the model we’ve

proposed, in Section 4 we shall present some

conclusions and further work.

2 THEORETICAL APPROACHES

In this paper we shall present our implementation of

such a system that aims to cover the aforementioned

requirements. Dante is a web based system designed

to be a good support for virtual teams in elaborating

scientific papers. It offers document management

facilities and process automation for repetitive tasks.

Since all the data the system uses is stored in XML

files, Dante can be document and user centered,

allowing authors to easily edit, review and export in

different formats their work. Teams are building

around the document allowing them to be goal

directed; all members of the team having the same

rights. Each author is responsible of editing different

chapters of the document, chapters on which the

others could only place comments, content changing

not being allowed, avoiding conflicts and stepping

on each others and allowing individuals to reconcile

with the teams goals. The application facilitates

communication through synchronous and

asynchronous channels. Documents being stored in

XML files and using a web based user interface

makes the system work on different platform and

allows users to export documents in different open

formats, representing a combination of best practices

specific for the previous discussed systems. At this

moment the application offers no version control

capabilities, and therefore authors could not revert

documents to older versions.

As mentioned earlier, Dante is both a user and

document centric system, supporting collaboration

in virtual teams and efficient document

management. As described by Millward and

Kyriakidou (2004), it is important for virtual teams

to be a “singular concrete entity” with the following

characteristics: stability, regular interaction,

symbiosis and member proximity. Following this

requirements, in Dante teams are organized around

the document that they are creating. Member

proximity results from the fact that each member can

view the most recent version of the chapters that the

others have created and that everybody knows who

is responsible for a particular chapter. Each author

can review others work and make suggestions

related to each piece of text using the commenting

tools. The symbiosis of the team is supported by the

fact that each person’s responsibilities are clearly

drawn and all members have the same rights, all

depending upon others in improving their work.

Interaction between members is supported by both

synchronous and asynchronous mechanisms.

In collaborative real time editing systems several

users can edit a file using different computers. An

important aspect for this type of systems is group

awareness (GA). GA provides users information

about the status of a document and changes made by

others. As shown in Raikundalia and Zhang (2004),

several GA techniques have been identified:

telepointers (multiple cursors are shown within the

document), radar views, multi-user scrollbars and

structured-based multi-page view, point jumping

mechanism and user info list. In Dante, the GA

problem is solved by using a structured-based multi-

page view panel for displaying a project. One of the

main downsides of all collaborative editing systems

(CES) is that the document is saved on the document

owner’s computer, all others participants being

allowed to save a copy of the document leading in

ICE-B 2008 - International Conference on e-Business

438

time to lost edits and versioning problems. Dante

being a web based system it has the advantage of

delivering file storage facilities. All changes made

by the authors are saved on the server which will

offer only the most recent version of the document.

Authors edit different parts of the document and

send changes to the server via AJAX.

As shown in Leone, Hodel and Gall (2005),

combining CES and DMS can result in greater

performance improvements. Dante takes a similar

approach providing solutions and process

automation for creating, storing retrieving, editing

and exporting documents collaborative, anytime on

the most used platforms. All documents have

metadata to easily manage them and provide extra

search capabilities. Another important aspect is the

presence on the internet and the collaboration with

persons from outside the team using e-prints.

Lawrence (2001) shows that articles freely available

online are more highly cited and recommends, in

order to achieve greater impact and faster scientific

progress, that authors should aim to make research

easy to access. Content can easily be transformed

into HTML and allow others to post comments on

the article if the team wishes so.

Figure 1: Dante architecture.

To summarize, the main characteristics that we

consider essential for an efficient collaborative

document creation and review system are:

support for virtual team efficiency

user and document centered

support users location and time independency

platform independent

efficient group interaction

allow users to concentrate on the content and

waste as little as possible time with the technology

that supports the process

document management facilities

automation for different parts of the process (i.e.

formatting the document according to different

templates)

As shown in Figure 1, Dante is a web based

system design upon de model-view-controller design

pattern which stores all data in XML files. This

approach allows a great flexibility in handling

document and user interaction. The presentation

layer sets the appropriate tools according to user

level and rights and manages the client-side

communication with the controller layer using

AJAX.

The presentation layer defines de following main

tools: Editor, Partners and Chat. The Editor allows

users to manage old projects, start new ones,

collaboratively edit current projects and export

documents in different open formats and format the

content using predefined templates. When starting a

new project, the project owner must define a set of

properties, like name (works like an identifier for the

project), type (according to the type that has been

selected, a certain template will be used to format

the content when users export the document) and a

list of members that will participate to the project

(after defining the project, all participants will have

the same rights) (Leone, Hodel and Gall 2005).

During development extra sets of metadata will be

added to the project allowing users to consult the

state of the project, last modified date etc. Each

author can edit one or more chapters from the

document and place comments on those chapters

edited by the others. Chapters are presented as

elements in a tree menu and can bee accessed in

different windows. When accessing a particular

chapter, the system checks the metadata associated

to it and determines if the current user is the author

of the chapter and displays a new menu that allows

him to edit the content if so, otherwise allowing him

only to add new comments to the component

elements or edit comments define earlier. The editor

is not made up from elements that will allow authors

to format the content (i.e. defining font types,

paragraph alignment etc.), but from elements that

represent structural components of document (i.e.

paragraph, note, quote, table etc.). The formatting of

document will be done automatically by the

application according to the template that the

document owner had defined. Each structural

component of the document can be commented by

the rest of the team, the editor defining special zones

at the end of each element where these comments

can be consulted. When accessing chapters for

whom the current users is not the owner, a different

editor is loaded allowing the user to define and edit

WEB BASED COLLABORATIVE DOCUMENT CREATION AND REVIEW SYSTEM

439

comments and reply to those that others have

defined. Before finalizing the project, the chapters

can be exported in different open formats (i.e. PDF)

or to a link in order to be accessed by those that are

not participating to the project.

The Chat sections allows user to communicate

using synchronous channels. The section defines two

main channels: Groups and Personal. The Groups

channel lists all active projects for the current user

and allows him to communicate with the members

of each project in different panels. When a particular

group is selected, the user can communicate with the

members of that project that are online, the messages

being available for all co-workers. All messages are

stored in XML files and associated to the project,

allowing user to consult the discussion archive or the

project anytime. On the other hand, the Personal

channel allows user to communicate with each

friend defined in the Agenda using private channels.

These messages are not stored in the message

archive, only if the users decide so.

The Partners section has two main

subcomponents: Agenda and Invitations. The

Agenda allows users to manage their partners’

contacts and export them to micro-formats like

HCard. The user’s personal data can be exported in

VCard format. The Invitations subcomponent

manages the user’s invitations to participate at

different projects (only after accepting an invitation

a user can actively participate to a project). All

personal data is stored in XML files, allowing the

system to easily integrate them into the projects

content when exporting a project in a final state

according to different templates.

In the next section we shall discuss the details

regarding the implementation of our web based

collaborative document creation and review system.

3 IMPLEMENTATION

As previously mentioned, Dante is a web based

system design upon de model-view-controller

design-pattern which stores all data in XML files.

The View layer personalizes the user interface

according to user’s rights and permissions and

displays the appropriate editor after reading the

documents metadata. It also implements a

communication module which transfers data to the

Controller layer using AJAX. The Controller layer

handles the events triggered from the UI and calls

the appropriate handler from the Model layer. We

will concentrate our attention on the Model layer

witch manages all XML documents.

The Model layer consists mainly from two

subcomponents: the Document manager and the

Repository. The Repository is a collection of DTD,

XML and XSLT files used by the Document

manager module.

The Model layer consists mainly from two

subcomponents: the Document manager and the

Repository. The Repository is a collection of DTD,

XML and XSLT files used by the Document

manager module.

Figure 2: Document workflow.

The document type definition files describe the

structure of the documents and all metadata that can

be added to them. In this case, to describe the

structure of the scientific papers we’ve created a

document type called xmlDocument. The structure

of this type is in a way similar to the more mature

DocBook xml vocabulary (Walsh 2005), being more

simplified and particularized for scientific papers.

The xmlDocument DTD defines the structure of the

whole document viewed as a project. The two main

elements of an xmlDocument are docInfo, which

describes the metadata associate with the document

(i.e. document name, authors, document type, last

update, version etc.) and docBody. The docBody

defines the content of the document (i.e. abstract,

references, appendixes etc.) and for to each chapter

the metadata associated (id, title and author) and a

link to the XML file that defines its content. This

separation of each chapter in different XML files

allows a more flexible management of content and

metadata. The vocabulary that describes the

chapter’s elements is defined by the xmlChapter

document type definition, implementing means of

storing and identifying elements. As mentioned

earlier, each author can define comments for the

chapters that he is not editing. These comments are

stored in separate XML files, and the vocabulary

describing them defines elements to uniquely

identify de comment, describe its characteristics (i.e.

author, date etc.) and link it to the element to which

it referees.

ICE-B 2008 - International Conference on e-Business

440

In order for the Document manager to easy set

the corresponding roles and rights, a document type

definition is implemented to store the projects. This

DTD allows defining for each user the projects for

which he is a collaborator or the owner. The

Document manager module consist in several

classes based on the DOM implementation which

automates the processes concerning the creation,

storage, retrieval and the export of the documents.

All server side scripts have been implemented in

PHP which includes a good support for XML,

complying with the commonly used standards (SAX,

DOM, SimpleXML, XMLReader, XMLWriter, and

the XSLT). Some of the classes defined in the

document manager come as a wrapper for the DTD

residing in the repository: the xmlChapter class, for

example, handles the creation and the update of

different document objects, prepares the content for

printing accessing the appropriate class when

exporting (i.e. the pdfPrinter class when exporting to

PDF). This module generates the appropriate user

interface according to the user’s role and rights,

loading the corresponding XSL and parsing the

requested XML file. For the same DTD the

Repository defines different XSL transformation

files for each role and user-level, restricting the

access to different actions that the user can perform

upon the document. Most of the UI used by Dante is

generated using XSL transformations.

The Chat module works in a similar way,

defining a DTD for storage purposes and different

XSL transformation schemes for display. As

mentioned previously, the user can communicate

using private channels or rooms dedicated to

different projects. The messages exchanged between

users in these project related rooms are bound to the

project using metadata (although physically residing

in different locations) allowing project based

message archiving. This approach allows members

to catch-up with the team when not being able to

join the group. The Agenda module stores member’s

personal data (this being done also using XML files)

and allows users to export this information in HCard

and VCard microformats. All templates require a

minimum of identification data for the authors, the

project importing all required data when formatting

from this module. We have tried to structure the

implementation as much as possible according to

functionality of the whole system, defining a

modular structure.

The user interaction is managed using

JavaScript. Because the communication from the

server to the client is mostly done using XML

chunks (or an entire file), the classes defined on the

client-side are in some manner a transposing of

those residing on the server being also a wrapper

upon the document. The Session Manager module

from the Controller layer determines which script to

be loaded according to conjuncture determining the

operations that a user can execute on a particular

document. When editing a particular chapter, all new

elements and all changes are stored in different

queues and only when the users decides to save the

document the content of these queues is send using

AJAX to the server. These classes are also based on

the DOM model, the content being sent to the server

representing a DOM node to be inserted in the

document residing on the server. The queues are

gradually discharged when a component element is

successfully saved on the server. This approach

allows defining specific UI elements and behavior to

each document type being treated. The user screen is

blocked while saving the content of the queues,

ousting the chances to make unsafe changes to the

content. When appropriate, an XSLT processor is

used on the client-side to reduce the charge on the

server.

We have chosen this modular approach based on

document types and user interfaces particularized on

roles/rights and document types in order to facilitate

further development of the system in such a manner

that it could handle a lot more document types.

4 CONCLUSIONS

The work reported in this paper has addressed the

problem of integrating document management

functionalities and workflow capabilities into single

system with support for virtual teams in order to

achieve a very efficient solution for collaborative

document editing. Allowing users to edit

collaboratively documents and supporting their

needs for time and location independence can result

in more team efficiency. Users have to be able to

concentrate their efforts on content and reduce as

much as possible the time used to handle and

integrate technologies. An approach that integrates

functionalities from document management systems,

workflow management and collaborative editing

proves to be a much more user centered and

supportive collaborative solution. As such an

implementation, Dante tries to offer sufficient

support in order for the teams to be goal directed and

efficient. Using a web based solution does not

impose restrictions on the users regarding operating

systems or software. Taking in consideration the

multitude of solutions available, we have to consider

WEB BASED COLLABORATIVE DOCUMENT CREATION AND REVIEW SYSTEM

441

using as much as possible open formats in order for

the content to be easily integrated. As shown in this

paper, using XML as building blocks allows the

system to be document and user centered. We have

presented a case study of a particular field where

such an implementation would be a great support.

As further work, we intend to develop our system

toward a framework that can handle a wide range of

documents, taking in consideration the increased

need for collaborative document editing in a

multitude of working fields.

ACKNOWLEDGEMENTS

The work presented has been funded by the research

grant “Intelligent System for Business Decisions

Support”, Director Nitchi Stefan, PhD., Professor,

PNII Program, 91-049/18.09.2007 supported by

Higher Education Ministry.

REFERENCES

Aversano, L, Canfora, G, Lucia, A & Gallucci, P 2001,

‘Integrating Document and Workflow Management

Systems’, Proceedings of the IEEE Symposia on

Human-Centric Computing Languages and

Environments,

http://csdl2.computer.org/comp/proceedings/hcc/2001/

0474/00/04740328.pdf

Guerrero, L & Collazos, C & Pino, J & Ochoa, S &

Aguilera, F 2004’Designing Virtual Environments to

Support Collaborative Work in Real Spaces’, in

Journal of Web Engineering, vol.2, no.4, pp. 282-294

Hampel, T, Keil-Slawik, R 2000, ‘A Collaborative

Document Management Environment For Teaching

And Learning’, in Proceedings of the Third

International Conference on Collaborative Virtual

Environments. San Francisco (Ca.), USA, pp.197-198.

Lawrence, S 2001, ‘Online or invisible?’, Nature, vol. 411,

no. 6837, p.521,

http://www.idemployee.id.tue.nl/g.w.m.rauterberg/pub

lications/CITESEER2001online-nature.pdf

Leone, S & Hodel, T & Gall, H 2005, ‘Concept and

Architecture of a Pervasive Document Editing and

Managing System’, in Proceedings of the 23rd annual

International Conference on Design of

Communication (SIGDOC), Coventry, United

Kingdom, ACM Press, pp. 41-47

Millward, L & Kyriakidou, O 2004, ‘Effective virtual

teamwork: a social-cognitive and motivational model’,

in Virtual and Collaborative Teams: Process,

Technologies and Practices, ed. Idea Group Inc.,

United Kingdom

Nallaparaju, V & Raikundalia, G & Brand, C & Bain, C &

Hutchinson, A 2005, ‘An Experimental Study of

Workflow and Collaborative Document Authoring in

Medical Research, in Proceedings of Tenth

Australasian Document Computing Symposium,

Sydney, Australia,

http://eprints.vu.edu.au/archive/00000648/

Raikundalia, G & Zhang, H 2004, ‘Novel Group

Awareness Mechanisms for Real-time Collaborative

Document Authoring’, The Ninth Australasian

Document Computing Symposium (ADCS 2004),

University of Melbourne, Australia,

http://www.cs.mu.oz.au/~alistair/adcs2004/papers/pap

er05.pdf

Spek, S 2008, ‘Wikis are good for knowledge

management’, submitted 6 February, 2008,

arXiv:0802.0745v1

Walsh, N 2005, DocBook 5.0: The Definitive Guide,

O'Reilly & Associates

ICE-B 2008 - International Conference on e-Business

442