<!DOCTYPE article PUBLIC "-//NLM//DTD JATS (Z39.96) Journal Archiving and Interchange DTD v1.0 20120330//EN" "JATS-archivearticle1.dtd">
<article xmlns:xlink="http://www.w3.org/1999/xlink">
  <front>
    <journal-meta />
    <article-meta>
      <title-group>
        <article-title>An extension of i* to Model CSCW Requirements Applied to a Collaborative Conference Review System</article-title>
      </title-group>
      <contrib-group>
        <contrib contrib-type="author">
          <string-name>Miguel A. Teruel</string-name>
          <email>miguel@dsi.uclm.es</email>
          <xref ref-type="aff" rid="aff0">0</xref>
        </contrib>
        <contrib contrib-type="author">
          <string-name>Elena Navarro</string-name>
          <email>enavarro@dsi.uclm.es</email>
          <xref ref-type="aff" rid="aff0">0</xref>
        </contrib>
        <contrib contrib-type="author">
          <string-name>Víctor López-Jaquero</string-name>
          <xref ref-type="aff" rid="aff0">0</xref>
        </contrib>
        <contrib contrib-type="author">
          <string-name>Francisco Montero</string-name>
          <email>fmontero@dsi.uclm.es</email>
          <xref ref-type="aff" rid="aff0">0</xref>
        </contrib>
        <contrib contrib-type="author">
          <string-name>Pascual González</string-name>
          <xref ref-type="aff" rid="aff0">0</xref>
        </contrib>
        <aff id="aff0">
          <label>0</label>
          <institution>LoUISE Research Group, I3A, University of Castilla- La Mancha</institution>
          ,
          <country country="ES">Spain</country>
        </aff>
      </contrib-group>
      <pub-date>
        <year>2011</year>
      </pub-date>
      <fpage>84</fpage>
      <lpage>89</lpage>
      <abstract>
        <p>In collaborative systems, users work together in order to collaborate, communicate and coordinate each other. To perform these tasks, users should be aware of other user's actions, usually by means of a set of awareness techniques. In this paper, CSRML (Collaborative System Requirements Modelling Language) is presented as an extension of i* to deal with the specification of the CSCW requirements. In these systems collaboration and awareness of other users' presence / actions are paramount. We apply CSRML to a conference review system, where papers are reviewed in a collaborative way.</p>
      </abstract>
      <kwd-group>
        <kwd>Collaborative systems</kwd>
        <kwd>Awareness</kwd>
        <kwd>Requirements Engineering</kwd>
        <kwd>Goal-Oriented</kwd>
        <kwd>i*</kwd>
        <kwd>CSRML</kwd>
      </kwd-group>
    </article-meta>
  </front>
  <body>
    <sec id="sec-1">
      <title>-</title>
      <p>Requirements elicitation can be considered the cornerstone to achieve the quality of
the developed systems. Failing in accomplishing this phase can make the rest of the
development process also fail, with the consequent cost in terms of time and money.
Therefore, a correct requirements specification is paramount for any kind of system.</p>
      <p>As in traditional single-user systems, CSCW (Computer Supported Cooperative
Work) systems are not exempt from this need. They are a special kind of software
whose users can perform collaboration, communication and coordination tasks. These
systems have to be specified by using a special set of requirements, usually of a
nonfunctional nature. They usually result from the users' need of being aware of the
presence and activity of other remote or local users, with who they perform the above
mentioned collaborative tasks. This is the so-called Workspace Awareness, which can
be defined as the up-to-the-moment understanding of another person’s interaction
within a shared workspace.</p>
      <p>Then, a proper specification of the system, identifying clearly the requirements of
the system-to-be, specially the awareness requirements, is one of the first steps to
overcome this problem.
2</p>
    </sec>
    <sec id="sec-2">
      <title>Objectives of the research</title>
      <p>
        In previous works [
        <xref ref-type="bibr" rid="ref1">1</xref>
        ], we analyzed which requirements engineering (RE) technique:
Goal-Oriented (GO), Use Cases or Viewpoints is more appropriate to specify the
requirements of collaborative systems, and we found that GO provides more facilities
to model the requirements of this kind of systems. Once we determined GO as the
most suitable technique, we analyzed which GO approach deals with CSCW systems
in a better way [
        <xref ref-type="bibr" rid="ref2">2</xref>
        ]. The analyzed approaches were NFR Framework, i* Framework
and KAOS Methodology for the specification of collaborative systems, paying special
attention to awareness requirements. As a result of this experiment, we concluded that
the analyzed GO approaches are not fully appropriate to model collaborative system
characteristics and its relationships with awareness and quality requirements. These
conclusions, together with the results of [
        <xref ref-type="bibr" rid="ref1">1</xref>
        ] support our initial hypothesis: a RE
technique to address the problems detected during this study is required. This
technique should adopt some features from the analyzed GO approaches and should
cover the lack of expressiveness in certain aspects that current GO techniques present.
This constitutes the main aim of this work: to adapt/extend a GO notation for this kind
of systems. Concretely, and according to the conclusions of our previous study [
        <xref ref-type="bibr" rid="ref2">2</xref>
        ] the
most appropriate approach to deal with this kind of systems is i*. Therefore, in this
paper CSRML (Collaborative Systems Requirements Modelling Language) [
        <xref ref-type="bibr" rid="ref3">3</xref>
        ] is
described, by extending i* to provide the required expressiveness to model the special
characteristics of CSCW stakeholder requirements.
3
      </p>
    </sec>
    <sec id="sec-3">
      <title>Scientific contributions</title>
      <p>Because of the special kind of requirements of CSCW systems, we present CSRML
as an extension of i* that includes some elements for modelling the special
collaboration features of CSCW systems. The elements of CSRML (Fig. 1), excluding
those whose meaning is the same as in i*, are:
· Role: A role is a designator for a set of related tasks to be carried out. The
difference between i* and CSRML is that an actor playing a role can participate in
individual or collaborative tasks (through participation links) and can be the
responsible for the accomplishment of a goal (through responsibility links). Thus,
an actor can both dynamically change the roles it plays, and simultaneously play
several roles. In addition, the graphical notation is also different from the i* role
(the concept of role/actor boundary is not used in CSRML).
· Actor: An actor is a user, program, or entity with certain acquired capabilities
(skills, category, and so forth) that can play a role in executing (using devices) or
being responsible for actions. An actor has to play a role (specified by means of a
playing link, see Fig. 1) in order to participate in the system.
· Task: The concept of task in CSRML is the same as in i*. They only differ in the
introduced notation to define the importance of a task: one, two or three
exclamation signs, depending on the importance of the task. Two kinds of CSRML
tasks have been identified:
! Abstract task: This kind of task consists in an abstraction of a set of concrete
tasks and, possibly, other elements. We are not able to assign participation links
directly to this kind of tasks.
! Concrete task: These are the tasks the participants are involved to. The abstract
tasks are refined in these ones. Participants will be assigned to the task through
participation links. There are four types of these tasks:
o Individual task is a task that an actor can perform without any kind of
interaction with other actors.
o Collaboration / Communication / Coordination task two or more actors are
involved in order to perform any kind of collaboration / communication /
coordination among them.
· Awareness softgoal: CSRML refines the i* concept of softgoal into a new
specialization: awareness softgoal, that represents a special need of perception of
other user’s presence / actions, without which the task the user wants to perform
would be affected negatively or even could not be done.
· Awareness resource: This special kind of resource corresponds to an
implementation or a design solution to accomplish an awareness softgoal.
· Playing link: A playing link is used to represent when an actor assumes a role. This
link has a guard condition that represent when a role can be played by an actor.
· Participation link: A participation link denotes who are involved in a task. This
link has an attribute to specify its cardinality, i.e., the number of users that can be
involved in a task.
· Responsibility link: A responsibility link assigns a role (played by an actor) to a
(soft)goal or task. This link represents who is the stakeholder responsible for a
goal/task accomplishment. It is not necessary that this stakeholder is involved in
the goal sub-tasks. Nevertheless, if the role is responsible for a goal or task, this
role is also responsible for the elements it is divided into, unless a responsibility
link reaches one of the elements it is divided into.</p>
      <p>Goal</p>
      <p>Softgoal
[!!! / !! / !] Task</p>
      <p>Resource
Individual Task</p>
      <p>Awareness</p>
      <p>Softgoal</p>
      <p>D
Dependency Link
[Guard]
Playing Link</p>
      <p>Communication</p>
      <p>Task
Awareness</p>
      <p>Resource
Means-end Link</p>
      <p>Task Decomposition Link
Responsability Link</p>
      <p>1..*
Participation Link</p>
      <p>Collaboration</p>
      <p>Task</p>
      <p>Coordination</p>
      <p>Task
Actor</p>
      <p>Role</p>
      <p>+
Contribution Link</p>
      <p>It is worth noting an additional difference between CSRML and i*: CSRML is
practically hierarchical (see Fig. 2 (a) (d)). Thus, it fosters the scalability of the model
created by using this notation. In a first level, we have the Responsibility diagram, in
which the system's main goal is decomposed into main tasks and quality softgoals.
Also, in this diagram, the goals and tasks responsibilities are defined.</p>
      <p>In a second level appear Task refinement diagrams, in which the system's main
tasks are decomposed into new goals, softgoals, tasks and resources, and roles are
assigned to tasks. This constitutes another difference between CSRML and i*.
Because CSRML has been thought for collaborative systems, i* boundaries for
actors/roles were discarded, since they would not support assigning a task to more
than one role. In addition, the Quality factors diagram completes the system
specification showing the quality softgoals and the elements that contribute to their
accomplishment.
3.1</p>
      <p>Case Study: Collaborative Conference Review System with CSRML
To check out the validity of our proposal, we are going to use the CSRML notation to
model a case study based on a collaborative conference review system in order to
illustrate its expressiveness capacity for CSCW systems. First, in Fig. 2 (a), we can
see the system goals diagram, in which the system main goals are defined. As shown,
we are going to achieve the system goals by means of the realization of the system's
main task: the preparation of the review process of papers for a conference by using
techniques of collaboration among users.</p>
      <p>Fig. 2 (b) shows the responsibility diagram with the main system’s task and its
decomposition in quality softgoals and tasks. In this figure, it can be observed that the
use of responsibility links shows who is responsible for goals and tasks. Note that if a
role is responsible for a goal or task, this role is also responsible for the elements it is
divided into, unless a responsibility link is specified to one of the elements it is
divided into. Also, the playing links are used to represent the condition that must be
met for an actor to play a role. For the sake of model readability, a task decomposition
will be shown in Fig. 2 (c).</p>
      <p>Fig. 2 (c) depicts Papers review task refinement diagram. In this figure, tasks are
refined into more specific ones or new goals, until individual or collaborative
(collaboration, coordination or communication) tasks are specified. It can be observed
that for collaborative tasks, more than an actor (playing a role) is involved through
participation links. This figure includes two awareness softgoals. One of them is
related to the knowledge of who reviews each paragraph, and the other one
corresponds to the use of remote cursors. In this figure, different cardinalities for
participation links are used. For example, for Paragraph review, three experts must
participate. Also, this figure illustrates some degrees of priority that can be assigned
to tasks: normal, high ([!]), very high ([!!]) and highest ([!!!]).</p>
      <p>Finally, Fig. 2 (d) depicts the Quality factors diagram. In this model, the quality
factors that contribute to achieve the conference review with a high quality level are
shown. These factors are represented as softgoals and they are related to the main
quality softgoal by means of contribution links with positive contributions. The
achievement of all these quality softgoals is obtained in different ways. For instance,
the Helpfulness softgoal is achieved by means of an awareness softgoal and its
corresponding awareness resource consists in a remote cursors implementation.</p>
      <p>Conference preparation by</p>
      <p>using techniques of
col aboration among users
Conference
creation</p>
      <p>Papers
submission</p>
      <p>Papers
assignation</p>
      <p>Papers
review</p>
      <p>Ensure compliance with</p>
      <p>quality standards
Chair CP</p>
      <p>Main author</p>
    </sec>
    <sec id="sec-4">
      <title>Conclusions and future work</title>
      <p>
        We found out in two previous works [
        <xref ref-type="bibr" rid="ref1 ref2">1,2</xref>
        ] that Goal-Oriented Requirement
Engineering techniques (and especially i*) can be used to deal with collaborative
systems requirements modelling. Nevertheless, we also found out that this kind of
specifications suffer from an important lack of expressiveness for some characteristics
related to user collaboration, awareness representation or quality factors. To address
these shortcomings, we propose CSRML, an extension of i* Goal-Oriented
specification to model CSCW systems requirements.
      </p>
      <p>In order to check out the suitability of this language, we have modelled a
collaborative system. For the sake of clarity, in this paper an excerpt of this system
consisting in a conference preparation system with collaborative reviews has been
presented. This case study was modelled because it has a set of characteristics that
were hard or impossible to be represented with the original i* notation. These
characteristics were properly described by introducing a set of new elements and links
into i* notation. The quality and awareness representation has been made possible by
means of new awareness elements and the inclusion of a new set of diagrams in order
to provide some structure to the specification.</p>
      <p>
        Resuming, CSRML helps in improving understandability [
        <xref ref-type="bibr" rid="ref3">3</xref>
        ] and maintainability
of requirements models for CSCW systems by adding new elements and relationships
to i*. These new elements facilitate the specification of awareness requirements,
which are paramount in the development process of any CSCW systems.
      </p>
      <p>One of our ongoing works is closely related to the development of e-learning
systems. Since LoUISE research group has been working during the last years in this
kind of systems, several patterns have been described up to date. One of the main
problems they have is that they have been specified in an informal way that cannot be
easily reused for the specification of different systems. Therefore, we are studying
how CSRML can be used to improve their specification.</p>
      <p>Another future work consist in a validation procedure to validate the developed
CSCW system against the initial set of requirements specified with CSRML and his
compliance with the ISO 25010 quality in use factors.</p>
    </sec>
    <sec id="sec-5">
      <title>Acknowledgements</title>
      <p>This work has been supported by the following projects: (PEII09-0054-9581) from
the Junta de Comunidades de Castilla-La Mancha and (DESACO,
TIN2008-06596C02-01) from the Spanish Government.</p>
    </sec>
  </body>
  <back>
    <ref-list>
      <ref id="ref1">
        <mixed-citation>
          [1]
          <string-name>
            <given-names>M.A.</given-names>
            <surname>Teruel</surname>
          </string-name>
          ,
          <string-name>
            <given-names>E.</given-names>
            <surname>Navarro</surname>
          </string-name>
          ,
          <string-name>
            <given-names>V.</given-names>
            <surname>López-Jaquero</surname>
          </string-name>
          ,
          <string-name>
            <given-names>F.</given-names>
            <surname>Montero</surname>
          </string-name>
          , and
          <string-name>
            <given-names>P.</given-names>
            <surname>González</surname>
          </string-name>
          , “
          <article-title>An Empirical Evaluation of Requirement Engineering Techniques for Collaborative Systems,” 15th Int. Conf. on Evaluation and Assessment in Software Engineering</article-title>
          , Durham, UK:
          <year>2011</year>
          .
        </mixed-citation>
      </ref>
      <ref id="ref2">
        <mixed-citation>
          [2]
          <string-name>
            <given-names>M.A.</given-names>
            <surname>Teruel</surname>
          </string-name>
          ,
          <string-name>
            <given-names>E.</given-names>
            <surname>Navarro</surname>
          </string-name>
          ,
          <string-name>
            <given-names>V.</given-names>
            <surname>López-Jaquero</surname>
          </string-name>
          ,
          <string-name>
            <given-names>F.</given-names>
            <surname>Montero</surname>
          </string-name>
          , and
          <string-name>
            <given-names>P.</given-names>
            <surname>González</surname>
          </string-name>
          , “
          <article-title>A Comparative of Goal-Oriented Approaches to Modelling Requirements for Collaborative Systems</article-title>
          ,
          <source>” 6th Int. Conf. on Evaluation of Novel Software Approaches to Software Engineering</source>
          , Beijing, China:
          <year>2011</year>
          .
        </mixed-citation>
      </ref>
      <ref id="ref3">
        <mixed-citation>
          [3]
          <string-name>
            <given-names>M.A.</given-names>
            <surname>Teruel</surname>
          </string-name>
          ,
          <string-name>
            <given-names>E.</given-names>
            <surname>Navarro</surname>
          </string-name>
          ,
          <string-name>
            <given-names>V.</given-names>
            <surname>López-Jaquero</surname>
          </string-name>
          ,
          <string-name>
            <given-names>F.</given-names>
            <surname>Montero</surname>
          </string-name>
          , and
          <string-name>
            <given-names>P.</given-names>
            <surname>González</surname>
          </string-name>
          , “
          <article-title>CSRML: A Goal-Oriented Approach to Model Requirements for Collaborative Systems</article-title>
          ,
          <source>” 30th Int. Conf. on Conceptual Modeling</source>
          , Brussels, Belgium: Springer Berlin,
          <year>2011</year>
          .
        </mixed-citation>
      </ref>
      <ref id="ref4">
        <mixed-citation>
          [4]
          <string-name>
            <given-names>M.A.</given-names>
            <surname>Teruel</surname>
          </string-name>
          ,
          <string-name>
            <given-names>E.</given-names>
            <surname>Navarro</surname>
          </string-name>
          ,
          <string-name>
            <given-names>V.</given-names>
            <surname>López-Jaquero</surname>
          </string-name>
          ,
          <string-name>
            <given-names>F.</given-names>
            <surname>Montero</surname>
          </string-name>
          , and
          <string-name>
            <given-names>P.</given-names>
            <surname>González</surname>
          </string-name>
          , “
          <article-title>Assessing the Understandability of Collaborative Systems Requirements Notations: an Empirical Study,” 1st Int</article-title>
          . Workshop on Empirical Requirements Engineering, Trento, Italy:
          <year>2011</year>
          .
        </mixed-citation>
      </ref>
    </ref-list>
  </back>
</article>