ADAPTIVE CONFIGURATION META-MODEL OF A GUIDANCE PROCESS

International Journal of Computer Science & Information Technology (IJCSIT) Vol 8, No 3, June 2016
DOI:10.5121/ijcsit.2016.8312 157
ADAPTIVE CONFIGURATION META-MODEL OF A
GUIDANCE PROCESS
Hamid Khemissa1
and Mourad Oussalah2
1
Computer Systems Laboratory, Computer Science Institute, USTHB University, Bab
Ezzouar Algeria
2
Computer Laboratory Nantes Atlantique, Faculty of science Nantes University
ABSTRACT
The current technology tend leads us to recognize the need for adaptive guidance process for all process of
software development. The new needs generated by the mobility context for software development led these
guidance processes to be adapted. In order to improve the performance of the deployed software
development, it is useful to manage the configuration of its evolving aspects. This paper deals with the
configuration management of guidance process or its ability to be adapted to specific development
contexts. The proposed adaptive configuration Meta-model is worked out on the basis of a Y description
for adaptive guidance process. This description focuses on three dimensions defined by the
material/software platform, the adaptation form and provided guidance service. Each dimension considers
several factors to develop a coherent configuration strategy and provide automatically the appropriate
guidance process to a current development context.
KEYWORDS
Guidance process, Configuration process, Adaptive meta-model, Development context.
1. INTRODUCTION AND CURRENT TEND
The software development organizations are actually confronted to difficulties regarding the
development of their applications. Due to technological progress, the developer is considered
nowadays as a mobile actor working in various development context using variable platforms.
This trend seems interesting from a user’s perspective, it poses a new problem in software
processes engineering. This concern denotes the adaptation ability to the possible variations of the
development context. The objective is to support the software process by providing software tools
to model, improve, assist and automate development activities [1, 2]. They agree on the following
goals like to facilitate the comprehension and communication process, to describe clearly the
roles, responsibilities and interactions between users, to automate the execution of repetitive tasks
that do not require the human actor intervention and to provide guidance to actors about modeling
and handling a software process. According to the aim and orientation given to the software
process, it is possible that other concepts such as strategy, organization and guidance can be
described in the software process meta-model [2,3,4].
For this, it is necessary to assist developers and to ensure configuration management of the
guidance processes [5, 6] by their ability to adapt to the current development context in respect of
their usefulness. By development context, we mean the triplet (material/software platform,
developer profile, activity context). Usefulness refers to the ability of a guidance process to allow
the developer to reach his objective preserving consistency and product quality in software
development.

158
In this perspective, a rigorous guidance process targets two basic aspects: 1) The progress control
of the software process development regarding the temporal constraints of the activity and the
consistency of the results, and 2) the guidance interventions adapted to the specific needs within
the development context in progress.
Several process-centered environments [7, 8, 9] deal with the guidance aspect in the support of
the software product development. Some Process Centered Software Engineering Environments
(PSEEs) use a guidance description structured in phases like prescribing systems or proactive
systems to control the operations carried out by the developer. Nevertheless, they are essentially
limited to the adaptive guidance aspect to current development context. However, the provided
guidance is not often adapted to the development context profile. The orientations of the guidance
are defined on the basis that the human actor, regardless of his profile (qualifications and
behaviour), has a central role in the progress of the development process.
Among this new generation of software process engineering, we can invoke the following meta-
models and modeling environments: SPEM [10] and APEL [8] considered as the most
representative in the software process modeling, RHODES [7, 11] that uses basic concepts closest
to those introduced in the proposed approach.
SPEM meta-model introduced the concept of "Guidance". According to SPEM, the guidance is a
describable element, which provides additional information to define the describable elements of
modeling. However, the proposed guidance is not suitable to the development context’s profile
(role, qualifications and behaviour). The guidance is rather defined in an intuitive way.
ADELE/APEL proposes a global guidance of proscriptive type without considering the
development context profile and automates part of the development process using triggers.
RHODES/PBOOL+ uses an explicit description of a development process. The activities are
associated to a guidance system with various scenarios of possible realization.
An effective support to software process depends on several factors, in particular the
personalization and adaptation factor. The definition of a process with an active guidance for
automation and coherency control would be effective if it can be adapted to each development
context. The platform, task and developer factors may considerably vary.
Taking into account specific criteria for an adaptive guidance, we have classified these limits
through several criteria describing explicitly the basic concepts linked to the adaptive guidance
[5, 6, 12]. To realize the effectiveness of the configuration concept of the guidance process
supported by its adaptation ability to current development context, we refer to the studied meta-
models and modeling environments [12, 13, 14, 15, 16]. The selected criteria are defined by:
• Global guidance core: The basic guidance is defined as a global orientation core regardless
the profile of both the activity and the actor.
• Developer profile oriented guidance: the guidance orientations are defined on the basis
that the human actor, regardless his profile, has a central role in the progress of the
development process.
• Context development guidance: The selection of the appropriate type of guidance is more
often not adapted nor suitable to a current context.
• Guidance types: the selection of guidance types remains defined in a manual and intuitive
way. It depends on the experience and on the informal personality of the project manager.
• Configuration of guidance process: the guidance functions are defined and offered on the
basis that the human actor always operates in a uniform development context.

159
To respond to these limits, one currently tries to offer more flexibility in the language of software
process modeling. This tendency results in the idea to define interventions of adaptive guidance in
particular contexts during the progress of software process. In considering the principal
limitations of PSEEs and essential characteristics of our approach in particularly the context
adaptation aspect, a comparative table of the studied meta-models is as follows.
Table 1. Comparative table of the studied meta-models.
The current tendency is that developers would like to have integrated environments that are
suitable to specific needs according to the development context factors. However, despite the
necessity imposed by technological evolution, the provided efforts to develop such environments
remain an insufficient contribution. This generation of guidance environment still interests
researchers in defining new concepts and objectives of the software process modeling [4, 17, 18].
Our work proposes an approach to define configuration of an adaptive guidance modeling in
software process. The concepts of proposed approach are described through a meta-model
denoted CGPM (Configuration of Guidance Process Meta-model). The information provided
must be adapted to the development context profile. They must guide the developer during the
software process development through suitable actions and decisions to undertake with
corrective, constructive or automatic intervention [12]. The three dimensions defined by the
development context, the adaptation form and the provided service explicitly describe our
guidance process adaptation.
Section 2 of this paper presents our approach of the Y adaptive guidance modeling. Section 3
describes the configuration management usefulness and the generic view of its operational model,
while section 4 describes the configuration management of the guidance process, as well as the
Configuration of Guidance Process Meta-model (CGPM) and section 5 presents the practical
cases study of the adaptive guidance process. It ends with a conclusion and future prospects.
2. THE ADAPTIVE GUIDANCE IN Y
A guidance process may be processed in many different ways according to the perspective
guidance to provide interveners with development context. Thus, there are generally several
possible guidance models, each of them with a particular relevance and need. This vision denotes
the configuration of guidance process, and its ability to adapt to its development context.
The configuration concept describes its capacity to adapt to the intrinsic variations of required
conditions in terms of usefulness [16, 19, 20].
Meta model
Criteria
ADELE/APEL RHODES / PBOOL+ SPEM
Global guidance core Global Global Global
Developer profile
oriented guidance
Not adapted
Considered strategy
Model
Not adapted
Context development
guidance
Not adapted Adapted Not adapted
Guidance types Not invoked
Associated with a
specific guide system
Intuitive selection
Configuration of
guidance process
Not covered
(Single Platform)
Not covered
(Single Platform)
Not covered
(Single Platform)

160
In this context, we propose a description in Y of the adaptive guidance. This description will
focus on the three considered dimensions. Each dimension considers several factors to deduce
automatically the appropriate guidance service according to the current context. Schematically,
we describe it as follow
Figure 1. Adaptive guidance description in Y.
The principle of our approach is to generate, from the development context related to the specific
data for each defined models according to the retained adaptation form, the guidance
interventions (corrective, constructive, automatic) adapted to the current development context.
2.1. The basic conceptual model description
The conceptual model highlights the guidance process aspect through the adaptation form
described by the inherent relationship between the three considered dimensions.
1. THE DEVELOPMENT CONTEXT
A guidance intervention is provided according to an object or set of objects. An object is
associated to the following:
• The material/software platform: described by the computing resources, the software
services as well as the interaction and communication modes.
• The activity context: models the structure and the workflow, they are defined by a
progression mode in the activity ensuring that all tasks can be performed under control in a
preset order established by the designer and a temporal progression mode specifying
deadlines for completion.
• The developer profile: defines the specific properties of each developer. These properties
can be either static or dynamic. The static aspect refers to the user characteristic as his
role, his business competence and his familiarity with the software process. The dynamic
aspect refers to the behaviour of using the guidance service, by the fact to execute, to
define or to complete the software process resource and the user’s reaction to a guidance
message.
The description performance rate of these factors is evaluated by considering each identified
object as concepts, principles, procedures, and resources. These guidance objects represent the
Adaptation form
+
- Adaptive
- Contextual
- Mixed
- Corrective
- Constructive
- Automatic
Provided service
Development
context
Description ratio
[0,1]
Adaptation ratio
[0, 1]
Description ratio
[0, 1]
Interpretation
sense
- Material/software Platform
- Developer profile
- Activity context

161
basis of different guidance interventions related to a particular situation. This performance serves
as the selection of the adaptation form to retain and guidance service to provide to the user.
2. THE ADAPTATION FORM
Each guidance intervention is done according to the retained adaptation form. It relates to a
specific situation described by the development context description. Our modeling approach
allows the following guidance adaptation forms:
• Contextual guidance: intervention is provided dynamically according to the
material/software platform and activity models. The adaptation rate is related to the
model description rate of the activity and the material/software platform. The guidance
intervention does not consider the developer model (e.g. to avoid inconsistency during
the affectation of a resource).
• Adaptive guidance: intervention is provided according to the developer model and the
material/software platform specificity (e.g.: the user asks for explanations on his choice).
The adaptation rate is related to the developer and material platform models description.
• Mixed guidance: intervention is provided according to the development context (e.g.: to
guide the developer on the sequencing principle during the software process progression).
This form describes the highest adaptation rate. This rate is evaluated on the basis of the
developer, activity and material/software platform models description.
The adaptation form performance is described by a strong coupling between the development
environment and guidance process. It determines the relevance and precision of the guidance
provided to developers.
This criterion is directly related to the adaptive guidance process concept. Through a strong
coupling, the process would deduce the guidance service and can therefore extract useful and
helpful information to the user.
3. THE PROVIDED SERVICE
The guidance process offers several service types in relation to a defined context by the current
development and adaptive form. The provided guidance services are corrective, constructive or
automatic order.
Control and taking corrective initiative: protect the user of his own initiatives when they are
inadequate under progress.
Control and taking constructive initiative: the ability to take positive initiatives, executing
and combining the performance of operations without the user intervention.
The guidance adaptation performance associated to a development environment is done by
enrichment or reduction of the possible offers of the guidance. Among these offers, we have:
• The directive guidance: to show the developer how to execute a task by an adaptive control
of the guidance system, specifying the steps of an activity or the whole process development.
• Retroaction, to offer the developer more information on the activities context (e.g. new
available resources) or on the progress state of his work (progression of an activity).
• Explanation, to offer explanations about a guidance object at the request developer. (e.g. the
activities coordination of the software process).
• Reminding, to remind the developer some principles or procedures on the sequencing of the
activities or their activation conditions when the system detects a conflict or inconsistency.

162
Automatic guidance: analyze the impact projection to define the solution to consider in order
to avoid deadlocks or delays, by the fact to start, suspend, discontinue or continue ongoing
actions to avoid conflict.
These guidance services can be combined. They may be temporary, permanent or left under the
developer control.
The usefulness rate is evaluated by the degree of the performance description of the development
context and adaptation form.
3. WHY PROCESS CONFIGURATION MANAGEMENT
The configuration management is a constant concern of the software development community,
classified as a fundamental discipline to software process development, and necessary for
providing all the time a coherent system support regardless the changes of the different
development context. This management process seeks adapted tools which allow to supervise
any entity evolution under the pre-established norms and rules of a software process. The tools
and techniques of initiating, evaluating and controlling change in the software development
environment where any entity managed in the software engineering process can potentially be
considered under configuration management control. The main purpose of a process management
configuration targets the establishment of the coherent support process synchronizing with
possible evolutions while maintaining the quality continuity of software process development [14,
21].
The configuration management is one of 17 process of ISO 12207 (Software life cycle support
processes) [22]. A specific standard, ISO 10007 is devoted to this process. This management
allows to identify, manage and control configuration items throughout the software development
process cycle. [23]. However, all the existing standards are still powerless to promote the best
practices in the development area.
The dynamic aspect of the configuration management process results from a development
perspective of a successive static description in time. It enables to coordinate the development
activities and supervise each profile entity evolution over the elaboration of software process to
promote coherence with the provided support. In an effective dynamic environment, efficient and
flexible guidance process is a suitable support for modeling and handling software process. The
guidance process management is an overall task supporting the whole lifecycle of software
process. Therefore, a guidance process needs to be planned, adapted and controlled in terms of
user's requirements. The operational process of the guidance is analyzed to identify each time the
guidance needs in order to improve the performance of the deployed software process. This
performance is due to innovations produced by developers, who are expecting to benefit from
using support services. This means that we should develop the configuration that fulfills their
exact needs, leading them to have a trend to benefit from a support solution, to preserve
consistency and product quality in software development [15, 16].
The configuration management plan describes the application scope and key terms. It identifies
the responsible who achieves the planned configuration management activities (Who manages?),
identifies the management activities to be performed (What activities?). Then identifies tools and
techniques required for the configuration planning (HOW planned?), establishes the sequence and
coordination of the management configuration activities (When scheduled?). Finally, identifies
activities and responsibilities on how the SCMP will be held active during the development
lifecycle. As regard the implementation of its process, the required information should be
collected by identifying the configuration items to be applied to the configuration management
and its application constraints. As well as, to identify the need for a change and how to provide a

163
service support. Moreover, assumptions that might have an impact on schedule and ability to
perform defined configuration activities [24].
The operational model of a configuration process is based on the fact that the software
configuration management plan might be extended or revised during the rest of the project
following any change at the developers’ request or by a trigger event that is assessed against the
project goals.
This configuration process is an application developed using a configuration server and based on
a set of constraints and / or preferences described by the characteristics of the current
development context. If one of these constraints and / or preferences causes a change or an
evolution notified event then it might mean that we do evolve the current configuration either
statically (inactive process) or dynamically (without interrupting the process progress). The
purpose of this activity is proceeding, every time, adaptation of the process support to the various
changes in the development context in order to ensure its deployment to software process. The
functional model description of the configuration method is given in the following figure:
Figure 2. The operational model of the configuration process.
4. THE CONFIGURATION MANAGEMENT OF A GUIDANCE PROCESS
The management discipline within the software engineering process allows to develop a reference
basis to any planned configuration. The configuration management is a complex activity that
covers all the steps in a software process, from the conceptual aspects to its configuration and
installation. It includes the initiation techniques and changes control that manage the software
process progress [14, 15, 16].
Process
support
Event
Configuration
Management
Configuration
server
Development
context
Uses
Is based
Configures
Adapted
system « Ready state »
Activates
Deployment
Evaluate
Change
Apply change
Developer’s
request
Occurs
Notifies change
Provides
Results

164
The role of configuration management is to identify the configuration items, manage the sequence
and ensure the evolution and any likely change. These activities are usually treated in different
ways depending on the nature and specificity of the objective [15, 16, 25, 26].
In this context, the configuration management of a software process and particularly a guidance
process is a configuration mechanism rigorously linked to the evolution of all development
context factors. This evolution process is influenced by the temporal and contextual elements.
Therefore, it is necessary to define the participating elements and trigger the configuration of a
guidance process, and clarify their functional interdependencies.
We describe graphically the configuration basis of our guidance process defined through five
interrelated elements in figure 3:
Figure 3. The configuration basis of guidance process.
This evolution basis allows, each time, to answer any manifest changes provoked statically or
dynamically in order to design the most appropriate guidance process related to the current
situation. The configuration management of guidance process is useful both for time optimizing
and for the guidance process quality.
The (Re) configuration is invoked by the basic guidance process actors, namely the guidance
process designer or at the user’s request, with different interests ensuring an equilibrium between
the user’s needs in terms of guidance, with either static adaptation at the guidance conceptual
level or dynamic process related to the development context change. The following illustration
defines the basis description of the guidance invocation
Why
Designer PG
User
Dynamic
Static
Event
« Context Change »
At the user’s
initiative
Configuration
Quality
Time optimizing
When
CONFIGURATIO
N BASIS
How
Who
What
Update
Configuration&
Selection
User initiative
UserContext change
Development context
Platform User profile Activity context

165
Figure 4. The basis description of the guidance invocation.
The involved configuration process deployment represents the core of the configuration
management. It describes the steps and the contextual sequencing that should be undertaken to
ensure this goal. Therefore, the deployment cycle of a guidance process configuration is
explained via all activities related to the configuration management and their functional
interdependencies (see figure 5.).
Figure 5. The guidance process configuration.
The first activity concerns the configuration and selection of the most coherent
composition, which best responds to the defined development context. Its objective is to
offer the most appropriate guidance to support the current situation.
For this, we should first deal with the dependencies between the guidance process and
description of the development context, to specify the expected guidance service and
constraints to be respected.
The transfer and activation activity consists in transferring the guidance support. It is
invoked at the initiative of the guidance process designer or at the request of the user.
Once installation and transfer problems solved, the activation operation role is to put the
guidance support in the active state, ready to be executed.
The update activity is to resume, adjust or to complete the guidance support
configuration. It is invoked at the user’s initiative or by the guidance process designer
following a change in the development context.
(Re) Configuration
&
Selection
Support end
Update
Deactivation
Transfer
&
Activation

166
This activity is almost similar to the (re) configuration, with the difference that the
guidance support is already installed. Finally, the implementation of the update uses the
deactivation activity.
The ''support end'' activity is relative to the end of guidance service offered to users of the
software process in progress.
Deactivation activity is essential before any configuration, update or support end
activities. Its role is to deactivate the support guidance component so that we can perform
the likely changes of some support components
These activities and their functional interdependencies give a rough estimation on the complexity
of the of the guidance process configuration management. This process aims to set up, in respect
of the development context constraints, a first core adaptation of guidance support. Subsequently,
it will have to activate the dynamic adaptation operation with regard to the likely evolutions of
the development environment.
4.1. THE CONFIGURATION OF GUIDANCE PROCESS META MODEL
Our modeling approach CGPM (Configuration of Guidance Process Meta model) is defined
with reference to studied PSEEs features. The aim of our approach is to better respond to the
factors of the development context defined by the three dimensions, the development context,
adaptation form and offered guidance. Each dimension is described through several basic factors
to develop a coherent configuration strategy and provide the most appropriate guidance support to
the current development context.
In this context, our meta-model is based on a conceptual model of a software process enriched by
a configuration model adapted to the guidance process. The configuration model detects any
event of the development context evolution and launches the configuration of guidance process.
The role of the first activity "configuration and selection" is to prepare the guidance support to the
current situation. The second service “transfer and activation” is to put the guidance support in
the active state. The third service "Update" is to provide a way to modify and / or complete the
guidance configuration following a development context change.
The configuration strategy evolves according to the political autonomy given to the guidance
system respecting the application conditions. The implementation of this policy is based on an
adaptation mode expressed by a set of rules of ECA form (Event, Conditions, Actions). For each
execution in the development context, if required conditions related to the context and the
adaptation form then launch guidance configuration strategy to generate the most appropriate
guidance service.
The proposed meta-model aims at generating the adapted guidance interventions to the
development context in relation to the considered factors and specific data for each defined
dimension (see Figure 6)

167
.
Figure 6. Configuration of Guidance Process Meta model.
5. THE PRACTICAL INTERPRETATION
Considering the software process model "Activity test", the process "Activity test" in the software
development is composed of several types of tests such as; Integration test and Unitary test. Each
receives as input a test plan and provides a test report. For each type of test, there is a manager,
responsible of the execution.
A performing tree given in Figure 7 describes the activity process «Activity test». We notice that
the activity test starts the execution of sub activities "Unitary test” then “Integration test". The
unitary test launches in parallel the execution of tasks "Test unit"
Development Context
Performance
Adaptation Form
Performance
Event
GuidanceManagement
Guidance Service
Performance
1 : Use
2 : Select
3 : Provide
Configuration
Strategy
Rules
« E.C.A » Consider
Launch
Caused by
Astrategyfor
Material/software
Platform
Developer profile
Activity context
ConditionsActions
Generate
Corrective
Constructive
Automatic
Adaptive
Contextual
Mixed
Update
Configuration
End support

168
Figure 7. The activity test process.
To simplify our example, we consider the execution process of the unitary component test. The
application of the activity “Unitary test", requires the list of components. It calls the tool that will
create the necessary environment to carry out the actual execution of the “Unitary test", as the
state diagram, the test variables, etc. ... the activity "Unitary test" launches in parallel the different
tasks "Test unit" where an event signals the beginning of the “Test unit” execution. Finally, the
ended event is broadcast.
The adaptive execution process of the activity "Unitary test", regarding our adaptive guidance
approach is described in Figure 8.
Figure 8. The adaptive execution process “ Unitary test”.
The adaptive guidance is linked to the manager or to each tester according to the current
development context profile defined by its material/software platform, the activity context and
developer profile. We explain this adaptive approach through the following situation; the testers
have the same role “test unit” with identical activity model. However, the developer’s
qualification and the material/software platform specificity differ from one development context
to another. According to the current context, it would require enrich or reduce the appropriate
guidance intervention or generate several possible forms of guidance services.
We consider three situations with tester’s qualification defined respectively as high, medium, and
low. The study case is related to launch the test unit without having all the input data, by selecting
the appropriate test variables and generating the unit test report. The adaptive guidance process
related to each qualification case is described as follows:
1. For a development context with high qualification tester and a high material platform
performance: the tester starts the test unit process on the basis of the defined plan by taking his
proper initiatives. The development context evaluation allows deducing the adaptation form to
Ordered
control flow
3: 3:3: launch
4: launch2: launch
Activity test
Unitary test Integration test
Test unit Test unit
1: launch
Parallel
control flow
Test unit
3: 3:
Test unit Test unit
Parallel
control
3:
Unitary test
AdaptiveGuidance
Tester
AdaptiveGuidance
AdaptiveGuidance
Tester Tester
AdaptiveGuidance
UT
Manager
Test unit
Activityperfor
mer+
Unitary test
+ControledActivit
y *
*1
*
*
+ ProvideguidanceAdaptiveGuidance
Test
*
Manager
Test unit
Test unit
Test unit
Taskperform
er
+
*1
Tester

169
retain and the guidance service to provide. In this case, we adopt the adaptation guidance form
and the provided guidance intervention is thus of a corrective order. The corrective
intervention is provided to inform the manager of the setback and remind him of the
corresponding unitary test diagram. The manager remains free to take into account the
intervention.
2. For a development context with an average skill tester and an acceptable material/software
platform performance: the tester starts the test unit process by applying rigorously the defined
test plan. The evaluation of such context results in a contextual guidance form and the
provided guidance intervention is thus of a constructive order. The guidance system analyzes
the current context of the task, evaluates the impact and consequence of the delay caused in
comparison with possible margins and offers a possible solution to the manager (solution: the
guidance proposes to cancel the launch of the current test unit and generates a new execution
plan according to the rate of delay and possible margins). The construction solution is not
definite; the manager should validate it.
3. For a development context with a low qualification tester and a reduced material/software
platform performance: the tester starts the test unit process by applying reliably the defined
test plan. The development context evaluation results in a mixed guidance adaption form and
the provided guidance intervention is thus of an automatic order. The guidance system
analyzes the current context, cancels the launch of the “test unit” task, evaluates the impact
and consequence of the delay caused in comparison with the possible margins and
automatically updates the execution plan of "unitary test" activity.
5.1. The digital application
The practical definition of the adaptive guidance type for each considered profile is deduced
through a quantitative process of the factors in relation to the basic models (material/software
platform, activity context, developer profile). The considered example is processed as follows :
Each profile is semantically described in table (see Table 2). The project manager under the
specification of an ongoing project [17] determines the semantic evaluation and the weighting. To
scan the semantics evaluation, we associate the weighting related to the interest granted to each
attribute.
Table 2. The profiles evaluation.
Development
context
Factors
Context
Profile 1
Context
Profile 2
Context
Profile 3
Context
Profile 4
Context
Profile 5
W[i]
Material/
software
Platform
Development
System Constraint
Low Medium High High Low P2
Software Tools Low Medium Low Provided Provided P1
Memory
Constraint
High Medium Medium High Medium P3
Developer
Profile
Role No effect Classic Critique Critique No effect P4
Competence High Medium Low Low High P1
Familiarity with
Software Process
Quite
Acceptable
Medium Low Low Acceptable P1
Behavior for
guidance
Most
Appropriat
e
Satisfying
Inadequat
e
Adequate Inadequate P2
Activity
Context
Density of tasks
in the activity
Acceptable
Acceptabl
e
Acceptabl
e
Acceptabl
e
Acceptable P3
Complexity Level Medium Medium Medium Medium Medium P2
With W[i]∈∈∈∈ [1, 5]. Where Pi represents the computing value.

170
Considering the similar principle such as the COCOMO model [26], the quantification of each
profile’s feature is on the data range] 0, 2 [, (see Table 3). This quantification is usually based on
the impact of each feature.
It is usually done through three levels, described by high, medium or low contribution, applying
the following rules:
1: middle order impact / <1: positive impact / >1: negative impact.
Table 3. The profiles quantification.
Development
context
Factors
Context
Profile 1
Context
Profile 2
Context
Profile 3
Context
Profile 4
Context
Profile 5
W[i]
Material/
software
Platform
Development
System
Constraint
0.75 1.00 1.30 1.30 0.75 1
Software
Tools 1.20 1.00 1.25 0.80 0.80 3
Memory
Constraint 1.40 1.00 1.00 1.60 1.00 2
Developer
Profile
Role 0.40 1.00 1.90 1.70 0.40 4
Competence 0.20 1.00 1.70 1.70 0.25 3
Familiarity
with
Software
Process
0.40 1.00 1.60 1.60 0.30 3
Behavior for
guidance 0.20 0.80 1.70 0.75 1.60 1
Activity
Context
Density of
tasks in the
activity
0.80 0.80 0.80 0.80 0.80 2
Complexity
Level 1.00 1.00 1.00 1.00 1.00 1
In this stage of profiles’ process, and in case of simple profiles’ samples, we can proceed to
associate each considered development context profile to the appropriate guidance adaptation
form and guidance service.
The guidance profile (GP) associated to each profile class is based on the following formula:
GP (Px) = Σ Ai Wi / 2*Σ Wi avec i=1 to n
With :
Ai: the feature value.
Wi: the associated weighting.
Px : the associated profile.
The adaptation form and the guidance profile of each considered development context profile
based on the evaluation of each model and GP value is given by (see Table 4).

171
Table 4. The associate guidance profile.
Context
Profile 1
Context
Profile 2
Context
Profile 3
Context
Profile 4
Context
Profile5
Associated
Adaptation
form
Adaptive Contextual Mixed Mixed Adaptive
Guidance
Profile (GP)
0.333 0.485 0.721 0.673 0.315
Associated
guidance profile
Corrective Constructive Automatic Automatic Corrective
It should be noted that the value of GP ranged from 0 to 1 and the range associated with each type
of guidance is defined by the fixed limits to each guidance type. If the range of corrective
guidance is fixed between 0 and 0.35 and the range of the constructive guidance is between 0.36
and 0.65, we automatically associate a corrective guidance to profile P1 and P5, a constructive
guidance to profile P2 and automatic guidance to profile P3and P4.
However, in case of a very important population, and for the aim of optimizing profile classes, it
is recommended to proceed in the gathering and classification of the provided development
profile and reasoning in relation to generated classes.
6. CONCLUSION
Our main purpose in this article is to propose a configuration management of a guidance process
for software process modeling. This configuration is highlighted through a Y description of our
adaptive guidance. This description will focus on three dimensions defined by the
material/software platform, the adaptation form and the provided service. Each dimension
considers several factors to deduce automatically the appropriate guidance process according to
the current development context. The proposed approach concepts are described through a meta-
model denoted CGPM (Configuration of Guidance Process Meta model). The proposed meta-
model aims to generate the adapted guidance support to the development context in relation to the
considered properties and specific data for each defined model.
The contribution of this approach is to institute the configuration management concepts, in order
to be able to proceed, every time, with the adaptation of the guidance support to the various
changes in the development context to ensure its adaptive deployment to software process.
The configuration strategy for the adaptation is done by the dynamic deployment of the provided
guidance process. Intuitively, we consider the defined guidance process configuration through the
sequencing of contextual activities. The guidance strategy evolves according to the political
autonomy given to the guidance process respecting the application conditions. The
implementation of this policy is based on an adaptation mode expressed by a set of rules of ECA
form.
A perspective to this work concerns, at first, the necessity to estimate the productivity and cost
due to the adaptation of guidance system.
In a second step, we will ensure the development of semantic rules that allow swapping through
different guidance profiles, either statically by adjusting the guidance parameters or dynamically
through the development context changes.

172
REFERENCES
[1] Ivan Garcia and Carla Pacheco « Toward Automated Support for Software Process Improvement
Initiatives in Small and Medium Size Enterprises ». Book chapter. Software Engineering Research,
Management and Applications 2009 Volume 253/2009, pp. 51–58. c_ Springer-Verlag Berlin
Heidelberg 2009. ISBN: 978-3-642-05440-2.
[2] Kirk, D.C, MacDonell, S.G., & Tempero, E. 2009 Modeling software processes - a focus on
objectives, in Proceedings of the Onward, 2009. Conference. Orlando FL, USA, ACM Press, pp.941-
948.
[3] Benoît COMBEMALE, Xavier CRÉGUT, Alain CAPLAIN et Bernard COULETTE. Towards a
rigorous process modeling with spem. Dans ICEIS (3), pages 530–533, 2006
[4] Hamid Khemissa, Mohamed Ahmed-Nacer, Mourad Daoudi, 2008. A Generic assistance system of
software process. In International Conference on Software Engineering: Software Engineering. SE
2008, Feb 12-14-2008, Innsbruck, Austria.
[5] Calvary, G., Coutaz, J., Thevenin, D., Limbourg, Q., Souchon, N., Bouillon, L., Florins, M.,
Vanderdonckt, J.: Plasticity of User Interfaces: A Revised Reference Framework. In: TAMODIA
2002 (2002).
[6] Joëlle Coutaz, EICS '10. User interface plasticity: model driven engineering to the limit!.
Proceedings of the 2nd ACM SIGCHI symposium on Engineering interactive computing systems.
June 2010.
[7] Coulette B., Crégut X., Dong T. B. T. and Tran D. T., “RHODES, a Process Component Centered
Software Engineering Environment”, ICEIS2000, 2nd International Conference on Enterprise
Information Systems, Stafford, pp 253-260, July 2000.
[8] Jacky Estublier, Jorge Villalobos, Tuyet Le Anh, Sonia Jamal-Sanlaville and German Vega. An
Approach and Framework for Extensible Process Support System. In Proceedings 9th European
Workshop on Software Process Technology (EWSPT 2003), Helsinki, Finland, 2003-09-01.
[9] Hans-Ulrich Kobialka, « Supporting the Software Process in A Process-centered Software
Engineering environment », Upgrade-cepis.org/issues/2004/5/upgrade-v VOL; V n° 5 October 2004.
[10] OMG. Inc. Software and System Process Engineering Meta-Model Specification version 2.0:
Formal/2008-04-01.
[11] Tran Hanh Nhi, Bernard Coulette, Xavier Crégut, Thuy Dong Thi Bich, Thu Tran Dan. Modélisation
du méta-procédé RHODES avec SPEM. Dans : Recherche Informatique Vietnam-Francophone
(RIVF'03), Hanoi, Vietnam, 2003.
[12] Hamid khemissa, Mohamed ahmed nacer & Mourad Oussalah «Adaptive Guidance System for
SPEM ». The First International Conférence on Information Technology Convergence and Services;
ITCS, SIP, JSE 2012 pp. 429-441, Bangalore, India.
[13] Hamid Khemissa, Mohamed Ahmed-Nacer, Mourad Oussalah «Adaptive Guidance based on
Context Profile for Software Process Modeling». Information Technology and Computer Science, ,
July 2012 in MECS 2012.
[14] C. Lueninghoener. "Getting Started with Configuration Management. ;login: issue: April 2011,
Volume 36, Number 2" (PDF). Retrieved 2012-11-23.
[15] IEEE Std-828-2012 IEEE Standard for Configuration Management in Systems and Software
Engineering. 2012. doi:10.1109/IEEESTD.2012.6170935. ISBN 978-0-7381-7232-3.
[16] Ali, U., Kidd, C. Barriers to effective configuration management. Application in a project context:
An empirical investigation, International Journal of Project Management, 2013b;
http://dx.doi.org/10.1016/j.ijproman.2013.06.005.
[17] Grambow, Gregor and Oberhauser, Roy and Reichert, Manfred (2011) Enabling Automatic Process-
aware Collaboration Support in Software Engineering Projects. In: Selected Papers of the
ICSOFT'11 Conference. Communications in Computer and Information Science(CCIS).
[18] Clarke, Paul and O'Connor, Rory (2011) An approach to evaluating software process adaptation. In:
11th International SPICE Conference on Process Improvement and Capability dEtermination, 30
May - 1 jun 2011, Dublin, Ireland. ISBN 978-3-642-21233-8.
[19] Sottet, J.-S., Calvary, G., Coutaz, J., Favre, J.-M. A Model-Driven Engineering Approach for the
Usability of User Interfaces. In Proc. Engineering Interactive Systems (EIS2007), J. Gulliksen et al.
(eds), LNCS 4940, (2007), 140-157

173
[20] Ferry, N. Hourdin, G., Lavirotte, S., Rey, G., Tigli, J.- Y., Riveill, M. Models at Runtime: Service
for Device Composition and Adaptation. In 4th International Workshop Models@run.time, Models
2009(MRT09).
[21] Leif Singer, Fernando Figueira Filho, Margaret-Anne Storey (2014) Software engineering at the
speed of light: how developers stay current using twitter. 36 th International Conference on Software
Engineering, ICSE May 2014. Pages 211-221 ACM New York, USA 2014, ISBN: 978-1-4503-
2756-5.
[22] International Standard ISO/IEC 12207 : (2008). Systems and software engineering -- Software life
cycle processes.
[23] International Standard ISO 10007:2003 for Configuration Management
[24] IEEE Standard 828-2012 - for Configuration Management in Systems and Software Engineering
[25] Lindkvist, C., A. Stasis, and J. Whyte, Configuration Management in Complex Engineering Projects.
Procedia CIRP, 2013, Volume 11(0): p. 173-176.
[26] Barry W. Boehm, Chris Abts, A. Winsor Brown, Sunita Chulani, Bradford K. Clark, Ellis Horowitz,
Ray Madachy, Donald J. Reifer, Bert Steece, 2009. Software Cost Estimation with COCOMO II.
Prentice Hall Edition, ISBN: 0137025769, 978013702576.
Authors’ profile
Hamid Khemissa is a full associate professor at Computer Systems Department, Faculty of Electronics and
Computer Science, USTHB University, Algiers. He is member of the software engineering team at
computer system laboratory LSI, USTHB. His current research interests include Software Process
Modeling and Software Modeling Guidace.
Mourad Chabane Oussalah is a full Professor of Computer Science at the University of Nantes and the
chief of the software architecture modeling Team. His research concerns software architecture, object
architecture and their evolution. He worked on several European projects (Esprit, Ist, ...). He is (and was)
the leader of national project (France Telecom, Bouygues telecom, Aker-Yard-STX, ...). He earned a BS
degree in Mathematics in 1983, and Habilitation thesis from the University of Montpellier in 1992.

ADAPTIVE CONFIGURATION META-MODEL OF A GUIDANCE PROCESS

More Related Content

What's hot (20)

Similar to ADAPTIVE CONFIGURATION META-MODEL OF A GUIDANCE PROCESS (20)

Recently uploaded (20)

ADAPTIVE CONFIGURATION META-MODEL OF A GUIDANCE PROCESS