ChapterPDF Available

Abstract and Figures

The business environment is rapidly undergoing changes, and they need a prompt adaptation to the enterprise business systems. The process models have abstract behaviours that can apply to diverse conditions. For allowing to reuse a single-process model, the configuration and customization features can support the design improvisation. However, most of the process models are rigid and hard coded. The current proposal for automatic code generation is not devised to cope with rapid integration of the changes in business coordination. Domain-specific rules (DSRs) constitute to be the key element for domain-specific enterprise application, allowing changes in configuration and managing the domain constraint within the domain. In this paper, the key contribution is conceptualization of the domain model, domain model language definition and specification of domain model syntax as a source visual modelling language to translate into domain-specific code. It is an input or source for generating the target language which is domain-specific rule language (DSRL). It can be applied to adapt to a process constraint configuration to fulfil the domain-specific needs.
Content may be subject to copyright.
Domain Model Definition
for Domain-Specific Rule Generation
Using Variability Model
Neel Mani, Markus Helfert, Claus Pahl, Shastri L Nimmagadda
and Pandian Vasant
Abstract The business environment is rapidly undergoing changes, and they need
a prompt adaptation to the enterprise business systems. The process models have
abstract behaviors that can apply to diverse conditions. For allowing to reuse a single
process model, the configuration and customisation features can support the design
improvisation. However, most of the process models are rigid and hard coded. The
current proposal for automatic code generation is not devised to cope with rapid
integration of the changes in business coordination. Domain-specific Rules (DSRs)
constitute to be the key element for domain specific enterprise application, allowing
changes in configuration and managing the domain constraint with-in the domain. In
this paper, the key contribution is conceptualisation of the do-main model, domain
model language definition and specification of domain model syntax as a source
visual modelling language to translate into domain specific code. It is an input or
source for generating the target language which is do-main-specific rule language
(DSRL). It can be applied to adapt to a process constraint configuration to fulfil the
domain-specific needs.
N. Mani ()M. Helfert
School of Computing, ADAPT Centre for Digital Content Technology,
Dublin City University, Dublin, Ireland
M. Helfert
C. Pahl
Faculty of Computer Science, Free University of Bozen-Bolzano, Bolzano, Italy
S. L. Nimmagadda
School of Information Systems, Curtin Business School (CBS), Perth, WA, Australia
P. Vasan t
Department of Fundamental and Applied Sciences, Universiti Teknologi PETRONAS,
Perak, Darul Ridzuan, Malaysia
© Springer International Publishing AG 2018
I. Zelinka et al. (eds.), Innovative Computing, Optimization and Its
Applications, Studies in Computational Intelligence 741,
40 N. Mani et al.
1 Introduction
We are primarily concerned with the utilisation of a conceptual domain model
for rule generation, specifically to define a domain-specific rule language (DSRL)
[1,2] syntax, its grammar for business process model and domain constraints man-
agement. We present a conceptual approach for outlining a DSRL for process con-
straints [3]. The domain-specific content model (DSCM) definition needs to consider
two challenges. The first relates to the knowledge transfer from domain concept to
conceptual model, where model inaccuracies and defects may have been translated
because of misunderstandings, model errors, human errors or inherent semantic mis-
matches (e.g., between classes). The other problem relates to inconsistency, redun-
dancy and incorrectness resulting from multiple views and abstractions. A domain-
specific approach provides a dedicated solution for a defined set of problems. To
address the problem, we follow a domain model language approach for developing
a DSRL and expressing abstract syntax and its grammar in BNF [4] grammar. A
domain-specific language (DSLs) [57] refers to an approach for solving insufficient
models by capturing the domain knowledge in a domain-specific environment. In
the case of semantic mismatches/defects, a systematic DSL development approach
provides the domain expert or an analyst with a problem domain at a higher level of
DSL is a promising solution for raising the level of abstraction that is easier to
understand or directly represent and analyse, thus, attenuating the technical skills
required to develop and implement domain concepts into complex system develop-
ment. Furthermore, DSLs are either textual or graphical language targeted to specific
problem domains by increasing the level of automation, e.g. through rule and code
generation or directly model interpretation (transform or translate), as a bridge, fill-
ing the significant gap between modeling and implementation. An increase in effec-
tiveness (to improve the quality) and efficiency of system process is aimed at rather
than general-purpose languages that associated with software problems. Behavioral
inconsistencies of properties can be checked by formal defect detection methods
and dedicated tools. However, formal methods may face complexity, semantic corre-
spondence, and traceability problems. Several actions are needed for implementation
of any software system. These are from a high-level design to low level execution.
The enterprises typically have a high level of legacy model with various designs
in a domain or process model. Automatic code generation [811]isawell-known
approach for getting the execution code of a system from a given abstract model. The
Rule is an extended version of code since code requires compiling and building, but
the rule is always configurable. Rule generation is an approach by which we trans-
form the higher-level design model as input and the lower level of execution code as
output. It manages the above mentioned constraint.
We structure the modelling and DSL principles in Sect. 1. In Sect. 2, we discuss
the State of-the-Art and Related Work. We give an overview of the global intelligent
content processing in a feature-oriented DSL perspective, which offers the domain
model and language definition in Sect. 3. Then, we describe the ontology-based con-
Domain Model Definition for Domain-Specific Rule Generation . .. 41
ceptual domain model in Sect. 4. The description of the domain model and language
expressed in terms of abstract and concrete syntax are given in Sect.5. As a part
of DSRL, the general design and language are presented in Sect. 6. Section 7pro-
vides details regarding implementation of a principal architecture of DSRL genera-
tion and how it translates from the domain model to the DSRL. We discuss analy-
sis and evaluation of the rule in Sect. 8. Finally, we conclude our work with future
scope. Throughout the investigation, we consider the concrete implementation as a
software tool. However, a full integration of all model aspects is not aimed at, and the
implementation discussion is meant to be symbolic. The objective is to outline the
principles of a systematic approach towards a domain model used, as source model
and the domain specific rule language, as a target for content processes.
2 Related Work
The web application development is described as a combination of processes,
techniques and from which web engineering professionals make a suitable model.
The web engineering is used for some automatic web application methodologies
such as UWE [12], WebML [13] and Web-DSL [14] approaches. The design and
development of web applications provide mainly conceptual models [15], focus-
ing on content, navigation and presentation models as the most relevant researchers
expressed in [16,17]. Now, the model driven approach for dynamic web applica-
tion, based on MVC and server is described by Distante et al. [18]. However, these
methods do not consider the user requirement on the variability model. To simplify
our description, we have considered the user requirement and according to the need
of the user, the user can select the feature and customize the enterprise application
at the dynamic environment.
A process modeling language provides syntax and semantics to precisely define
and specify business process requirements and service composition. Several graph
and rule-based languages have been emerged for business process modeling and
development, which rely on formal backgrounds. They are Business Process Mod-
eling Notation (BPMN) [19], Business Process Execution Language (BPEL)/WS-
BPEL, UML Activity Diagram Extensions [20], Event-Driven Process Chains (EPC)
[21], Yet Another Work ow Language (YAWL) [22], WebSphere FlowMark Defin-
ition Language (FDL) [23], XML Process Definition Language (XPDL) [24], Java
BPM Process Definition Language (jPDL) [25], and Integration Definition for Func-
tion Modeling (IDEF3) [26]. These languages focus on a different level of abstraction
ranging from business to technical levels and have their weaknesses and strengths
for business process modeling and execution. Mili et al. [27] survey the major busi-
ness process modeling languages and provide a brief comparison of the languages, as
well as guidelines to select such a language. In [28], Recker et al. present an overview
of different business-process modeling techniques. Among the existing languages,
BPMN and BPEL are widely accepted as de facto standards for business process
design and execution respectively.
42 N. Mani et al.
Currently, there is no such type of methodology or process of development for
creating a rule-based system in a web application (semantic-based). Diouf et al. [29,
30] propose a process which merges UML models and domain ontologies for busi-
ness rule generation. The solution used for semantic web has ontologies and UML;
to apply to the MDA approach for generating or extracting the rules from high level
of models. Although, the proposed combination of UML and semantic based ontolo-
gies is for extracting the set of rules in target rule engine, they only generate the first
level of the abstraction of the rules.
Our approach provides the systematic domain-specific rule generation using vari-
ability model. The case study uses intelligent content processing. Intelligent content
is digital that provides a platform for users to create, curate and consume the con-
tent in dynamic manner to satisfy individual requirements. The content is stored,
exchanged and processed by a dynamic service architecture and data are exchanged,
annotated with metadata via web resources.
3 Business Process Models and Constraints
We use the intelligent content (IC) [31] processing as a case study in our applica-
tion. The global intelligent content (GIC) refers to digital content that allows users
to create, curate and consume content in a way that fulfills dynamic and individual
requirements relating to information discovery, context, task design, and language.
The content is processed, stored and exchanged by a web architecture and the data
are revised, annotated with metadata through web resources. The content is deliv-
ered from creators to consumers. The content follows a certain path that consists
of different stages such as extraction and segmentation, named entity recognition,
machine translation, quality estimation and post-editing. Each stage, in the process,
comes with its challenges and complexities.
The target of the rule language (DSRL) is an extendable process model notation
for content processing. Rules are applied at processing stages in the process mode.
The process model that describes activities remains at the core. It consists of
many activities and sub-activities of reference for the system and corresponds to the
properties for describing the possible activities of the process. The set of activities
constitutes a process referred to as the extension of the process and individual activ-
ities in the extension are referred as instances. The constraints may be applied at
states of the process to determine its continuing behaviour depending on the current
situation. The rules combine a condition (constraint) on a resulting action. The target
of our rule language (DSRL) is a standard business process notation (as shown in
Fig. 1).
The current example is a part of digital content (processing) process model as
shown in Fig. 1, a sample process for the rule composition of business processes
Domain Model Definition for Domain-Specific Rule Generation . .. 43
Fig. 1 Process model of global intelligent content
and domain constraints that conduct this process. The machine language activ-
ity translates the source text into the target language. The translated text quality
decides whether further post-editing activity is required. Usually, these constraints
are domain-specific, e.g., referring to domain objects, their properties and respective.
4 Ontology-Based Conceptual Domain Model
We outline the basics of conceptual domain modelling, the DSRL context and its
application in the intelligent content context. The domain conceptual models (DCM)
are in the analysis phase of application development, supporting improved under-
standing and interacting with specific domains. They support capturing the require-
ments of the problem domain and, in ontology engineering. A DCM is a basis for the
formalized ontology. There are several tools, terminologies, techniques and method-
ologies used for conceptual modelling, but DCMs help better understanding, repre-
senting and communicating a problem situation in specific domains. We utilise the
conceptual domain model to derive at a domain-specific rule language.
A conceptual model can define concepts concerning a domain model for a DSL,
as shown in the class model in Fig. 2, which is its extended version described in
[3]. A modeling language is UML-based language defined for a particular domain,
defining relevant concepts as well as a relation (intra or inter-model) with a meta-
model. The metamodel consists of the concrete syntax, abstract syntax and static
semantics of the DSL. The abstract syntax defines modelling element such as classes,
nodes, association, aggregation and generalisation, and relationships between the
modelling elements [3].
A DSRL reuses domain model elements or define new modelling element depend-
ing on the domain concept and its relations with other elements. Modelling of ele-
ments is done with two fundamental types: concepts and relations. The concepts are
44 N. Mani et al.
Fig. 2 A domain model based DSRL concept formalization
based on domain concepts such as entity, state, action, location, risk, menu and rela-
tions are used to connect elements. The relations are also of two types: generalisation
and association. Aggregation is a special type of an association. The association has
domain-specific relations like conditional flow, multiplicity and aggregation. Fur-
thermore, relations may have properties such as symmetry, reflexivity, equivalence,
transitivity and partial order.
5 Language Definition of a Domain Model
Domain model serves as the very basis of all types of Business Applications that run
on a Domain, both individual as well as Enterprise applications. The objective is to
define the language for Domain Model and recognises the internal data structures or
schema used in it. It is easy to transform or translate the graphical Domain Model into
Textual Rule language for a particular domain. In this scenario, the objective data
structure refers to the storability of a domain model in a vulnerable environment, as
in the case with a rule language. This is because the target of a language for mapping
the translated domain model’s knowledge into XML schema of DSRL that follows
the rule paradigm.
Domain Model Definition for Domain-Specific Rule Generation . .. 45
1Domain ::= <Domain model> Domain definition
2Concept ::= <Concept> Concept definition
3Class ::= <Attributes>,< Operation>, < Receptions>,
<Template Parameters>, < Component>,
<Constraints>, <Tagged Values>
Class Definition
4<Relations> ::= <Association>|<DirectedAssociation>|<Reflexive
Class relationships
5<Association> ::=
| ‘
| ‘::’
Structural relationship
between objects (classes)
of different type
6<Type> ::= <BuiltinType>|<UCase Ident >|<EnumType> Domain model type
concept type or extended
type enumeation type
list type
7<PrimitiveTypes > <String>,<Integer>,<Boolean>, ...,<Date> Domain model primitive (
built-in) types
Fig. 3 Syntax definition of domain model language
5.1 Language Description
Metamodeling is used to accomplish specifications for the abstract syntax. We intro-
duce the Domain Model language by analyzing its syntax definition (Fig.3shows in
EBNF notation). The language with its basic notions and their relations are defined
with structural constraints (for instance to express containment relations, or type cor-
rectness for associations), multiplicities, precise mathematical definition and implicit
relationships (such as inheritance, refinement). The visual appearance of the domain
specific language is accomplished by syntax specifications, which is done by assign-
ing visual symbols to those language elements that are to be represented on diagrams.
5.2 Syntax
For describing the language in general, the rule language checks various kinds of
activities. The primary requirement is to specify the concept of the syntax (i.e.
abstract and or concrete syntax) and develop its grammar. The semantics is designed
to define the meaning of the language. The activities are completed by concepting,
designing and developing a systematic domain-specific rule language systems, defin-
ing the functions and its parameters, priorities or precedence of operators and its
values, naming internal and external convention system. The syntaxes are expressed
with certain rules, conforming to BNF or EBNF grammars that can be processed
by rules or process engine to transform or generate the set of rules as an output.
The generated rules follow the abstract syntax and gram-mar to describe the domain
46 N. Mani et al.
concepts and domain models because both the artefacts (abstract syntax and gram-
mar) are reflected in the concrete syntax.
5.3 Abstract Syntax
The abstract syntax refers to a data structure that contains only the core values set in
a rule language, with semantically relevant data contained therein. It excludes all the
notation details like keywords, symbols, sizes, white space or positions, comments
and color attributes of graphical notations. The abstract syntax may be considered as
more structurally defined by the grammar and Meta model, representing the structure
of the domain. The BNF may be regarded as the standard form for expressing the
grammar of rule language, and some type describes how to recognize the physical set
of rules. Analysis and downstream processing of rule language are the main usages
of Abstract syntax. Users interact with a stream of characters, and a parser compiles
the abstract syntax by using a grammar and mapping rules.
For example, we do process activities in our case domain (Global Digital Con-
tent): Extraction and Machine Translation (MT). The list of the process model, event
and condition are following:
List of Process
List of Event
<EventList> ::={
gic:Text -->SourceTextSegmentation,
List of Conditions
Domain Model Definition for Domain-Specific Rule Generation . .. 47
IF(<SingleLangugeDetection((gic:Text)::== True
IF(<MultiLanguageText(gic:Text)::== True|False>)
<gic:MachineTranslation.Condition>::= IF (<gic:Translation
(Source.Lang, TargetLang,gic:Text) ::= True|False>)|
IF (<gic:Translation.Memory ::= <TM)(Mem Underflow)|
IF (<gic:Translation.Memory ::= >TM)(Mem Overflow)|
where L is length of text and TM is the specific memory size.
5.4 Concrete Syntax
Rule languages use textual concrete syntax, which implies that a stream of charac-
ters expresses the program syntax. The modelling languages traditionally have used
graphical notations and primarily in modelling languages. Though textual domain-
specific languages (and mostly failed graphic based general-purpose languages) have
been in use for a long time only recently, the textual syntax has found a prominent use
for domain-specific modelling. Textual, concrete syntax form have been traditionally
used to store programs, and this character stream is transformed using scanners and
parsers into an abstract syntax tree for further processing by the Programming lan-
guages. In the modelling languages, editors have found a major usage, as it directly
manipulates the abstract syntax and uses projection to render the concrete syntax in
the form of diagrams.
The concrete syntax of DSLs is expected to be textual by default. If good tool
support is available, the textual support has been found to be adequate for compre-
hensive and complex software systems. The programmers write lesser code in DSL
as compared to a GPL for expressing the same functionality—because the available
abstractions are quite similar to the domain. An additional language module suitable
for the domain is defined easily by the programmers.
6 Rule Language Definition
Now we go back to the full rule definition. The DSRL grammar [2] is defined as
follows. We start with a generic skeleton and then map the globic domain model
<DSRL Rules> ::= <EventsList>
48 N. Mani et al.
<EventLists> ::= <Event> | <Event> <EventLists>
<Event> ::= EVENT <EventName> IF <Expression> |
<RulesList> ::= <Rule> | <Rule> <RuleList>
<Rule> ::= ON<EventName>
<ActionList> ::= <ActionName> |
<ProcessModelList> ::= <ProcessModel> |
<ProcessModel> ::= ProcessModel <ProcessModelName>
::= <ProcessModelName>
tions of a process model.
The description of DSRL contains lists of events, condition, an action of the rules
and process model states. An event can be an internal or external (for rules gener-
ated as an action, it may be the INTERNAL or EXTERNAL term) or generated
when the expression should have been satisfied by the condition or becomes true.
An event name activates with ON syntax, which is a Boolean expression to deter-
mine the conditions that apply and the list of actions that should be per-formed when
event and condition are matched or true (preceded by DO syntax). The process mod-
els contain the state name. A certain policy is decided in the process model when
sequential, and parallel actions are performed or sent in that state. An action is an
executable program or set of computation decussation. The action provides methods
or function invocation, creating, modifying, updating, communicating or destroying
an object. DISCARD allows discarding the instructions, and DELAY allows delay-
ing the instructions, but one.
For example, the gic:Extraction is used in an event on Text Input by user as a
source data.
EVENT gic:TextInput::BOOL IF TextInput_Get
EVENT gic:TextInput::BOOL IF TextInput_ON
ON presence
IF (gic:SourceLang:EN) DO
ON presence
IF (gic:TextLength <X) DO
Domain Model Definition for Domain-Specific Rule Generation . .. 49
Notification to user (Text LENGTH LESS THAN X)
Notification to user(Source language is invalid)
7 Implementation of Principle Architecture of DSRL
The principle architecture of a domain-specific rule (DSR) is the automated model to
text generator on accessible domain models, extracting information from them, and
translating it into output in a specific target syntax. This process follows the con-
cept of Model-Driven Architecture which depends on the metamodel. The modeling
language with its concepts, the source syntax, semantics and its rules are required
by the domain-specific framework and target environment. We present the process
architecture of a domain model translation and the target rule environment in Fig.4.
Fig. 4 MDA organisation view of models approach and artifacts of DSRL generator
50 N. Mani et al.
7.1 Architecture
The architecture of the DSRL generator follows the MDA as four-level model orga-
nization, presented by Bzivin [32] as illustrated in Fig.4. At the top level, the M3
is the Syntax Definition Formalism (SDF) metametamodel which is the grammar of
the SDF. This level is also known as Computational Independent Model (CIM) or
metametamodel as defined (and thus conforms to) itself [33]. A self-representation
of the BNF notation takes some lines. This notation allows a defining infinity of
well-formed grammars. A given grammar allows description of the infinity in syn-
tactically correct DSR configuration.
At the M2 level, we describe the DSRL metamodel, i.e., the grammar of DSRL
with ECA as defined in SDF and this level is called Platform Independent Model
(PIM). The metamodel conforms to the metametamodel at level M3.
At the M1 level, we describe DSRL models for configuration applications. It is
known as Platform Specific Model (PSM) consisting of entity and definitions. The
model conforms to the metamodel at level M2. The bottom level is called M0, we
define the configuration of BPM customization consisting of DSR and XML rules,
which represent the models at the M1 level.
7.2 Mappings Domain Model and Domain-Specific Rule
A mapping is description of mapping rule definitions, generation, configuration and
execution of order specification. Each mapping rule specifies what target model frag-
ment is created for the given DSR. The mapping rule body contains one or more class
of the domain model occurrences (Sect. 4) with all attribute and operational value set.
Expressions for attribute, functional and operational setting are based on a specific
source metamodel, Fig. 5shows the corresponding domain model of gic:extraction
sub type of digital content process used as a source metamodel to describe the DSRL
conceptualization as illustrated in Fig. 2(Sect. 3). Although the given source meta-
model is completely translated into graphical model to text rule by using the grammar
or language definition of source and target metamodel as given in Sects.4and 5,it
is sufficient to show all basic mapping constructs.
7.3 Domain Model Translation into DSR
A rule generation is made automatic such that domain models are accessed in a
way to extract information and translate it into output in a specific syntax based
on feature model, as selected by the domain user. This process model is guided by
the metamodel, the modeling language with its high level of concepts, syntactical,
Domain Model Definition for Domain-Specific Rule Generation . .. 51
Fig. 5 Source metamodel of gic:Extraction as example DSRL used for mapping
Fig. 6 Domain model to DSR translations
and semantics rules. The input required by the user (selection of the feature model)
needs a domain to process the domain model, target environment as rule generation
and configuration. We present the process of domain model translation and target
rule environment in Fig. 6.
52 N. Mani et al.
In an example of a rule generation from a domain model, we propose an approach:
the domain model would be translated into a domain-specific rule language through
a model transformation or translation, and then the DSRL meta-metamodel would
be synthesized into DSR text by means of a rule generator. It is advantageous to have
syntactical and semantic domain translation achieved by a graphical model to text
model translation. It is a dedicated technology, because rule generators deal with the
abstract and concrete syntaxes of the target language (in Sect.3) directly. The entire
process separates two distinct tasks (translation and synthesis) that are performed
using appropriate tools.
For translation, the models need to be expressed in a modeling language (e.g.,
UML for design models, and programming languages for source models). A meta-
mode expresses the modeling languages syntax and semantics by themselves. For
example, the syntax of the Domain-metamodel has feature notations expressed using
class diagrams, whereas its semantics is described by well-defined rules (expressed
as OCL constraints) and a mixture of natural languages [34]. Based on the language
in which the source and target models of a translations (grammar of model change) or
transformation are expressed, a distinction is made between endogenous and exoge-
nous transformations. The endogenous transformations are expressed between mod-
els in the same languages (when the grammar and structure are same).
The exogenous transformations are conversions made between models and
expressed using different languages (the grammar and syntax are different), which
is also known as translation. Essentially the same distinction was proposed in [35],
but ported to a model transformation setting. We use the exogenous transformation
that taxonomy and graphical model to text rule, whereas the term translation is used
for an exogenous transformation.
8 Generated Rule Analysis and Evaluation
In rule generation evaluation, we validate the type of generated output concerning
the correctness, completeness, output effectiveness and efficiency. Our primary goal
is to have a proof of fully functional and operational correctness, and completeness of
the rule for its feature requirement selected by the domain user. Our rule evaluation
consists of following:
Validation of rule generation concerning under- and over-generation.
Under generation—We define under generation as missing instance (for exam-
ple events, actions etc.) at the time of generation or after generation.
Over generation—This is identified as an added information regarding syntax
and semantics (functional and operational information).
Evaluation of syntactical and semantical correctness of generated rule fulfills our
goal. The above results imply that if one knows, for example, how to formulate
partial correctness of a given deterministic algorithm in predicate mathematics,
Domain Model Definition for Domain-Specific Rule Generation . .. 53
the formulation of many other properties of the algorithm in predicate mathemat-
ics could have been straightforward. As a matter of fact, partial correctness has
already been formulated in predicate mathematics and manual rule templates of
many feature deterministic algorithms.
Syntactical correctness means correct use of keys, functions, values, and gram-
matical rules.
– Semantical correctness is important for functional and operational point of
view; here, we validate the correct order of generated rule and grammatical
Grammatical correctness means the generated rules follow SDF grammar which
is used in M3 level of MDA model as shown in Fig. 4.
Comparing automated and hand-written rules.
Evaluation of completeness of generated rule.
Completeness: Most rule languages are not designed as targets for rule genera-
tion, because of lack of functional and operational parameters in program frag-
ments. Major challenging part in the rule sets has two constraints: dead-lock
situation and live lock situation.
Identification of rule deadlock.
Identification of rule live lock.
It also needs to be identified if rules of the application require any additional set of
rules to function as desired.
9 Discussion
In this paper, we have proposed a syntax definition for domain model language for
rule generation and presented a DSR generation development for domain model
through MDA approach using domain variability. We have presented a novel approach
of handling knowledge transfer from domain concept (domain model) to conceptu-
ally configurable rule language, avoiding the inaccuracies and misunderstanding,
model error, human error or semantic mismatches during translation of graphical
abstract model to text rule. We have added adaptivity to the domain model. We pro-
vide a conceptual view of domain-specific rule generation and manage the domain
model, and variability model using MDA. It helps in managing frequent changes of
the business process along with variability schema of a set of structured variation
mechanisms for the specification. The domain user can generate the DSRs and con-
figure domain constraint in a dynamic environment. They can generate and config-
ure DSRs without knowing any technical and programming skill. The novelty of our
approach is a variability modelling usage as a systematic approach to transforming
(generate) domain-specific rule from domain models.
We plan to extend this approach in combination with our existing work on busi-
ness process model customization based on user requirement (feature model, domain
54 N. Mani et al.
model and process models) so that a complete development life cycle for the cus-
tomization and configuration of the business process model is supported. We explore
further research that focuses on how to define the DSRL concerning abstract and
concrete syntactical description with grammar formation across different domains,
converting conceptual models into generic domain-specific rule language which are
applicable in other domains. So far, it is a model for text translation but has the poten-
tial to serve as a system that learns from existing rules and domain models, driven by
the feature model approach with automatic constraints configuration that is resultant
to an automated DSRL generation.
Acknowledgements This research is supported by Science Foundation Ireland (SFI) as a part
of the ADAPT Centre for Digital Content Technology at Dublin City University (Grant No:
13/RC/2106) and EAI COMPSE 2016, Penang, Malaysia.
1. Mani, N., & Pahl, C. (2015). Controlled variability management for business process model
constraints. ICSEA 2015, The Tenth International Conference on Software Engineering
Advances. IARIA XPS Press.
2. Mani, N., Helfert, M., & Pahl, C. (2016). Business process model customisation using
domain-driven controlled variability management and rule generation. International Journal
on Advances in Software,9(3, 4), 179–190.
3. Tanrıöver, Ö. Ö, & Bilgen, S. (2011). A framework for reviewing domain specific conceptual
models. Computer Standards. Interfaces,33(5), 448–464.
4. Knuth, D. E. (1964). Backus normal form vs. backus naur form. Communications of the ACM,
7(12), 735–736.
5. Deursen, A. V., Klint, P., & Visser, J. (2000). Domain-specific languages: an annotated bibli-
ography. SIGPLAN Not.,35(6), 26–36.
6. Fowler, M. (2010). Domain-specific languages. Pearson Education.
7. Mernik, M., Heering, J., & Sloane, A. M. (2005). When and how to develop domain-specific
languages. ACM Computing Surveys (CSUR),37(4), 316–344.
8. Hudak, P. (1997). Domain-Specific Languages. Handbook of Programming Languages,3,
9. Ringert, J. O., et al. (2015). Code generator composition for model-driven engineering of robot-
ics component connector systems. arXiv:1505.00904.
10. Edwards, G., Brun, Y., & Medvidovic, N. (2012). Automated analysis and code generation for
domain-specific models. 2012 Joint Working IEEE/IFIP Conference on, Software Architecture
(WICSA) and European Conference on Software Architecture (ECSA). IEEE.
11. Prout, A., et al. (2012). Code generation for a family of executable modelling notations. Soft-
ware Systems Modeling,11(2), 251–272.
12. Koch, N., et al. (2008). UML-based web engineering. Web Engineering: Modelling and Imple-
menting Web Applications (pp. 157–191). Springer.
13. Ceri, S., Fraternali, P., & Bongio, A. (2000). Web modeling language (WebML): a modeling
language for designing web sites. Computer Networks,33(1), 137–157.
14. Groenewegen, D. M., et al. (2008). WebDSL: a domain-specific language for dynamic web
applications. Companion to the 23rd ACM SIGPLAN Conference on Object-Oriented Pro-
gramming Systems Languages and Applications.ACM.
15. Ceri, S., Fraternali, P., & Matera, M. (2002). Conceptual modeling of data-intensive Web appli-
cations. IEEE Internet Computing,6(4), 20–30.
Domain Model Definition for Domain-Specific Rule Generation . .. 55
16. Moreno, N., et al. (2008). Addressing new concerns in model-driven web engineering
approaches. International Conference on Web Information Systems Engineering.Springer.
17. Linaje, M., Preciado, J. C., & Sánchez-Figueroa, F. (2007). Engineering rich internet applica-
tion user interfaces over legacy web models. IEEE Internet Computing,11(6), 53–59.
18. Distante, D., et al. (2007). Model-driven development of web applications with UWA, MVC
and JavaServer faces. International Conference on Web Engineering.Springer.
19. White, S. A. (2004). Introduction to BPMN. IBM Cooperation,2.
20. Dumas, M., & Ter Hofstede, A. H. (2001). UML activity diagrams as a workflow specification
language. ł UML 2001—The Unified Modeling Language. Modeling Languages, Concepts, and
Tools (pp. 76–90). Springer.
21. Davis, R. (2001). Business process modelling with ARIS: a practical guide. Springer Science
Business Media.
22. van der Aalst, W. M. P., & ter Hofstede, A. H. M. (2005). YAWL: yet another workflow lan-
guage. Information Systems,30(4), 245–275.
23. IBM. (December 2010). WebSphere©MQ Workow FlowMareket©Definition Language
24. Zeng, L., et al. (2004). Qos-aware middleware for web services composition. IEEE Transac-
tions on, Software Engineering,30(5), 311–327.
25. Boss, J. (January 2008). jBPM Process Definition Language (jPDL).
26. Maker, R., et al. (1992). IDEF3-Process Description Capture Method Report. Information
Integration for Concurrent Engineering (IICE), Armstrong Laboratory, Wright-Patterson AFB:
27. Mili, H., et al. (2010). Business process modeling languages: Sorting through the alphabet
soup. ACM Computing Surveys (CSUR),43(1), 4.
28. Recker, J., et al. (2009). Business process modeling-a comparative analysis. Journal of the
Association for Information Systems,10(4), 1.
29. Diouf, M., Maabout, S., & Musumbu, K. (2007). Merging model driven architecture and
semantic web for business rules generation. International Conference on Web Reasoning and
Rule Systems.Springer.
30. Musumbu, K., Diouf, M., & Maabout, S. (2010). Business rules generation methods by merg-
ing model driven architecture and web semantics. 2010 IEEE International Conference on
Software Engineering and Service Sciences. 2010. IEEE.
31. Pahl, C., Mani, N., & Wang, M. -X. (2013). A domain-specific model for data quality con-
straints in service process adaptations. Advances in Service-Oriented and Cloud Computing
(pp. 303-317). Springer.
32. Bzivin, J. (2005). On the unification power of models. Software Systems Modeling,4(2),
33. Visser, E. (1997). Syntax definition for language prototyping. Eelco Visser.
34. Group, O. M., Unified Modeling Language specification version 1.5. formal. 2003.
35. Visser, E. (2001). A survey of rewriting strategies in program transformation systems. Elec-
tronic Notes in Theoretical Computer Science,57, 109–143.
ResearchGate has not been able to resolve any citations for this publication.
Full-text available
Business process models are abstract descriptions and as such should be applicable in different situations. In order for a single process model to be reused, we need support for configuration and customisation. Often, process objects and activities are domain-specific. We use this observation and allow domain models to drive the customisation. Process variability models, known from product line modelling and manufacturing, can control this customisation by taking into account the domain models. While activities and objects have already been studied, we investigate here the constraints that govern a process execution. In order to integrate these constraints into a process model, we use a rule-based constraints language for a workflow and process model. A modelling framework will be presented as a development approach for customised rules through a feature model. Our use case is content processing, represented by an abstract ontology-based domain model in the framework and implemented by a customisation engine. The key contribution is a conceptual definition of a domain-specific rule variability language.
Full-text available
Business process models are abstract descriptions and as such should be applicable in different situations. In order for a single process model to be reused, we need support for configuration and customisation. Often, process objects and activities are domain-specific. We use this observation and allow domain models to drive the customisation. Process variability models, known from product line modelling and manufacturing, can control this customisation by taking into account the domain models. While activities and objects have already been studied, we investigate here the constraints that govern a process execution. In order to integrate these constraints into a process model, we use a rule-based constraints language for a workflow and process model. A modelling framework will be presented as a development approach for customised rules through a feature model. Our use case is content processing, represented by an abstract ontology-based domain model in the framework and implemented by a customisation engine. The key contribution is a conceptual definition of a domain-specific rule variability language.
Full-text available
Engineering software for robotics applications requires multidomain and application-specific solutions. Model-driven engineering and modeling language integration provide means for developing specialized, yet reusable models of robotics software architectures. Code generators transform these platform independent models into executable code specific to robotic platforms. Generative software engineering for multidomain applications requires not only the integration of modeling languages but also the integration of validation mechanisms and code generators. In this paper we sketch a conceptual model for code generator composition and show an instantiation of this model in the MontiArc- Automaton framework. MontiArcAutomaton allows modeling software architectures as component and connector models with different component behavior modeling languages. Effective means for code generator integration are a necessity for the post hoc integration of applicationspecific languages in model-based robotics software engineering.
Conference Paper
Full-text available
Business process models are abstract descriptions that are applicable in different situations. To allow a single process model to be reused, configuration and customisation features can help. Variability models, known from product line modelling and manufacturing, can control this customisation. While activities and objects have already been subject of similar investigations, we focus on the constraints that govern a process execution. We report here on the development a rule-based constraints language for a workflow and process model. The aim is a conceptual definition of a domain-specific rule variability language, integrated with the principles of a common business workflow or process notation. This modelling framework will be presented as a development approach for customised rules through a feature model. Our use case is content processing, represented by an abstract ontology-based domain model in the framework.
Full-text available
Requirements capture is arguably the most important step of software engineering and yet the most difficult and the least formalized one [Phalp & Shepperd, 2000]. Enterprises build information systems to support their business processes. Software engineering research has typically focused on the development process, starting with user requirements—if that—with business modeling often confused with software system modeling [Isoda, 2001]. Researchers and practitioners in management information systems have long recognized that understanding the business processes that an information system must support is key to eliciting the needs of its users (see e.g. [reference]), but lacked the tools to model such business processes or to relate such models to software requirements. Researchers and practitioners in business administration have long been interested in modeling the processes of organizations for the purposes of understanding, analyzing, and improving such processes [reference], but their models were often too coarse to be of use to software engineers. The advent of eCommerce and workflow management systems, among other things, has led to a convergence of interests and tools, within the broad IT community, for modeling and enabling business processes. In this paper, we present an overview of business process modeling languages. We first propose a categorization of the various languages. Then, we describe representative languages from each family.
Full-text available
Domain-Specific Languages are used in software engineering in order to enhance quality, flexibility, and timely delivery of software systems, by taking advantage of specific properties of a particular application domain. This survey covers terminology, risks and benefits, examples, design methodologies, and implementation techniques of domain-specific languages as used for the construction and maintenance of software systems. Moreover, it covers an annotated selection of 75 key publications in the area of domain-specific languages. 1998 ACM Computing Classification System: D.3 Keywords and Phrases: Example DSLs, DSL design, DSL implementation, survey. Note: Work carried out under project SEN 1.5, Domain-Specific Languages, sponsored by the Telematica Instituut. 1 Introduction In all branches of science and engineering one can distinguish between approaches that are generic and those that are specific. A generic approach provides a general solution for many problems in a certain ar...
Conference Paper
Full-text available
Service processes are often enacted across different boundaries such as organisations, countries or even languages. Specifically, looking at the quality and governance of data or content processed by services in this context is important to control different constraints in this cross-boundary processing. In order to provide a context-aware solution that takes into account data and data processing requirements, a rule-based constraints specification and adapation of processes shall be proposed. A domain ontology shall capture the key data/content data types, activities and constraints, which forms the basis of a rule-based policy monitoring solution. A provenance model is at the core of this ontology solution. The key contribution is a domain-specific model and specification template for constraint policy definition, which can be applied to adapt service processes to domain-specific needs.
1994 marks the year that the international version of ARlS Toolset - a software system capable of analysing, modelling and navigating business processes - was developed and released by the software and consulting company IDS Prof. Scheer GmbH in Saarbriicken, Germany (from 1999 known as IDS Scheer AG). By the end of 2000 an estimated 24,000 copies of this product will have been sold and installed worldwide, with many additional copies supplied to universities for teaching and research projects. The system's connectivity to existing integrated standard software products, such as System Rl3 and by SAP AG, is another factor that is expected to contribute to its fast disseminahon worldwide. The development of ARlS Toolset evolved from results which were obtained at the research institute "Institut ftir Wirtschaftsinformatik" (German for "Institute for Business Process Engineering"). The Institute has a staff of about 50 people and is based at the University of the Saarland. The history of ARlS Toolset is interesting because it demonstrates the benefits provided by a closely-knit mesh of research and product development. Creating such a mesh was only possible by founding IDS Scheer AG in parallel with acting as head of the research institute. Today, this software and consulting firm has a staff of more than 1600 people.
In this chapter I discuss the need for structure in modelling and describe horizontal and vertical structuring. I describe simple ways of linking models together in a horizontal structure and more advanced techniques for vertical structuring. I describe some additional models that support the creation of model structure.