BORM (Business Object Relationship Modeling) is a development methodology used to store knowledge of process-based business systems. BORM is based on the combination of object-oriented approach and processbased modelling. Especially, in the case of agricultural, food and environmental enterprise informational systems we need to work with new flexible modelling tools, because processes and data are instantly changing and modifying through the whole life cycle of such systems. In this paper, we present BORM use as a tool for capturing process information required in the initial phases of information system development.
The attitude of business towards Information Technology (IT) is constantly changing, and increasingly sophisticated. New systems and tools are becoming available. Additionally, there is a constant exchange of ideas between the IT and the business communities, arising out of the development of knowledge-based systems. Today, when modern visual programming tools, combined with the support of rapid webbased application development environments and sophisticated end-user hardware technologies, are available, it would appear that the whole software development process is becoming easier. However, this statement can apply only in those cases where the system complexity of the solution and of the users’ requirements is relatively small.
The aim of our projects was to analyse and suggest improvements to business processes, company structure, data flows, organizational structure; as well as providing IT support for them. Especially, in the case of agricultural, food and environmental enterprise informational systems we need to work with new flexible modelling tools, because processes and data are instantly changing and modifying _________________________________ through the whole life cycle of such systems. We soon realized that we needed to carry out analyses of the problems that were supposed to be solved in order to be able to design the system and properly test their solution.
First, it is important to identify, document, and test a system in order to be able to
analyse and design a more elaborated system. There are different methods and tools.
On one hand, there are methods and tools oriented towards business modelling such
as EPC (EPC, 2008) or varieties of Flow Charts. However, these tend to have
paradigmatic gap in weak relationships with the software engineering. On the other
hand, there are methods and tools of software engineering based on UML, which
assume that the business requirements have already been correctly formulated and
verified
The second is the construction of an initial object model, often called an essential object or conceptual model, built from of a set of domain specific objects known as essential objects. Both these tasks should be carried out with the active participation of the stakeholders, in order to ensure that the correct system is being developed. Consequently, any tools or diagrams used at these early stages should be meaningful to the stakeholders; many of them are not 'computer system literate'.
The most common technique for specification of requirements in current
objectoriented methodologies is Use Case modelling, and subsequent use of Sequence,
Collaboration and State-Chart Diagrams. This is the foundation of most
ObjectOriented development methods. However, this approach is often insufficient by itself
to fully support the depths required for initial system specification. Business analyst
highlights some deficiencies in this approach. There are many views on the
effectiveness of Use Cases and related tools as a first stage in System Design.
Finally, we conclude this discussion by a brief comparison of the BORM method with selected well-known other business (or software engineering) modelling methods presented in Table 1. It is clear that all compared methods have some deficiencies. More precisely, they do not cover all required features. Naturally, the primary objective of the developing of the BORM method is covering as many required features as possible.
RUP objectoriented (UML) poor, use-case modelling only
BORM combination of FSM and objectoriented yes theory behind business modelling support requirement support
ARIS Petri nets yes software engineering support verification and validation has textual phase before diagraming consistency of concepts during modelling modelling phases no no no process simulations no no no there are no phases there are no phases process simulations using prototyping yes no using the MDA 7, internally detailed yes using process simulations, instance-level modelling there is a use of process scenarios constrains rules in the BORM metamodel 5, internally detailed 2.
The BORM (Business Object Relation Modelling) method is in continuous development since 1993 when it originally was intended as an instrument to provide support for building object-oriented software systems based on pure object-oriented languages such as Smalltalk and object-oriented databases. It has now evolved into a system development methodology that has been used successfully in about 30 projects. These systems range through all sizes of software development.
The most common technique for requirements specification in current software
development methodologies is Use Case modelling as the start of UML
documentation process. Indeed Use Cases are often the foundation of most
objectoriented development methods
Our experience in system modelling suggests that UML is not suitable for the first
stages of analysis, where business processes need to be recognized. UML diagrams
are too complex for the business community as they often contain too much detail
concerning potential software implementations. This means classes, inheritance,
public/private methods, attributes, link classes, etc. Here we got almost the same
experience as documented in
We believe that the business community needs a simple yet expressive tool for modelling; able to play an equivalent role to that played by the Entity-Relation Diagrams, Data-Flows Diagrams or Flow-Charts over the past decades. One of the strengths of all these approaches was that they contained only a limited set of concepts (about five) and were comprehensible by problem domain experts after a few minutes of study. Unfortunately the UML/BPML approach lost this advantage of simplicity. 3.
The BORM method has been developed on academic grounds since the 1990s. It
unifies the MDA principle, using an object-oriented paradigm and a unified approach
to business and IT system modelling. For more on the BORM method, see
The three primary theoretical BORMs fundamentals are as follows. • • •
MDA approach that separates a specification of business system description
(CIM – Computation Independent Model) from its computer
implementation specification (PIM – Platform Independent Model); and
this computer specification from the final solution on a concrete
technological platform (PSM – Platform Specific Model). MDA is created
and maintained by the Object Management Consortium
Automata theory is a study of abstract finite-state automata and the problems they can solve. An automaton is a mathematical model for a device that reacts to its surroundings, gets an input, and provides an output. An automaton's behaviour is defined by a combination of its internal state and its accepted input. The automata theory is a basis for system behaviour descriptions.
The first part of the method covers the CIM field, i.e. business engineering. It transforms a project assignment into a model described by data hierarchies, process participants, process scenarios, various diagrams and generated reports. The main instrument of verification and validation is the process simulator.
For the following purposes, it is possible to use this part of BORM without any relation to a software engineering phase:
Organizational consultancy projects. These are process analysis, organizational structure analysis, and drafts for processes or organizational structure improvement.
Projects documenting processes and organizational structure. These are, for instance, projects whose aim is knowledge management, creating training materials, knowledge visualization, etc.
Projects for preparing the groundwork for selection procedures for organizational consultancy, or other consultancy services.
Projects for preparing the groundwork for selection procedures for the delivery of information systems, or other software engineering projects
In BORM each concept has some of the following:
A Set of predecessor concepts from which it could be derived by an appropriate technique and a Set of successor concepts, which could be derived from it by an appropriate technique.
A validity Range - The phases where it is appropriate. In each phase of
BORM modelling, only limited set of concepts is recommended.
A Set of techniques and rules, which guide the step-by-step
transformation and the concept revisions between the system
development phases. There are refactoring techniques; data
normalization, design patterns and other programming-related
techniques
BORM respects UML and MDA, but uses an original diagram for business process modelling. It conveys together information from three separate UML diagrams: state, communication and sequence. The BORM group has found that it is clearly understood by business stakeholders. • •
Each subject participating in a process is displayed in its states and transitions.
This diagram expresses all the possible process interactions between process participants. The business process itself consists of a sequence of particular communications and data flows among participating subjects.
More formally, BORM process diagrams are graphical representations of
interconnected Mealy-type finite state machines of particular subjects. Visual
simulation of a business process is based on market-graph Petri net. Almost the same
approach is described in detail by
The Craft.CASE tool has been designed according to state-of-the-art principles of
object-oriented software development, and uses a special Smalltalk programming
language technology that has been evolving since the 1970's. The program consists
of many user functions, has its own internal object-oriented database, several graphic
editors and a programmable interface. The Craft.CASE tool provides all instruments
for CIM (as business models) and PIM (as conceptual models), including their
mutual interconnection and the possibility to undertake thorough testing
The Craft.CASE can be used in process and organizational consultancy and in analytical projects and information system drafts while identifying requests on newly designed systems; also in component modelling and service oriented architecture. It works with a system model and its processes in an original way that is based on: 1. 2.
An ability not only to visualize processes and systems in the form of diagrams, but also to test them by means of cross-references and graphic simulations.
An ability to model not only symbolic terms (for example, a customer, order, payment, etc.) as drawn symbols in a diagram, but also the possibility of working with real objects (several concrete customers, orders, payments, etc. that differ in their real values such as a date, name, price, etc.). Craft.CASE supports both class-level and instancelevel of modelling.
A method having diagrams as the result of a gradual deriving and checking.
More information about Craft.CASE such as its programming facilities, metamodel
etc. is described in
Our example of modelling using Craft.CASE is to model an information system for a company called FD (Food Delivery); dealing with organic food products delivered from suppliers to local customers. A typical example of such activity is shopping and delivering the goods to a home, or home delivery of pastries and milk and similar products. The information system is primary designed for communication between company and its delivery employees. It also provides an information portal for customers to whom the products are to be delivered.
The case study presented subsequently is structured according to three main phases of the BORM method development process, as briefly depicted in Fig. 1.
The first stage of business modelling is the initiation of work on a project. At the beginning of this phase there are interviews, and then an initial structure and relations are modelled. The testing of generated reports is executed as the final part of this stage. This task is supported by free model drawing in Craft.CASE, see Fig. 2. Interview & Requirements. Since kick-off meeting with the client, we have negotiated that the project should focus on ordering a meal through a customer process. The client expressed a preference for communication with customers via web pages. The client also wants work activity descriptions for the logistics manager and van drivers.
Structure. Structure definition is the first formal step of the method. An analyst should be capable of finding key subjects and process fields on the basis of performed interviews. Participants. Based on the scope of the project and the allowed project time, eight participants were found (See Table 2). These are the job positions of the logistics manager and the van driver, as well as future software components of the web interface and database system; and, of course, a customer and some suppliers, since they take part in the processes.
Hierarchies. For the detailed mapping of problems to be solved and future use of information system creating, two hierarchies of concrete testing data were recognized and modelled. They are food delivery product catalogue and a list of suppliers for these products, see Fig 3. Architecture. Five functional areas were developed, see Fig. 4. It was then decided that three functions (marked as external functions) related to care, and that virtual food products warehouse, advertising activity and web page maintenance will not be included in the concrete scope of our project. The project was limited to activities related to ordering and delivering (marked as internal functions).
Relationships. Several bindings were modelled between elements of the product hierarchy and the product supplier hierarchy to finalize system requirements. These data relationships (elements of the Cartesian product, more formally) show a concrete supplier of concrete products that can be used later for instance-level testing.
Processes. First, the ordering process was modelled, see Fig. 5. Based on the project focus, the food delivery company, a new property for communications was created, with „manual" or „automatic" values, with a relation to the graphic interface where, for automatic communication -- expecting a software realization -- a thick arrowed line between oval activities is drawn. Manual communications are visualized in the standard way of thin arrowed line between oval activities. Modelling cards. The modelling cards of selected objects were presented during a meeting with the client. Based on these cards (generated as a result of the simulation) a process model was finally confirmed by clients/stakeholders for subsequent software implementation.
Fig. 4. Business architecture diagram (Use-Case clone) of the whole system requirements Process testing. Detailed modelling cards and a simulator were used for testing in this step. Craft.CASE simulator displays particular simulation steps and allows dialogue with the user, see Fig. 6. Especially in the area of agriculture, food and environment, where our customers are mostly not technically educated persons, it turned out that fair testing of correctness of all discovered business processes is an absolutely must. si mulati on di alog
Currently there is not a 'standard solution' to the problem of gathering and representing business knowledge. The agricultural, food and environmental business systems are specially such cases, when we need ‘smart’ modelling tools, because processes and data are instantly changing and modifying through the whole life cycle of such systems. Our approach, described here, developed out of business experience and enhanced by graphic models with clear connection towards system development seems to be a promising candidate for such a standard. The approach we propose may serve not only as a tool for formal representation of modelled information, but also as we have demonstrated as a useful tool for communicating with developers and experts from the problem domain (managers, employees, etc.). The key advantages of BORM are its graphic models of knowledge representation, which provides easy and effective feedback. There are also clear rules how to progress through the system development process using this knowledge representation.
The number of projects, including projects in agriculture, food and environment
area, executed in past 10 years gives us an important feedback. The clients say our
analysis gives them a complex and context view of issues they did not see before.
Clients appreciate BORM models having collection of business elements and their
relationships being visualized and simulated together. Moreover, several clients use
miscellaneous legacy Process Modelling Systems for historical reasons (e.g.
EPCbased ARIS, for example).
Our next work will concentrate on elaboration of the BORM, its concept transformations from and to other methodologies and research in the area of business process patterns.