<!DOCTYPE article PUBLIC "-//NLM//DTD JATS (Z39.96) Journal Archiving and Interchange DTD v1.0 20120330//EN" "JATS-archivearticle1.dtd">
<article xmlns:xlink="http://www.w3.org/1999/xlink">
  <front>
    <journal-meta />
    <article-meta>
      <title-group>
        <article-title>Normalization of Processes - Toward an Integrated View of Business Process System</article-title>
      </title-group>
      <contrib-group>
        <contrib contrib-type="author">
          <string-name>Václav Řepa</string-name>
          <xref ref-type="aff" rid="aff0">0</xref>
        </contrib>
        <aff id="aff0">
          <label>0</label>
          <institution>Department of Information Technologies, Faculty of Informatics and Statistics, University of Economics</institution>
          ,
          <addr-line>Prague</addr-line>
        </aff>
      </contrib-group>
      <abstract>
        <p>In spite of the unquestioned importance of business process modeling in both the information systems development and the enterprise development the methodology standards in this field are still hardly insufficient. One of the most important aspects which has to be covered with the process models is the essential unity of object oriented and process oriented views of a business system. In this paper we introduce so-called Process Normalization technique, a part of the MMABP methodology, as a particular way of methodical covering of this essential unity. This technique is freely inspired with the famous 'ancient' Normalization of Data Structures technique which we regard as relevant right because of the essential unity of objects and processes in the business system. The process of the normalization of processes is explained and illustrated by example. Important connected problems of structural algorithmic thinking as well as the other ways of manifestation of the mentioned essential objectprocesses unity in the business system are also discussed in the paper.</p>
      </abstract>
      <kwd-group>
        <kwd>business process model</kwd>
        <kwd>normalization</kwd>
        <kwd>structured programming</kwd>
        <kwd>object orientation</kwd>
        <kwd>process orientation</kwd>
      </kwd-group>
    </article-meta>
  </front>
  <body>
    <sec id="sec-1">
      <title>-</title>
      <p>
        In spite of the unquestioned importance of business process modeling in both the
information systems development and the enterprise development the methodology
standards in this field are still hardly insufficient. As a good example of this
insufficiency we can mention even the modeling standard – BPMN language [
        <xref ref-type="bibr" rid="ref10">10</xref>
        ]. Besides
the serious internal contradictions BPMN is still able to cover just a small part of
needed contents. One of the most important aspects which has to be covered with the
process models is the essential unity of object oriented and process oriented views on
a business system.
      </p>
      <p>
        The ideas expressed in this paper are based on the Methodology for Modeling and
Analysis of Business Processes – MMABP [
        <xref ref-type="bibr" rid="ref1">1</xref>
        ]. The first basic principle of this
methodology is the Principle of Modeling which expresses the presumption that the
objective basis for the implementation of the business system in the organization must be
constituted by real facts existing outside of, and independently of, the organization. In
other words, every organization as an implementation of some business system
(business idea) must be based on the model of the relevant part of the Real World.
      </p>
      <p>MMABP distinguishes between two basic dimensions of the Real World: structure
(object view) and behavior (process view). In both dimensions there are two basic
types of model: global (system) view on the system as a whole and detailed
(particular) view on just one element of the system. The main general difference between the
global (system) and the detailed (particular) views is in the factor of time. System
(global) view always tries to abstract this factor and focus on stable, time independent
aspects of the modeled system while the detailed model respects it anyway by
focusing mainly on it. Global view then can be characterized as object-oriented while
detailed view as process (algorithmic)-oriented (see Figure 1).</p>
      <p>Sufficient business system modeling methodology thus has to support both types of
models (global vs. detailed) in both dimensions (structural vs. behavioral). Moreover,
both dimensions have to be always taken in their mutual interconnections. At the
same time the methodology has to integrate also both essential components of a
business system: its managerial and technology contents. Proposed Process Normalization
technique tries to cover all those aspects offering a certain way of integration of the
global process map with the process details, intentional processes with conceptual
objects, and also the process structure with their managerial meaning.
2</p>
    </sec>
    <sec id="sec-2">
      <title>Business process system and its managerial context</title>
      <p>
        By the 'business process system' we mean the system of mutually collaborating
business processes. To understand the 'business essence' of the collaboration of
processes in terms of ideas of process-driven management [
        <xref ref-type="bibr" rid="ref5">5</xref>
        ] one primarily has to
differentiate between two basic functional types of processes: key ones versus support ones.
As customer needs are constantly changing, the processes in the organization should
change as well. This means that any process in the organization should be linked to
the customer needs as directly as possible. Thus, the general classification of
processes in the organization distinguishes mainly between:
• Key processes, i.e. those processes in the organization which are linked directly to
the customer, covering the whole business cycle from expression of the customer
need to its satisfaction with the product / service.
• Support processes, which are linked to the customer indirectly - by means of key
processes which they are supporting with particular products / services.
      </p>
      <p>The value of the key process is given by its direct contact with the value for the
customer as it is its the main goal. The values of other (support) processes are given
by the services by which these processes support other processes. This way every
process is ultimately connected to the value for the customer either directly (key
process) or through its services for other processes. Key processes thus represent a
specific enterprise's way of satisfying the customer needs while support processes
represent rather standard functionality often connected with some technology.
Consequently, key processes are very dynamic, often changing, permanently developing, every
instance is an original. Support processes are mostly static, stable, offering
standardized and multiply usable services.</p>
      <p>So the main effort in the process of creating the conception of the system of
processes must be establishing the equilibrium of needed dynamics of key processes on
one hand and the necessary stability of the system ensured with its maximally
standard support processes.
3</p>
    </sec>
    <sec id="sec-3">
      <title>Process normalization</title>
      <p>
        Normalization of processes is a technique freely inspired with the Normalization of
Data Structures technique firstly introduced by E.F.Codd in [
        <xref ref-type="bibr" rid="ref2">2</xref>
        ], then elaborated in
further detail with R.F.Boyce in [
        <xref ref-type="bibr" rid="ref3">3</xref>
        ] and comprehensively explained in [
        <xref ref-type="bibr" rid="ref9">9</xref>
        ]. Although
the original Codd's intention was mainly technical in terms of a database system
design, this technique started uncovering the essential Principle of Modeling in the field
of information systems development which has been later defined by P.Chen in [
        <xref ref-type="bibr" rid="ref6">6</xref>
        ].
As this principle is essential also in terms of its validity in all dimensions of the Real
World models, it has to be valid even for process descriptions. Regarding this fact
together with the aforementioned essential unity of objects and processes in the
business system as it is defined in MMABP [
        <xref ref-type="bibr" rid="ref1">1</xref>
        ] we can define the goals of the
Normalization of Processes as follows:
(a) Reduce redundancy of process activities.
(b) Ensure that all activities and non-initial events are dependent on the initial event.
(c) Eliminate unnecessary hidden dependency relationships within a process.
      </p>
      <p>
        The Processes Normalization goals support mainly the basic ideas of
processdriven management excellently expressed in [
        <xref ref-type="bibr" rid="ref5">5</xref>
        ]. The goal (a): redundancy of process
activities means unnecessary repeating of activities, with essentially the same content
and meaning in different processes. Unrecognized redundant occurrence of activities
with the same contents in different processes is one of typical consequences of
hierarchical organization mentioned by Hammer in [
        <xref ref-type="bibr" rid="ref5">5</xref>
        ] as a 'symptom of broken processes'.
The goal (b): unconditional clear dependency of all activities on the initial event is a
certain way to ensuring the relationship to the customer-oriented value of the
performance of the company. This goal, together with the goal (c): no hidden dependencies,
support mainly the principle of key processes as centerlines of the final meaning of all
activities in the organizations. Using the Process Normalization technique the key
processes are relieved of all activities which signalize the existence of possible
process goals, other than the primary process goal expressed with the dependency on the
initial event. These supportive process sub-goals usually signalize the existence of
other, more general, sub-processes which are focused on some specific goal and are
hidden in the body of the normalized process. From the point of view of the goal of
the normalized process these processes represent the set of supporting actions with
more general meaning which are needed as a step on the way to the final process
target. Their general meaning is then the reason for their removal from the body of the
normalized process and establishing them as standalone support processes. This way
the normalized process is freed of all non-essential activities which are removed to
more general support processes and the necessary relationships between the original
process and the new ones are uncovered in terms of the meaning of support activities
in the context of the goal of the normalized process.
      </p>
      <p>Consequently, the input assumptions for the normalizing procedure are:
• The logical process represents a part of the Real World consisting of natural
process chains and their relationships.
• Each activity in the process represents an activity from some natural process chain
or relationship among process chains.
• Each event in the process represents an activity of some external actor or related
(collaborating) process chain.
• Each natural process chain hidden in the logical process can be uniquely identified
by some event or by a logical structure of events. Such event (structure of events)
is called an 'initial event'.</p>
      <sec id="sec-3-1">
        <title>The procedure of process normalization</title>
        <p>The procedure of process normalization is defined as a sequence of steps by
particular normal forms. The initial condition for each step is that the process is in the
previous normal form (i.e. fulfills its required characteristics). Particular normal forms
are defined as follows:</p>
      </sec>
      <sec id="sec-3-2">
        <title>1st Normal Form (iterative generalizable structures free)</title>
        <p>The process is in the 1st Normal Form if the bodies of all its repeating
nonelementary structural parts (iterations) have been removed to standalone processes
and replaced with process states. Each removed part of the process has been identified
with the corresponding business system object. Its starting event has been defined as a
request from the original process. Its product has been defined as a service and
corresponding event from the point of view of the original (i.e. receiving) process.</p>
      </sec>
      <sec id="sec-3-3">
        <title>2nd Normal Form (alternative generalizable structures free)</title>
        <p>The process is in the 2nd Normal Form if it is in the 1st Normal Form and the
bodies of all its mutually alternative non-elementary structural parts (selections) have
been removed to standalone processes and replaced with process states. Each
removed part of the process has been identified with the corresponding business system
object. Its starting event has been defined as a request from the original process. Its
product has been defined as a service and corresponding event from the point of view
of the original (i.e. receiving) process.</p>
      </sec>
      <sec id="sec-3-4">
        <title>3rd Normal Form (parallel generalizable structures free)</title>
        <p>The process is in the 3rd Normal Form if it is in the 2nd Normal Form and the
bodies of all its mutually parallel non-elementary structural parts (simultaneities) have
been removed to standalone processes and replaced with process states. Each
removed part of the process has been identified with the corresponding business system
object. Its starting event has been defined as a request from the original process. Its
product has been defined as a service and corresponding event from the point of view
of the original (i.e. receiving) process.</p>
      </sec>
      <sec id="sec-3-5">
        <title>4th Normal Form (hidden generalizable sub-structures free)</title>
        <p>The process is in the 4th Normal Form if it is in the 3rd Normal Form and the
bodies of all its non-elementary structural parts (sequences) which are not fully specific to
the starting event of the process have been removed to standalone processes and
replaced with process states. Each removed part of the process has been identified with
the corresponding business system object. Its starting event has been defined as a
request from the original process. Its product has been defined as a service and
corresponding event from the point of view of the original (i.e. receiving) process.</p>
        <p>
          From the point of view of the process system the Process Normalization can be
regarded as a way of uncovering the natural supporting processes which are hidden in
the body of the given process. In the first step all repeatedly used support services are
uncovered in the process. The second step uncovers all conditionally used support
services and the third step uncovers all substantially different (parallel) services
hidden in the same process step. In the last step all remaining parts of the process which
can be also regarded as standalone services are uncovered. Ordering of steps is
important for similar reasons as it is in the normalization of data structures. To be able to
correctly interpret alternative sub-structures (branches) in the process one has to
remove all repeated structures of activities first (1st to 2nd Normal Form) as the decision
about the repetition (end of the loop) might be incorrectly interpreted as a fork
signalizing mutually alternative sub-structures. Seeking for parallel structures does not
make sense between different alternative structures so the structure has to be in the 2nd
normal form before it can be transformed to the 3rd normal form. Similarly, identified
sub-structure cannot overstep the border between alternative nor parallel branches of
the process. These rules generally follow from the theory of structured thinking most
comprehensively described in [
          <xref ref-type="bibr" rid="ref7">7</xref>
          ] and [
          <xref ref-type="bibr" rid="ref8">8</xref>
          ]. Relationships of the process normalization
technique to this theory are also discussed in the paragraph Discussion.
        </p>
      </sec>
      <sec id="sec-3-6">
        <title>Example of Process Normalization</title>
        <p>
          Figure 2 shows unnormalized process Customer Order Management in the BPMN
[
          <xref ref-type="bibr" rid="ref10">10</xref>
          ] notation. The way of process description follows MMABP methodology [
          <xref ref-type="bibr" rid="ref1">1</xref>
          ]
caring especially of the communication with external collaborators. In this, unnormalized
form the process collaborates just with the customer. See the process states (Parallel
AND gates) at the beginning (negotiation of the order) and before the end (payment)
of the process.
        </p>
        <p>
          The notation used in example is a strong reduction of BPMN [
          <xref ref-type="bibr" rid="ref10">10</xref>
          ] standardly used
by MMABP methodology as it follows from the business process meta-model [
          <xref ref-type="bibr" rid="ref1">1</xref>
          ]. To
make BPMN methodically consistent with the MMABP principles it has to be
reduced here just on the basic set of language constructs: activities, two essential types
of events (ad-hoc and timer), decisions and logical connectors, and inputs and outputs.
Events are defined as always external influences of the process. Missing concept of
process state is compensated with the AND join construction expressing the meaning
of the process state as a necessary synchronization of the process flow with the
external influence (set of events). More detailed explanation of the need to improve BPMN
can be found in [
          <xref ref-type="bibr" rid="ref11">11</xref>
          ].
        </p>
        <p>Figure 3 shows the changes in the process caused by transforming it into the first
normal form. Repeating parts which represent handling the objects Order, Product
and Payment have been identified, removed to standalone support processes and
replaced with the new process states each representing the waiting for the external
service.</p>
        <p>The process at Figure 4 is in the second normal form after the removal of two
conditional parts which can be regarded as standalone standard support services. Product
Preparation is performed only if the customer requires nonstandard product and
Order Finalization branch only if the customer adopted the delivered product. As there
are no remaining parallel parts in this process it is also in the third normal form.</p>
        <p>In the fourth normal form (Figure 5) the process has been exempt of two parts
representing standalone standard services: Production Technology and
Expedition&amp;Delivery. All remaining parts of the process are then fully specific to the starting
event Customer request.</p>
        <p>
          Figure 6 shows the resulting map of processes in the Eriksson/Penker notation [
          <xref ref-type="bibr" rid="ref4">4</xref>
          ]
after the full normalization of the original one key process. Besides the newly created
support processes discussed in the previous comments to the particular normal forms
of the process there are also several other new processes created by the normalization
of new support processes: Product Design and Production Planning removed from
the Production Preparation support process, and former three parallel parts of the
Production Step Management process. The process Payment Management originally
created from the key process on the way to its first normal form has been completely
removed in the second normal form together with the process Order Finalization and
is no more connected directly with the key process.
        </p>
        <p>Process map shows the structure of mutually collaborating processes following the
common goal – the final product of the key process Customer Order Management.
The structure perfectly reflects the main characteristics of the process driven
organization:
• All processes are organized in the network structure which means that they are
generally independent, their relationships reflect just their collaboration: mutually
supporting each other with the services. There is no general hierarchy of processes.
• Common meaning of all processes in the map is given by the key process which is
a final consumer of the supporting services thus its final product defines the
meaning of the presence of all processes in the map.
• Such structure of generally independent processes which however follow the
common goal can be called purpose-centric collaboration network structure.
• The process in the center of the structure (key process) represents the direct contact
with the Customer, a representative of the common purpose of the whole structure.
As the main idea of a process-driven management is to make the organization as
much flexible a possible towards the customer needs, the central (key) process is
naturally the most dynamic and specific (purpose-oriented) process in the structure.
• Consequently, other (support) processes the bigger is their distance from the center
of the structure, the more they are universal, generally oriented and naturally static.
For example see at Figure 6 the 'first order' support processes Product Preparation,
Production Step Management and Order Finalization which are obviously closer
to the specific customer aspects than more general 'second order' support processes
like Product Design, Purchase of Part or Enforcement Process.</p>
      </sec>
    </sec>
    <sec id="sec-4">
      <title>Discussion</title>
      <p>
        The essential meaning of the Process Normalization can be characterized as a
summarization of the needed decisions and their mutual dependencies led by the
intention to distinguish as much as possible between the individuality of the normalized
process and the natural generality of the processes which support it. Such principal
differentiation is a certain way to achieve the needed equilibrium of the maximal
flexibility on one hand and maximal stability of the enterprise performance at the other
hand; every needed process which does not directly fulfill the primary function of the
organization should be stabilized as much as possible up to its complete outsourcing.
This way the process outsourcing can be regarded as a border of the maximal possible
stabilization of the process. Often discussed automation of processes then should be
regarded as a special case of the outsourcing where the 'service supplier' is a machine.
The significant outcome of this short reflection is the fact that the complete
automation of the enterprise including also the automation of key processes which is often
regarded as a finite goal of the Business Process Re-engineering is a clear nonsense as
it is in contradiction with the main idea of the process driven management ([
        <xref ref-type="bibr" rid="ref5">5</xref>
        ]). Once
the process is outsourced it cannot be more regarded as a process in terms of the
organization's performance. From the organization's point of view such original process
then exists just as a single service delivered by the external supplier which means that
the organization is no more responsible for its performance as it now has the role of
customer.
      </p>
      <p>
        The crucial condition for the proper use of Process Normalization is the ability to
think about the process a structural way which means to be able to see the process in
terms of the hierarchical structure of sequences, selections and iterations as it is
defined in [
        <xref ref-type="bibr" rid="ref7">7</xref>
        ] and generalized in [
        <xref ref-type="bibr" rid="ref8">8</xref>
        ]. The Structure Diagram introduced by Jackson in
[
        <xref ref-type="bibr" rid="ref7">7</xref>
        ] is the best tool for the description of the process such way. Figure 7 shows the
description of the Customer Order Management process by the Structure Diagram.
Nevertheless, a structural viewing of a business process for the purpose of the Process
Normalization technique requires expressing also possible parallel actions in the
process. For this purpose we enrich the original concept of three basic structure types
(sequence, selection and iteration) from [
        <xref ref-type="bibr" rid="ref7">7</xref>
        ] by the fourth type 'simultaneity' which
represents the parallel running of actions synchronized at the beginning and at the
end. In the Jackson's Structure Diagram we label the elements of this additional
structure with the plus sign ('+') in the upper right corner (see Figure 7). The structural
view of an algorithm prevents parallelism in general as it is a determinant of the
border between a single algorithm and the system of cooperating algorithms (see [
        <xref ref-type="bibr" rid="ref8">8</xref>
        ] for
instance). In spite of it the parallelism can be accepted as a fourth basic type of
structure if it is controlled: a simultaneity can be used only if beginnings as well as ends of
all parallel actions are synchronized in a single point of time which allows regarding
the whole structure as a single element from the point of view of the upper level of the
structure. Such encapsulation of parallel actions in a single structure element prevents
the possible corruption of the nature of an algorithm1. Another complement to the
Jackson's Structure Diagram is the predefined action 'End' which means the
termination of the process before its regular end. Use of the 'End' expresses the situation
when the process has to be terminated for some specific objective purpose in the
particular point of its structure, an exception from the regular process flow. Therefore,
the End operation can be used only as a part of a selection structure under the
condition which represents the specific objective reason for the termination of the process.
Regularly, such situation should be expressed in the Structure Diagram as an
upperlevel selection in order to keep the principle that the structure has just one end. From
the point of view of the time flow such description however expresses the decision
before the action which allows this decision. The Jackson's method generally solves
this problem with the special technique called 'backtracking' [
        <xref ref-type="bibr" rid="ref7">7</xref>
        ] which allows the
needed exceptional jump from the structure in the implemented program together
with keeping the clearly static 'conceptual' description of the program structure.
Nevertheless, for the purpose of the Process Normalization we need to describe the
process in structural way but as a process, respecting primarily the time flow which is the
main reason for this complement to the original Structure Diagram.
      </p>
      <p>
        Process Normalization is also closely connected with the object-oriented view of a
business system which represents the “causal” view of the Real World. Each removed
part of the normalized process should be identified with some business system object.
This object points to the general meaning of the new support process created from the
removed sub-process. General meaning of this object can be subsequently used for the
generalization of this new process in order to achieve its needed stability. For
exam1 For example, Tse in [
        <xref ref-type="bibr" rid="ref12">12</xref>
        ] considers even five basic structure types: 'sequence, selection,
parallelism, iteration and procedure call' and mentions also the rule that 'every selection or
parallelism must have a single input state and a single output state' [
        <xref ref-type="bibr" rid="ref12">12</xref>
        ]. It can be seen that
the Tse's 'procedure call' represents in fact any sub-structure invoked by the 'call' operation,
and that the same is implicitly ensured in Jackson's diagram with its hierarchical nature. The
simultaneity thus remains the only natural structure type not covered by the original
Jackson's diagram.
ple processes like Expedition&amp;Delivery, Payment Management, Production Planning,
etc. from the Process Map at Figure 6 are typical representatives of standard processes
absolutely independent of any specific key process which can use their services.
      </p>
      <p>
        The title of this paper is a paraphrase of the title of the crucial work in the field of
conceptual modeling [
        <xref ref-type="bibr" rid="ref6">6</xref>
        ]. In this article Peter Chen expressed the need for perceiving
the data base of the organization as a model of a relevant part of the Real World in
terms of objects (entities), their mutual associations (relationships) and classifications
(types) which is nowadays known in informatics as a 'Principle of Modeling'. This
principle should not be limited just to the field of data (i.e. Real World objects, Real
World ontology) but regarded as a universal principle valid even in the field of
operations (i.e. business processes, Real World behavior). Just this universality of the
Principle of Modeling is an original reason for such title of this paper and it also uncovers
the essential relationship between these two parts of the Real World: objects and
processes, alias ontology of the Real World and behavior in the Real World. Looking at
this relationship from the philosophy point of view one can see also general
consequences of the causality and the intentionality in the Real World.
5
      </p>
    </sec>
    <sec id="sec-5">
      <title>Conclusions</title>
      <p>
        The purpose and practical meaning of the Process Normalization are similar to the
technique for “normalization of data structures” ([
        <xref ref-type="bibr" rid="ref2">2</xref>
        ], [
        <xref ref-type="bibr" rid="ref9">9</xref>
        ]). Like in the case of the
Data Normalization technique the main value of the Process Normalization does not lie
in the procedure and its particular steps in terms of the strict instruction for making
the decisions in the defined order. Once the user of this technique understands its
essential meaning he/she usually looses the ability to use it as a directive procedure
because he/she anticipates what should be the result of the next steps which influences
his/her way of thinking and disallows him/her to keep the prescribed procedure.
Instead of being the prescription of mechanical making the decisions in defined order it
rather serves as a 'training program' for understanding the essential meaning and
purpose of the proper division of process steps into the natural process chains related to
existing business objects.
      </p>
      <p>In the near future we suppose possible development of succeeding process normal
forms and also the analysis of relationships among the principles expressed in the
Process Normalization and some relevant aspects of the conceptual analysis, for
example the problem of generalization structures in the conceptual model and its
manifestation in the structure of supporting processes. Such reflections may also take the
form of new succeeding normal forms.</p>
      <p>All these visions of possible future development emphasize the essential meaning
of the process normalization It is not only the training procedure for understanding the
essence of proper division of activities to processes but also the framework for the
generation of stimuli for the further research in the field of business process modeling
and its relationships to other fields like the conceptual modeling, for instance.</p>
    </sec>
  </body>
  <back>
    <ref-list>
      <ref id="ref1">
        <mixed-citation>
          1.
          <string-name>
            <given-names>Business</given-names>
            <surname>System</surname>
          </string-name>
          Modeling Specification: http://opensoul.panrepa.org/metamodel.html (
          <issue>2000-2016</issue>
          ).
        </mixed-citation>
      </ref>
      <ref id="ref2">
        <mixed-citation>
          2.
          <string-name>
            <surname>Codd</surname>
            ,
            <given-names>E. F.</given-names>
          </string-name>
          :
          <article-title>A Relational Model of Data for Large Shared Data Banks</article-title>
          .
          <source>In Communications of the ACM</source>
          , vol.
          <volume>13</volume>
          , No.
          <volume>6</volume>
          ,
          <issue>(</issue>
          <year>June 1970</year>
          ).
        </mixed-citation>
      </ref>
      <ref id="ref3">
        <mixed-citation>
          3.
          <string-name>
            <surname>Codd</surname>
            ,
            <given-names>E. F.</given-names>
          </string-name>
          <article-title>"Recent Investigations into Relational Data Base Systems"</article-title>
          .
          <source>IBM Research Report RJ1385 (April</source>
          <volume>23</volume>
          ,
          <year>1974</year>
          ).
        </mixed-citation>
      </ref>
      <ref id="ref4">
        <mixed-citation>
          4.
          <string-name>
            <surname>Eriksson</surname>
            ,
            <given-names>H.E.</given-names>
          </string-name>
          ,
          <string-name>
            <surname>Penker</surname>
            ,
            <given-names>M.</given-names>
          </string-name>
          :
          <article-title>Business Modeling with UML: Business Patterns at Work</article-title>
          , Wiley (
          <year>2000</year>
          ).
        </mixed-citation>
      </ref>
      <ref id="ref5">
        <mixed-citation>
          5.
          <string-name>
            <surname>Hammer</surname>
            ,
            <given-names>M.</given-names>
          </string-name>
          ,
          <string-name>
            <surname>Champy</surname>
          </string-name>
          , J.:
          <article-title>Reengineering the Corporation: A Manifesto for Business Revolution</article-title>
          . London: Nicholas Brealey Publishing (
          <year>1993</year>
          ).
        </mixed-citation>
      </ref>
      <ref id="ref6">
        <mixed-citation>
          6.
          <string-name>
            <surname>Chen</surname>
          </string-name>
          ,
          <string-name>
            <surname>P P-S.:</surname>
          </string-name>
          <article-title>The Entity-Relationship Model-Toward a Unified View of Data</article-title>
          .
          <source>In ACM Transactions on Database Systems</source>
          , Vol.
          <volume>1</volume>
          , No.
          <volume>1</volume>
          , (March
          <year>1976</year>
          ).
        </mixed-citation>
      </ref>
      <ref id="ref7">
        <mixed-citation>
          7.
          <string-name>
            <surname>Jackson</surname>
            ,
            <given-names>M.A.</given-names>
          </string-name>
          :
          <source>Principles of Program Design</source>
          , Academic Press, London,
          <string-name>
            <surname>UK</surname>
          </string-name>
          , (
          <year>1975</year>
          ).
        </mixed-citation>
      </ref>
      <ref id="ref8">
        <mixed-citation>
          8.
          <string-name>
            <surname>Jackson</surname>
            ,
            <given-names>M.A.</given-names>
          </string-name>
          :
          <string-name>
            <surname>System</surname>
            <given-names>Development</given-names>
          </string-name>
          , Prentice Hall International Series in Computer Science, Prentice Hall, London,
          <string-name>
            <surname>UK</surname>
          </string-name>
          , (
          <year>1983</year>
          ).
        </mixed-citation>
      </ref>
      <ref id="ref9">
        <mixed-citation>
          9.
          <string-name>
            <surname>Kent</surname>
            ,
            <given-names>W.:</given-names>
          </string-name>
          <article-title>A Simple Guide to Five Normal Forms in Relational Database Theory</article-title>
          , Sept. (
          <year>1982</year>
          ).
        </mixed-citation>
      </ref>
      <ref id="ref10">
        <mixed-citation>
          10. Object Management Group:
          <article-title>Business Process Model and Notation (BPMN) Specification Version 2</article-title>
          .0.2., http://www.omg.org/spec/BPMN/
        </mixed-citation>
      </ref>
      <ref id="ref11">
        <mixed-citation>
          11.
          <string-name>
            <surname>Řepa</surname>
          </string-name>
          , V.:
          <article-title>'Business Process Modeling Notation from the Methodical Perspective</article-title>
          . In:
          <article-title>Towards a Service-Based Internet</article-title>
          . Berlin: Springer, LNCS, pp.
          <fpage>160</fpage>
          -
          <lpage>171</lpage>
          ,
          <year>2011</year>
          . ISBN 978-3-
          <fpage>642</fpage>
          - 22759-2.
        </mixed-citation>
      </ref>
      <ref id="ref12">
        <mixed-citation>
          12.
          <string-name>
            <surname>Tse</surname>
            ,
            <given-names>T.H.</given-names>
          </string-name>
          :
          <article-title>Integrating the structured analysis and design models: a Category-Theoretic Approach</article-title>
          , Australian
          <source>Computer Journal</source>
          , Vol.
          <volume>19</volume>
          , No.
          <issue>1</issue>
          , pp.
          <fpage>25</fpage>
          −
          <lpage>31</lpage>
          , (
          <year>1987</year>
          ).
        </mixed-citation>
      </ref>
    </ref-list>
  </back>
</article>