JP2008140264A - Automatic model generation device and program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To support efficient software development reusing an integrated model by identifying variability and commonality of a plurality of object-oriented models and automatically generating the model having a plurality of models integrated therein. <P>SOLUTION: A model element analysis part 6 receives a model A and a model B as input, extracts relation between classes from the two models, refers to a collation pattern included in a conversion rule definition part 8 to the pair of the extracted inter-class relation, and selects a collation pattern to be applied. An integration model generation part 7 converts and integrates the pair of the inter-class relation by use of a conversion pattern corresponding to the selected collation pattern to generate the integration model 3 wherein the model A and the model B are integrated. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to an object-oriented analysis / design model automatic generation apparatus that supports reusable software development, and a program for causing a computer to function as each unit included in the automatic generation apparatus.

  In object-oriented software development, analysis and design for realizing the functions required as software specifications are described as an analysis / design model using UML (Unified Modeling Language), and the program is based on the contents. It is generally done to create. UML is a model notation standardly used in object-oriented development. The analysis / design model is constructed by dividing the entire software into units called classes and realizing the functions of the software with a structure composed of classes and relationships between the classes. The relationship between classes is often expressed in accordance with the notation of class diagrams defined by UML. One of the prior arts for such object-oriented development is IBM's “Rational Unified Process”. In addition, several CASE (Computer Aided Software Engineering) tools that support object-oriented development have been commercialized, and one of them is Rational Rose (registered trademark) of IBM (registered trademark).

  Usually, when developing one software, it is necessary to create tens to hundreds of classes, assign appropriate names to each of these classes, and connect the classes in an appropriate relationship. One of the relationships between classes defined in UML is “relation”. An association is a relationship that represents a semantic connection between two classes. In the UML class diagram, a class and an association are described as a square meaning a class and a straight line connecting the squares. The connected straight line can also indicate directionality by an arrow.

  The relationship between classes defined in UML includes “generalization” in addition to association. Generalization is a relationship representing the definition of a more general class, focusing on the commonality between different classes. Inheritance is the process of inheriting all the properties of a class and defining a new class with additional properties. By defining a general class having commonality among a plurality of different classes and inheriting the class, the structure of the analysis / design model can be expressed hierarchically. A class inherited in UML is defined as a superclass, and a class inherited is defined as a subclass. In the UML class diagram, generalization is described by drawing a white arrow from the subclass toward the superclass.

  In the class diagram, in addition to the class, the software structure is described using an interface. An interface defines how to use and call a class. An interface is represented by a rectangle like a class, and << interface >> is added to the top of the name. An implementation relationship can be established between an interface and a class defined in UML. The realization relationship is described by a white dotted arrow from the class toward the interface. A class tied to an interface in an implementation relationship means that it can be used and called through the interface.

  The representation of software functions based on the structure composed of the class and interface using the class diagram described above and the relationship between them is useful for facilitating the understanding of the software and improving the efficiency of development work such as analysis and design.

  When developing multiple similar software used in the same field, rather than building individual software models, it is possible to develop a model that integrates individual software models and reuses the integrated models. Efficient. A common model can be constructed by analyzing and extracting the commonality of multiple similar softwares and integrating the extracted commonality.

  As a conventional technology that supports the development of reusable software (see, for example, Patent Document 1 or 2), a framework is generated by managing the variability of requirements and identifying hot spots and frozen spots. There is technology to do. Here, the hot spot is a specification portion that varies for each application in the framework, and the frozen spot is a specification portion that is common among applications in the framework. These conventional techniques can mechanically manage and utilize the commonality and variability identified by the developer.

Japanese Patent Laid-Open No. 11-237982 JP 2002-157117 A

  However, according to the above-described conventional technology that supports the development of reusable software, it is necessary for a developer to manually perform the task of analyzing similar systems in the same field to recognize commonality and variability. there were. Therefore, it is difficult to efficiently construct an integrated model, and there is a problem that the efficiency of software development based on the integrated model is reduced.

  The present invention has been made in view of the above, and identifies commonality and variability of a plurality of models based on object orientation, and automatically generates a model in which a plurality of models are integrated, It is an object of the present invention to obtain an automatic model generation apparatus and program for supporting efficient software development by reusing the integrated model.

  In order to solve the above-described problems and achieve the object, an automatic model generation apparatus according to the present invention is an automatic model generation that generates a model that expresses a software structure including a relationship between classes using an object-oriented notation. In the generation device, a plurality of matching patterns expressing the degree of identity of the relationship between two classes as a combination pattern including matching and mismatching of model elements that are components of the model, and according to each of the matching patterns A conversion rule definition part in which a conversion pattern for defining a relationship between classes after conversion for integrating and converting a relationship between the two classes is defined, and a pair of relationships between classes respectively extracted from two models For each model element, and an analysis for selecting a matching pattern that matches the comparison result from the conversion rule definition unit. By selecting a conversion pattern corresponding to the collation pattern selected by the model element analysis unit and the conversion rule definition unit, and generating a relationship between classes according to the selected conversion pattern, And an integrated model generation unit that generates an integrated model that integrates the two models.

  According to the present invention, when two models that express the structure of software including relationships between classes using object-oriented notation are integrated, pairs of relationships between classes extracted from the two models, respectively. The degree of identity is evaluated by comparing with the matching pattern, and the pair of relationships are integrated and converted to generate the relationship between the converted classes according to the conversion pattern corresponding to each matching pattern. It is possible to easily create a model integrated from the two models. Therefore, in developing similar software in the same field, it is not necessary to newly create a model for each piece of software, and software can be developed based on an integrated model. In addition, efficient software development can be performed by reusing the integrated model.

  Embodiments of a model automatic generation device, a model automatic generation method, and a program according to the present invention will be described below in detail with reference to the accompanying drawings. Note that the present invention is not limited to the embodiments.

Embodiment.
FIG. 1 is a block diagram showing a configuration of an automatic model generation device according to an embodiment of the present invention. As shown in FIG. 1, the automatic model generation device 10 according to the present exemplary embodiment inputs a model A and a model B, which are object-oriented analysis / design models, respectively, and a conversion rule in which a matching pattern and a conversion pattern are defined. Referring to the definition unit 8 and the matching pattern defined in the conversion rule definition unit 8, with respect to a pair of relationships consisting of relationships between classes extracted from the model A and the model B (hereinafter simply referred to as relationships), The model element analysis unit 6 that performs analysis for comparing the two relationships that constitute the pair of the relationship and selects a matching pattern that matches the comparison result from the conversion rule definition unit 8, and the model element analysis unit 6 The corresponding conversion pattern is selected from the conversion rule definition unit 8 in accordance with the collation pattern thus selected, and the relationship pair is integrated according to the selected conversion pattern. By generating a relationship after conversion converted to, an integrated model generating unit 7 integrates the model A and model B to generate an integrated model 3, configured with a.

  Here, the object-oriented analysis / design model is a model including information defined by UML (Unified Modeling Language) including classes and relationships between the classes. Various components constituting the object-oriented analysis / design model are referred to as model elements, and the relationship is defined by a plurality of model elements including information such as classes and relationship types. Hereinafter, the model element included in the relationship is also referred to as a relationship element.

  Furthermore, the integrated model generation unit 7 includes a display device 4 and an input device 5, and when there are a plurality of conversion patterns corresponding to the selected matching pattern, the integrated model generation unit 7 notifies the user of this apparatus via the display device 4. A request is made to select one conversion pattern, and the user notifies the apparatus of which conversion pattern to select via the input device 5.

  FIG. 7 is a class diagram showing an example of the model A input to the present embodiment, and FIG. 8 is a model element that defines the relationship with respect to the relationship included in the model A shown in FIG. It is a figure which shows the list of related elements. FIG. 9 is a class diagram showing an example of model B input to the present embodiment, and FIG. 10 is a diagram showing a list of related elements included in model B shown in FIG.

  In FIG. 7, there are seven classes of “zoo”, “monkey”, “monkey mountain”, “giraffe”, “芋”, “banana”, and “elephant”, and “association” exists between them. Shows when to do. Of the relationships, “relation” is expressed as a straight line that connects squares that represent classes, and the connected straight line can be expressed by adding an arrow. The class at the starting point of the arrow is called a source class, and the class at the destination of the arrow is called a target class. The arrow means that the target class instance can be accessed from the source class instance. For example, “monkey” and “banana” are connected by “related”, the source class is “monkey”, and the target class is “banana”. In addition, a “relation name” that is a word indicating a relationship can be given near the center of a straight line connecting classes. For example, for the relationship between “monkey” and “banana”, the relationship name “eat” Is given. On the other hand, no relation name is given to the relation between “monkey” and “芋”, and in the following, such a case is referred to as “anonymous”. In addition to the relation name, a role name and multiplicity can be given to the relation. For example, for the relationship between “monkey” and “banana”, the roll name “favorite” is described near the end of the “banana” side of the straight line connecting “monkey” and “banana”. Means “favorite” of “monkey”. In addition, near the end of the “banana” side of the straight line connecting “monkey” and “banana”, a multiplicity “20” is further described, indicating the number of instances that may be connected. Yes. The multiplicity is described as “lower limit .. upper limit”. For example, for the relationship between “monkey” and “芋”, the multiplicity of “芋” is “2.4”. Yes. FIG. 9 shows a class diagram for model B having six classes.

  Next, in FIGS. 8 and 10, each relationship is given an ID number, and each relationship is a relationship element, such as relationship type, source class name, target class name, role name, multiplicity, and relationship. It is prescribed from the name. For example, in FIG. 8, the relationship of ID = 5 has the relationship type “relation”, the source class name “monkey”, the target class name “banana”, the roll name “favorite”, and the multiplicity “20”. "The relationship name is" eat ".

  The conversion rule definition unit 8 defines a plurality of verification patterns and conversion patterns corresponding to the verification patterns. The collation pattern is a pattern that is collated with the comparison result of the two relations, and represents the degree of identity of the two relations by a combination pattern including matching and mismatching of the model elements. Hereinafter, the details of the matching pattern will be described based on a specific example. In the present embodiment, model elements for comparing two relationships are, for example, “source class name”, “target class name”, “role name”, and “relation name” that are the relationship elements in FIG. It is. FIG. 3 is a diagram showing an example of a collation pattern in the present embodiment. As shown in FIG. 3, in the present embodiment, one collation pattern is composed of information on collation pattern numbers, source class names, target class names, role names, relation names, and priorities. In the example of FIG. 3, 13 collation patterns are defined. The matching pattern is used to take one arbitrary relationship from each of the two comparison models and evaluate whether the two relationships can be converted as the same or similar on the integrated model. The evaluation method is matched by comparing the source class name, the target class name, the role name, and the relation name for the relational elements of the two relations that are taken out, and evaluating whether the relational element matches, anonymously matches, or does not match. A matching pattern can be found and a matching pattern number can be obtained. Here, “match” means that the relevant relationship elements match in both models, “nameless match” means that the relevant relationship elements match both models anonymously, and “mismatch” means that the relevant relationship Means that the elements do not match in both models. The priority indicates the priority with which the matching pattern is referred to by the model element analysis unit 6, and the smaller the number, the higher the priority. In other words, a reference pattern having a priority of 1 is referred to from a verification pattern having a verification pattern number of 1.

  The conversion pattern gives a definition for converting two relations according to the collation pattern and generating a relation after conversion. 4A and 4B are diagrams illustrating an example of the format of the conversion pattern table according to the present embodiment. As shown in FIGS. 4A and 4B, the format of the conversion pattern table is a conversion pattern that is a number for identifying a matching pattern number in FIG. 3 and a series of conversion processes applied to a model that matches the matching pattern. Numbers, source class names of relationships created by conversion, source class additional elements that are elements (for example, stereotypes, constraints, etc.) added to source classes, target class names of relationships created by conversion, and target classes Target class addition elements that are elements to be added to (for example, stereotypes, constraints, etc.), role names on the target side of relations created by conversion, and types of relations to be created by conversion (eg, relation, generalization, realization) , Dependency, etc.), the relationship name (relation name in the case of a relationship) Additional elements (e.g., stereotypes, generalization, etc.) configured and related additional element is, from a multiplicity of target-side relationship to create the conversion.

  The model element analysis unit 6 arbitrarily extracts a relationship from each of the models A and B, compares the two relationships extracted from each model with each other for each of a plurality of relationship elements, and matches the comparison result. It has a function of performing an analysis for selecting a matching pattern from the conversion rule definition unit 8. As described above, in the example shown in FIGS. 8 and 10, the comparison elements used for comparing the two relationships are the relationship elements “source class name”, “target class name”, “role name”, and “relation name”. Is.

  The integrated model generation unit 7 selects a conversion pattern corresponding to the verification pattern from the conversion rule definition unit 8 according to the verification pattern selected by the model element analysis unit 6. Then, the integrated model generation unit 7 integrates the model A and the model B by integrating and converting the pair of relationships that match the matching pattern according to the selected conversion pattern, thereby generating an integrated model. 3 is generated.

  Next, the operation of the present embodiment, that is, the object-oriented development support method according to the present embodiment will be described with reference to the flowchart of FIG. 2 based on the example shown in FIGS. FIG. 2 is a flowchart showing the flow of processing for inputting two object-oriented analysis / design models and outputting an integrated model in the present embodiment. The two models that are input are model A and model B.

  As shown in FIG. 2, in step S <b> 101, the model element analysis unit 6 refers to the conversion rule definition unit 8 and selects one matching pattern to be applied to the process in order from the highest priority. As shown in FIG. 3, the priority is a priority given in advance to the collation pattern. In subsequent steps, processing is performed for one matching pattern selected in step S101, and processing is sequentially applied from a matching pattern with a higher priority to a matching pattern with a lower priority.

  In step S <b> 102, the model element analysis unit 6 checks whether there is a relationship pair that matches the matching pattern selected in step S <b> 101. Here, the relationship pair means a combination of any one relationship included in the model A and any one relationship included in the model B. As can be seen from FIGS. 8 and 10, the relationship pair is a relationship pair.

  Now, when the collation pattern of collation pattern number 1 which is the collation pattern of the first priority is selected in step S101, the model element analysis unit 6 in step S102, the source class name, target class name, role name and Whether there is a related pair that matches all of the relationship names is searched from the relationship elements of the model A shown in FIG. 8 and the model B shown in FIG. As a result, it can be seen that there is no related pair corresponding to collation pattern number 1. In that case, the process returns to step S101 again, the matching pattern having the next highest priority is selected, and the subsequent processing is continued.

  Next, when the collation pattern of the collation pattern number 2 that is the collation pattern of the second priority is selected in step S101, the model element analysis unit 6 in step S102, the source class name, the target class name, and the role name Are searched for from the relational elements of the model A shown in FIG. 8 and the model B shown in FIG. As a result, as a pair of relations corresponding to the collation pattern number 2, a pair having an ID of 1 in the model A (FIG. 8) and a pair having an ID of 1 in the model B (FIG. 10) are found.

  Therefore, the process proceeds to step S103. In step S103, the processing in the subsequent steps is applied to all pairs that match the matching pattern found in step S102.

  In step S104, the integrated model generation unit 7 checks whether or not there are a plurality of conversion patterns corresponding to the matching pattern selected in step S102. Since the conversion pattern corresponding to the collation pattern selected in step S102 is only one conversion pattern number 2-1, as shown in FIG. 6A, this conversion pattern is selected and the process proceeds to step S106. 6A, 6B, and 6C are examples of the conversion pattern table in the present embodiment, and are given in the order of the collation pattern numbers according to the format of the conversion pattern table shown in FIG. . That is, in FIGS. 6A, 6B, and 6C, a conversion pattern corresponding to the matching pattern shown in FIG. 3 is given for each matching pattern number. One or more conversion patterns correspond to one collation pattern number. FIG. 5 is a diagram for explaining the format of contents that can be entered in each item of the conversion pattern table. As shown in FIG. 5, in the conversion pattern with the conversion pattern number 2-1 in FIG. 6-1, the relationship type after conversion is related, and the source class name is the relationship of interest in model A The source class name (A. source class name), the target class name is the target class name of the relationship focused on model A (A. target class name), and the role name is the role name of the relationship focused on model A ( A. Role name), the relation name is the relation name of the relation being noticed in model A (A. relation name), and the multiplicity is the multiplicity of the relation being noticed in model A (A. multiplicity) And the related multiplicity (B. multiplicity) that is focused on.

  In step S106, the integrated model generation unit 7 applies a conversion pattern to the related pair found in step S102 based on the conversion pattern table shown in FIG. 6 and adds it to the integrated model. Now, the conversion pattern 2-1 corresponding to the matching pattern 2 is applied to the pair of the related element having the ID of 1 in the model A and the relationship having the ID of 1 in the model B. In the conversion pattern 2-1, the source class name, the target class name, and the role name are adopted from the model A side of the pair, and the multiplicity is a multiplicity including the multiplicity in the model A and the model B. It means that. FIG. 11 shows the relationship elements included in the integrated model after the relationship converted in step S106 is output. As shown in FIG. 11, ID = 1 is assigned as a relation element of the integrated model, the relation type is “relation”, the source class name is “zoo”, the target class name is “monkey”, and the role name is “small animal”. It is. Further, the multiplicity is the inclusion of the multiplicity “10..20” of ID = 1 in FIG. 8 and the multiplicity “30..50” of ID = 1 in FIG. 10..20, 30..50 ".

  In step S107, it is determined whether or not processing has been performed for all the relational element pairs found in step S102. If there is an unprocessed relational element pair, the process returns to step S103. Otherwise, the process proceeds to step S108, and the loop 2 between S103 and S106 is terminated.

  In step S109, it is determined whether or not processing has been performed for all the matching patterns. If there is an unprocessed matching pattern, the process returns to step S101. Otherwise, the process proceeds to step S110, and the loop between S101 and S109 is terminated. Now, the processing of the collation pattern 2 that is the second priority is completed, but the processing is not yet performed for the collation patterns after the third priority, and the process returns to step S101.

  Hereinafter, since the flow of the loop processing from step S101 to step S109 is the same as described above, the description of the processing flow will be continued focusing on the difference in pattern.

  The collation pattern of the third priority is the collation pattern of the collation pattern number 3. In model A of FIG. 8 and model B of FIG. 10, there is no related pair that matches this pattern. For this reason, the process of the conversion pattern corresponding to this collation pattern is not implemented.

  The fourth priority collation pattern is a collation pattern of collation pattern number 4. In the model A in FIG. 8 and the model B in FIG. 10, a related element having an ID of 5 in the model A and a related pair having an ID of 4 in the model B are found as related pairs that match this pattern. As shown in FIG. 6A, there are two types of conversion patterns corresponding to the collation pattern number 4: a conversion pattern 4-1 and a conversion pattern 4-2. In this case, it is determined that there are a plurality of conversion patterns in step S104, and a conversion pattern is selected in step S105. Selection of the conversion pattern is performed by displaying on the display device 4 requesting the user to select a conversion pattern, and the user selects one conversion pattern through the input device 5. In step S106, the relationship is converted according to the selected conversion pattern and added to the integrated model. Now, assume that the conversion pattern 4-2 is selected in step S105. FIG. 12 shows the relationship elements included in the integrated model after the relationship converted in step S106 is output. In FIG. 12, the related element with ID 2 is a relationship added corresponding to the collation pattern of the fourth priority. For example, the role name “favorite_main staple” is a combination of “favorite” which is the role name from model A and “main staple” which is the role name from model B, according to the conversion rule shown in FIG. It can be seen that the role name is a variable element indicating variability when the two models are compared. As described above, in the integrated model, a label indicating the identity such as matching or mismatching is added to each function element to express commonality and variability.

  The collation pattern of the fifth priority is the collation pattern of collation pattern number 5. In the model A in FIG. 8 and the model B in FIG. 10, as a pair of relations matching this pattern, a relation having an ID of 3 in the model A and a pair having an ID of 3 in the model B are found. Further, as shown in FIG. 6A, there is only one conversion pattern corresponding to the collation pattern number 5 of the conversion pattern 5-1. FIG. 13 shows the relationship elements included in the integrated model after the relationship converted in step S106 is output. In FIG. 13, the relationship between IDs 3 and 4 is a relationship added corresponding to the collation pattern of the fifth priority. For example, the multiplicity is “0, 1” in the relation of ID 3 and the multiplicity is “0, 3” in the relation of ID 4, which varies by allowing the multiplicity 0 respectively. Label sex.

  The collation pattern of the sixth priority is the collation pattern of collation pattern number 6. In model A of FIG. 8 and model B of FIG. 10, there is no related pair that matches this pattern. For this reason, the process of the conversion pattern corresponding to this collation pattern is not implemented.

  The collation pattern of the seventh priority is the collation pattern of the collation pattern number 7. In model A of FIG. 8 and model B of FIG. 10, there is no related pair that matches this pattern. For this reason, the process of the conversion pattern corresponding to this collation pattern is not implemented.

  The matching pattern of the eighth priority is a matching pattern of matching pattern number 8. In the model A in FIG. 8 and the model B in FIG. 10, a pair having an ID of 2 in the model A and a pair having an ID of 2 in the model B are found as a pair that is compatible with this pattern. Further, as shown in FIG. 6B, there are three conversion patterns corresponding to the collation pattern number 8: a conversion pattern 8-1, a conversion pattern 8-2, and a conversion pattern 8-3. In this case, it is determined that there are a plurality of conversion patterns in step S104, and a conversion pattern is selected in step S105. The selection of the conversion pattern is performed by displaying on the display device 4 requesting the user to select a conversion pattern, and the user selects one conversion pattern through the input device 5. In step S106, the relationship is converted according to the selected conversion pattern and added to the integrated model. Assume that the conversion pattern 8-3 is selected in step S105. FIG. 14 shows the relationship elements included in the integrated model after the relationship converted in step S106 is output. In FIG. 14, the relation of ID 5 and the generalization relation of IDs 6 and 7 are relations added corresponding to the collation pattern of the eighth priority. Since the target class names of the related pairs are different from “Kirin” and “Elephant”, respectively, in the relationship of the integrated model, the related pairs are integrated via the class “abstract_large animal” that abstracts the common part. Yes.

  The collation pattern of the ninth priority is the collation pattern of the collation pattern number 9. In model A of FIG. 8 and model B of FIG. 10, there is no related pair that matches this pattern. For this reason, the process of the conversion pattern corresponding to this collation pattern is not implemented.

  The collation pattern with the tenth priority is the collation pattern with the collation pattern number 10. In the model A in FIG. 8 and the model B in FIG. 10, as a pair of relations matching this pattern, a relation having an ID of 6 in the model A and a pair having an ID of 5 in the model B are found. Further, as shown in FIG. 6-3, there are three conversion patterns corresponding to the collation pattern number 10: a conversion pattern 10-1, a conversion pattern 10-2, and a conversion pattern 10-3. In this case, it is determined that there are a plurality of conversion patterns in step S104, and a conversion pattern is selected in step S105. The selection of the conversion pattern is performed by displaying on the display device 4 requesting the user to select a conversion pattern, and the user selects one conversion pattern through the input device 5. In step S106, the relationship is converted according to the selected conversion pattern and added to the integrated model. Assume that the conversion pattern 10-2 is selected in step S105. FIG. 15 shows the relationship elements included in the integrated model after the relationship converted in step S106 is output. In FIG. 15, the association with ID 8 and the realization relationship with IDs 9 and 10 are relation elements added corresponding to the collation pattern of the tenth priority. In FIG. 15, the commonality of related pairs is extracted as the interface “I_eat”.

  The collation pattern of the eleventh priority is the collation pattern of the collation pattern number 11. In model A of FIG. 8 and model B of FIG. 10, there is no related pair that matches this pattern. For this reason, the process of the conversion pattern corresponding to this collation pattern is not implemented.

  The collation pattern of the twelfth priority is the collation pattern of the collation pattern number 12. In model A of FIG. 8 and model B of FIG. 10, there is no related pair that matches this pattern. For this reason, the process of the conversion pattern corresponding to this collation pattern is not implemented.

  The 13th priority collation pattern is the collation pattern of collation pattern number 13. This is a case where it does not apply to all the collation patterns of the collation pattern numbers 1-13. In the model A shown in FIG. 8 and the model B shown in FIG. Further, as shown in FIG. 6-3, there are two conversion patterns corresponding to the collation pattern number 13: a conversion pattern 13-1 and a conversion pattern 13-2. In this case, it is determined that there are a plurality of conversion patterns in step S104, and a conversion pattern is selected in step S105. The selection of the conversion pattern is performed by displaying on the display device 4 requesting the user to select a conversion pattern, and the user selects one conversion pattern through the input device 5. In step S106, the relationship is converted according to the selected conversion pattern and added to the integrated model. Now, it is assumed that the conversion pattern 13-1 is selected in step S105. FIG. 16 shows the relationship elements included in the integrated model after the relationship converted in step S106 is output. In FIG. 16, the association with ID 11 is a relationship added corresponding to the collation pattern of the 13th priority.

  As described above, when the processing for all the matching patterns is completed, the process proceeds to step S111. In step S111, the integrated model created so far is output as a UML class diagram. FIG. 17 shows a class diagram of the integrated model output in step S111. For example, the class “Elephant” and the interface “I_Eat” are connected by “Relation” with the relationship name “Eat”, the role name of the interface “I_Eat” is “Favorite_Nutrition”, and the multiplicity is “50,100”. This association is a UML representation of the association with ID 8 in FIG. The arrow connecting the class “elephant” and the interface “I_eat” is given “pattern 10” via a dotted line in order to indicate that this has been converted according to the collation pattern of priority 10. Yes. The relationship between other classes is the same, and is illustrated according to the normal UML notation.

  As described above, in the present embodiment, two object-oriented analysis / design models are input, and the commonality and variability of the two input models are analyzed by referring to the matching pattern. An integrated model that integrates the input model is automatically generated. Further, by repeating such operations sequentially, an integrated model that integrates three or more object-oriented analysis / design models can be automatically generated. That is, first, two input models are selected, and an integrated model that integrates these two models is generated. Next, one model other than the previously selected model is selected to generate an integrated model that integrates the selected model and the integrated model. Such an operation can be repeated to generate an integrated model that integrates all input models.

  The model automatic generation apparatus shown in the present embodiment can be configured by an information processing apparatus such as a personal computer having storage means such as a hard disk, memory, and a CPU (Central Processing Unit). For example, the conversion rule definition unit 8 is realized by a storage unit such as a hard disk, and the model element analysis unit 6 and the integrated model generation unit 7 are realized by a CPU, a memory, a hard disk, and the like. In addition, a program for causing a computer to function as the conversion rule definition unit 8, the model element analysis unit 6, and the integrated model generation unit 7 in the present embodiment is stored in the storage unit of the information processing apparatus, and is called and executed on the memory. This makes it possible to execute the above processing. This program can be read by a computer such as a CD-ROM (Compact Disk Read Only Memory), a flexible disk, a DVD (Digital Versatile Disc or Digital Video Disc), or a memory card as an installable or executable file. It can be provided by being stored in a recording medium. The program can also be provided by being stored in a computer connected to a network such as the Internet and downloaded via the network.

  According to the embodiment of the present invention, it is possible to create an object-oriented analysis / design model that is easily integrated from a plurality of individually created object-oriented analysis / design models. In development, it is not necessary to newly create an object-oriented analysis / design model for individual software, and software can be developed based on an integrated object-oriented analysis / design model. Therefore, efficient software development can be performed by reusing the integrated object-oriented analysis / design model.

  Note that the conversion rule definition unit 8 can include an input means for the user of this apparatus to input and set a collation pattern and a conversion pattern. Thereby, a collation pattern and a conversion pattern can be input or their contents can be changed. Since the user can appropriately input and change the collation pattern and conversion pattern defined in the conversion rule definition unit 8, an integrated model can be created according to the purpose of the user.

  As described above, the model automatic generation apparatus and program according to the present invention are suitable for reusable software development work.

It is a block diagram which shows the structure of the model automatic generation apparatus concerning embodiment of this invention. 4 is a flowchart showing a flow of processing for inputting two object-oriented analysis / design models and outputting an integrated model in the embodiment of the present invention. It is a figure which shows an example of the collation pattern in embodiment of this invention. It is a figure which shows an example of the format of the conversion pattern table | surface in embodiment of this invention. FIG. 4 is a diagram for explaining items in FIG. It is a figure explaining the format of the content which can be entered in each item of the conversion pattern table | surface in embodiment of this invention. It is a figure which shows an example of the conversion pattern table | surface in this Embodiment. It is a figure which shows an example of the conversion pattern table following FIG. It is a figure which shows an example of the conversion pattern table | surface following FIG. 6-2. It is a class diagram which shows an example of the model A input into embodiment of this invention. It is a figure which shows the list of the relational elements contained in the example of the model A shown in FIG. It is a class diagram which shows an example of the model B input into embodiment of this invention. It is a figure which shows the list of the relational elements contained in the example of the model B shown in FIG. It is a figure which shows the related element of the integrated model after a 2nd priority collation pattern process. It is a figure which shows the related element of the integrated model after a 4th priority collation pattern process. It is a figure which shows the related element of the integrated model after a 5th priority collation pattern process. It is a figure which shows the related element of the integrated model after an 8th priority collation pattern process. It is a figure which shows the related element of the integrated model after a 10th priority collation pattern process. It is a figure which shows the related element of the integrated model after a 13th priority collation pattern process. It is a class diagram showing the integrated model produced | generated by this Embodiment integrating the model A and the model B. FIG.

Explanation of symbols

DESCRIPTION OF SYMBOLS 3 Integrated model 4 Display apparatus 5 Input device 6 Model element analysis part 7 Integrated model production | generation part 8 Conversion rule definition part 10 Model automatic generation apparatus

Claims (10)

In a model automatic generation device that generates a model that expresses the structure of software including relationships between classes using object-oriented notation,
A plurality of matching patterns that represent the degree of identity of the relationship between two classes as a combination pattern including matching and mismatching of model elements that are components of the model, and the two classes according to each matching pattern A conversion rule definition part in which a conversion pattern that defines a relationship between classes after conversion to integrate and convert the relationship between
A model element analysis unit that compares a pair of relationships between classes extracted from two models for each model element and performs an analysis to select a matching pattern that matches the comparison result from the conversion rule definition unit;
Integration that integrates two models by selecting a conversion pattern corresponding to the collation pattern selected by the model element analysis unit from the conversion rule definition unit and generating a relationship between classes according to the selected conversion pattern An integrated model generation unit for generating a model;
A model automatic generation device comprising:   The model automatic generation apparatus according to claim 1, wherein the relationship between the classes extracted from the two models is a relationship between classes.   The model element includes a source class name that represents the origin of the relationship, a target class name that represents the destination of the relationship, a role name, and a relationship name, and the matching pattern includes the source class name, the target The model according to claim 1 or 2, wherein each of the class name, the role name, and the relation name is a combination pattern including matching, mismatching, and anonymous matching in which names are not defined in both. Automatic generator.   The matching pattern includes priority information, and the model element analysis unit matches the comparison result of the pair of relationships by sequentially referring to the matching pattern with the higher priority to the matching pattern with the lower priority. 4. The model automatic generation apparatus according to claim 3, wherein a collation pattern is selected from the conversion rule definition unit. The model automatic generation apparatus according to claim 1, wherein the conversion rule definition unit includes an input unit configured to input and set the collation pattern and the conversion pattern.   The model element analysis unit extracts all the relationships between classes from the two models, creates first and second model element lists that are lists of model elements, and the first and second models. The model automatic according to any one of claims 1 to 5, wherein an element list is used to find a pair of relationships whose comparison results match the matching pattern defined in the conversion rule definition unit. Generator.   The integrated model generation unit adds a label indicating a degree of identity of the relationship pair to a model element that defines a relationship between the classes after the integrated conversion according to the matching pattern. The model automatic generation device according to any one of claims 1 to 6.   The integrated model generation unit includes an input unit and a display unit, and the integrated model generation unit is a user of the automatic model generation device when a plurality of corresponding conversion patterns exist for one matching pattern. The model automatic generation apparatus according to claim 1, wherein a plurality of conversion patterns are displayed on the display unit, and one conversion pattern is selected through the input unit. .   According to the corresponding matching pattern, the conversion pattern is obtained from the multiplicity in the relationship between one class constituting the pair of relationships and the multiplicity in the relationship between the other classes. The model automatic generation device according to claim 1, wherein the multiplicity in the relationship is defined.   The program for functioning a computer as the said conversion rule definition part, the said model element analysis part, and the said integrated model production | generation part described in any one of Claims 1-9.

Images