Reusability: Component Retrieval based on MDL Format

Transcription

1 Reusability: Component Retrieval based on MDL Format Deepak Kumar Assistant Professor Computer Science & Engineering Department Panipat Institute of Engineering. & Technology, Samalkha ABSTRACT The keyword based approach that is primarily used in descriptive methods for component retrieval operation is not precise. In keyword-based approach, a candidate asset is selected whenever the keywords that form its representation match all or most of the keywords of the query. Keyword search retrieve large number of candidate components, it is the responsibility of the (re) user to manually interpret the component that is most suitable to be adapted for reuse. To locate the best-fit reusable asset is a tedious task for the (re) user. Hence although keyword based approach is widely used for descriptive methods, it yields results that are far too vague. Here we propose a retrieval approach of components based on MDL file format. The UML models that are used for modeling are stored as MDL file format. These MDL file formats are semantically very information rich and contains lot of valuable information about the asset. The information can be structural as well as behavioral. The class diagram MDL file format contains valuable information about the structural description and contents of a class, i.e. class name, attributes, behavior, relationships, generalization etc. If we search the repository on the basis of attributes of MDL file descriptions, the search result would be better and thus giving higher precision, as compared to keyword based search. Moreover if we assign some numeric weights to different contents of a class, and arrange the search results in descending order, we would be able to find out the precision of the components in descending order of percentage match. Hence the role of (re) user to find the best-fit component from the search results would be much easier. Keywords MDL, UML Models, Search Engine. 1. INTRODUCTION Software reuse refers to using software modules that were developed on a previous software project as part of a new software development project. Software reuse is a worthwhile goal since it has shown to reduce software costs, and improve software quality as well as programmers productivity. Software components have played an important role in modern software and system development. The main contribution of software components is reuse which helps reduces development cost and time, and increase productivity. While libraries of Unified Modeling Language (UML) diagrams and source codes do exist, one of the challenges that still remain is to locate suitable designs and source codes, and adapt them to meet the specific requirements of the software designer. Traditional approaches to component retrieval are keyword-based; which resulted in the retrieval of many irrelevant components. A more promising approach is retrieval based on MDL format, where the contents of MDL file of the UML diagrams are matched to retrieve the components. The UML models that are used for modeling are stored as MDL file format. These MDL file formats are semantically very information rich and contains lot of valuable information about the asset. The information can be structural as well as behavioral. The class diagram MDL file format contains valuable information about the structural description and contents of a class, i.e. class name, attributes, behavior, relationships, generalization etc. These attributes can be used for specification matching with the contents of the repository. The Use case diagram MDL file format contains valuable information about the requirements specification of software. These include use cases and actors. If we search the repository on the basis of attributes of MDL file descriptions, the search result would be better and thus giving higher precision, as compared to keyword based search. Moreover if we assign some numeric weights to different contents of a class, and arrange the search results in descending order, we would be able to find out the precision of the components in descending order of percentage match. Hence the role of (re) user to find the best-fit component from the search results would be much easier. In this paper we describe a tool named as Component Retrieval Search Engine, which assists the software designers in the retrieval of the designs as well as source codes. 2. LITERATURE SURVEY 2.1 Existing Retrieval Techniques In [6], an empirical study of methods for representing reusable software components is described. Thirty-five subjects are searched for reusable components in a database of UNIX tools using four different representation methods: attribute-value, enumerated, faceted and keyword. Precision, Recall and Overlap are used for measuring searching effectiveness.this paper also compares search time for the four methods. Conclusions of this study say that there was no significant difference in search effectiveness, as measured by recall and precision, between the four methods. In [12], an automated support system for developers who search large repositories for the appropriate reusable software objects is discussed. Search is characterized for repository objects in terms of a multistage model involving screening, identification, and the subsequent choice between new object construction or reusable object implementation. Automated support tools, including ORCA, a software Object Reuse Classification Analyzer and AMHYRST, an Automated Hypertext-based Reuse Search Tool, that are based on this model are proposed and discussed. ORCA utilizes a faceted classification approach that can be implemented using 1

2 hypertext. An aspect of AMHYRST s architecture which can automatically create hypertext networks that represents and link objects in terms of a number of distinguishing features is also described. It describes the drawbacks in keyword-based search. In [18], matching to find those components that are similar across the libraries is discussed. Automation to help developers to directly compare and contrast libraries requires matching of similar components (such as classes and functions) across libraries. This paper discusses this kind of matching. It discusses extraction of information about the component i.e. class, functions, relationships, inheritance and composition. Two types of matching i.e. name matching and similarity matching are discussed. 2.2 Gaps in Existing Techniques Through the analysis of the four current major approaches of software classification and retrieval,we think that the behavior based approaches and the denotation semantics approaches are not suitable for classifying and retrieving object-based components, they are more appropriate for functional-based component retrieval ; the formal specification-based methods are difficult to implement ; the advantage techniques of the descriptive classification methods are suitable for object-based component retrieval. The major drawbacks of the traditional descriptive classification schemes for software component retrieval are: There is Ambiguity problem in keyword based search; substantial disagreement over the choice of keywords can occur when different words mean different things to different people. Effort to narrow down the search results is quite high. Precision and Recall is not high. They are based on a controlled vocabulary that must be constructed manually for each application domain. Both classification and retrieval require important human effort because users must select appropriate terms for each facet in the classification scheme from usually list of terms in the controlled vocabulary. This Thesis proposes an approach combining main advantages of descriptive classification methods for component retrieval in order to improve retrieval effectiveness and provide a friendlier user interface through the use of queries in MDL format. 3. EXPERIMENT The keyword based approach that is primarily used in descriptive methods for component retrieval operation is not precise. In keyword-based approach, a candidate asset is selected whenever the keywords that form its representation match all or most of the keywords of the query. Keyword search retrieve large number of candidate components, it is the responsibility of the (re) user to manually interpret the component that is most suitable to be adapted for reuse. To locate the best-fit reusable asset is a tedious task for the (re) user. Hence although keyword based approach is widely used for descriptive methods, it yields results that are far too vague. The primary goal of this thesis is to identify a retrieval approach by using the MDL format. The UML models that are used for modeling are stored as MDL file format. These MDL file formats are semantically very information rich and contains lot of valuable information about the asset. The information can be structural as well as behavioral. The class diagram MDL file format contains valuable information about the structural description and contents of a class, i.e. class name, attributes, behavior, relationships, generalization etc. The use case diagram MDL file format contains valuable information about the requirements specification of the software, i.e. it contains use cases, actors (which are special kind of class), and their relationships. If we search the repository on the basis of attributes of MDL file descriptions, the search result would be better and thus giving higher precision, as compared to keyword based search. Moreover is we assign some numeric weights to different contents of a class, and arrange the search results in descending order, we would be able to find out the precision of the components in descending order. Hence the role of (re) user to find the best-fit component from the search result would be much easier. To successfully combine two paradigms software reuse and UML, for Component Retrieval, an automated tool must be designed, named as Component Retrieval Search Engine. The Purpose of Component Retrieval Search Engine is to retrieve best-fit or most reusable Component from the Repository as intended by the (re) user. Moreover the search results to be displayed in descending order of percentage match with the input query. Hence the role of (re) user to find the best-fit component from the search results would be much easier. To model UML diagrams some software for modeling is required. In this work the software used is the Rational Rose, which is the product of IBM Corporation. In Rational Rose all the diagrams of the UML can be modeled and stored in a single file of MDL file format. Hence various sample cases are taken and are modeled in Rational Rose. The repository should be indexed and classified according to the projects. Here our repository includes two assets, namely source codes and designs. Designs are the diagrams, which are modeled in Rational Rose and stored in the Repository. Source codes consist of programs in java. Designs and Source Codes about a project are stored in the Repository. The search engine should search upon the queries built to match the relevant components. The search input consists of MDL file. The output can be designs or source codes, as required by the user. The resultant should not only be most relevant matched items, but also some relevant matched items for browsing. 3.1 Experimental Setup Component Retrieval Search Engine is developed in VB.Net and MS Access. The repository can be changed or integrated with any other standard databases available like SQL Server or Oracle. For the reporting purposes the data reports of VB.Net are used. The input provided to search engine is modeled in Rational Rose to produce MDL file. To provide the setup of experiment the repository is populated with 1 project, 10 MDL files, and 120 source codes. 3.2 Proposed System The system developed is called as Component Retrieval Search Engine, which stores MDL files as well as source codes in a Repository. The tool has the search engine for retrieving MDL files as well as source codes from the Repository. The System Working is shown in Figure 1. 2

3 Input Query MDL File Repository Repository Manager Re (user) MDL File Reader Search Engine Developer Retrieval of designs and source code Figure 1. Reusability: component retrieval based on MDL format The main components of Component Retrieval Search Engine are MDL file Reader, The Repository, and Search Engine. First the Repository Manager stores the MDL Files as well as source codes in the Repository, i.e. the Repository is populated with the deliverables of software development process. Now when a query is fired for retrieval of a particular component, MDL file Reader reads the MDL file provided as input to the system, and then stores the contents of the MDL file to be used in the search operation in a temporary storage. Based on the contents of MDL File the Search Engine is there to Search for the desired results from the Repository and shows the results back to the developer. The search is done and some weights are applied to search results, such that search results are displayed in decreasing order of percentage match. Hence the developer finds the exact or most relevant match from the repository. The search input query is MDL file to search for relevant components according to the design of system MDL File Reader Diagrams are the pictorial representation of the system. Here we apply some principles to reveal the details from the diagrams. The diagrams can be easily modeled in some software like Rational Rose or Visualizer. After developing the designs of software in UML, store the file. By specifying and storing all the diagrams in repository, we can read and extract the details. Taking the example of class diagram, in a class diagram, there are different classes, operators, attributes, relationships and documentation or comments provided for explanation and behavior. There is relationship between the classes like inheritance, association or any other dependency. Then they are related with the Cardinality. These files must be read by applying some algorithm, to extract all the details. But one must keep trace of classes, operations, attributes and their relations. The attributes of one class should not mix with other class. To implement such algorithm, MDL file Reader is used to extract the necessary information. The MDL file can be read by treating it as a simple text file. Start reading the file and whenever a match is found for object class, object attributes and object operation store the string following next in the temporary storage. Likewise all other attributes can be stored by reading the file. The contents are stored in a temporary storage and used for match when the search is to be performed The Repository Design The purpose of Repository is to store designs as well as source codes. Designs are stored in MDL format and source codes consist of programs in Java. The architecture of Repository is prepared for storing both designs as well as source codes. Both Designs in MDL file format and source codes in Java are stored as simple text files Search Engine To extract diagrams or components from repository, to facilitate software reuse a search engine is required to fetch data as per requirements of the (re) user. As there are many diagrams in UML the search for each diagram is different and the technique required is different. To search from the repository, different search techniques are utilized to extract relevant data. To support the search, there are reports generated which gives the description of the output of the search query. The format of report is same but data varies according to the search technique used. MDL File Class Diagram Search Search Technique In this case, search is made on the MDL file Class Diagram. Now the search is not defined on the keywords or comments of a class. In this case the search is performed on the structure of the whole software i.e. class name, attributes name and operations name. The class diagram of new software is modeled in Rational Rose and the file is saved in MDL file format. Now for the search, this MDL file would be the input to the query. The software would read the class structure modeled and depending upon these structures will search for structure in repository and would result back with preexisting MDL files as well as source codes from the repository. Moreover some weights are assigned to different contents of a class as class name, class attributes and class operations. These weights are computed and stored along with the component path in a temporary storage.the search results are displayed in the decreasing order of their weights i.e. percentage match of the component. The component with highest weight is displayed first and then next and so on. Thus the best- fit component is that which is having highest percentage match. Hence (re) user can easily find the best-fit component from the search results. 3

4 MDL File Use Case Diagram Search: Search Technique In this case, search is made on the MDL files Use Case Diagram. Now the search is not defined on the keywords or comments of a Use Case Diagram. In this case the search is performed on the functional requirements of the whole software i.e. use cases and actors. The use case diagram of new software is modeled in Rational Rose and the file is saved in MDL file format. Now for the search, this MDL file would be the input to the query. The software would read the requirements specification modeled and depending upon these requirements specification will search for requirement specifications in repository and would result back with preexisting MDL files from the repository. Moreover some weights are assigned to different contents of a use case diagram as actors (which is a special kind of class) and use case. These weights are computed and stored along with the component path in a temporary storage.the search results are displayed in the decreasing order of their weights i.e. percentage match of the component. The component with highest weight is displayed first and then next and so on. Thus the best- fit component is which is having highest percentage match. Hence (re) user can easily find the best-fit component from the search results. Example Cases: As per the data in the database, required is a search of a data from the repository. The search technique and the search query depend upon the data to be extracted and the data present in the repository. The functionality of the search engine is to fetch the data from the repository to the user in the desired output. Here, we are dealing with designs and source codes. Designs itself cannot be reused as it is, as there is no physical component that can be extracted, although source codes can be reused as it is. The output form would be the information about the components or the details provided by the repository manager at the time of insertion of the component in the repository. The form of output of the search query is the file itself. Next level of data extraction is to be provided by the developer as how to use the component. Also the output can provide the information about the components, that component could be found from the Repository and then reuse principles are applied. Here the main motive at the design and coding level is to know about the similar kinds of components as per the search query, present in the Repository. To support the explanation provided above, here we provide a case study for Library Management System. First we would populate the Repository with the Diagrams and Source Codes for Library Management System. Then we will try to retrieve the diagrams and source codes from the Repository with search queries. 4. EXPERIMENTAL RESULTS The data about the Library Management System is already populated in the repository with other source codes and diagrams; the process required is the extraction of the data from the repository with the search engine. In this process a search query is fired to the search engine and the respective data is searched from the database. 4.1 Input MDL File The search Input in this search engine does not consists of keywords, rather it consists of MDL file from which information is extracted for the search query input, which is to be given to the search engine. To explain the process, let us take a case study of Course Registration System. In this we have modeled the Class diagram and Use case diagram for Course Registration System on Rational Rose, and then they are saved in MDL file format. These MDL files can be saved as simple text files. Then the information, which is to be extracted for the search query input, from the MDL is shown. Class diagram when modeled in Rational Rose produces the MDL file. This MDL file is a simple text file and contains valuable information about the contents of the class diagram as class name, attributes, operations, relationships etc. This MDL File is supplied as an input to the search engine. Extracted Part of MDL file used for retrieval of Class Diagrams and Source Codes: From the input MDL file, part containing information about the Logical View is extracted as shown below. It contains valuable information about the structural specification for Course Registration System. (Object Petal (Object Design "Logical View" Root category (object Class Category "Logical View" (Object Class "Student" Operations (list Operations (Object Operation "getstudentdetails" (Object Operation "registerforcourses" class_attributes (list class_attribute_list (Object ClassAttribute "name" (Object ClassAttribute "rollno" (Object Class Attribute "dept" (Object Class Attribute "year" (Object Class Attribute "address" (Object Class Attribute "dateofbirth" (Object Class "Course" Operations (list Operations (Object Operation "getcoursedetails" Class attributes (list class_attribute_list (Object Class Attribute "name" (Object Class Attribute "coursed" (Object Class Attribute "credits" (Object Class "Professor" Operations (list Operations (Object Operation "getprofessordetails" (Object Operation "selectcoursestoteach" (Object Operation "submitgrades" Class attributes (list class_attribute_list (Object Class Attribute "name" (Object Class Attribute "empid" (Object Class Attribute "dept" (Object Class Attribute "address" (Object Class "Dayscholar" (Object Class "Hosteller" Keywords Extracted by MDL File Reader: From this, the MDL File Reader extracts the keywords, which are input to the search engine. Keywords consist of class name, attributes and operations. Keywords in this example includes Class1: Student Attributes: name, rollno, dept, address, year, dateofbirth Operations: getstudentdetails, registerforcourses Class2: Course Attributes: name, courseid, credits Operations: getcoursedetails Class3: Professor Attributes: name, empid, dept, address 4

5 Operations: getprofessordetails, submitcoursestoteach, submitgrades Class4: Dayscholar Class5: Hosteller Now these keywords will be the input to search query fired to search engine. Use Case Diagram for Course Registration System: This use case diagram when modeled in Rational Rose produces the MDL file. This MDL file is a simple text file and contains valuable information about the contents of the requirements specification captured by software as actors, use cases and relationships. Extracted Part of MDL file used for retrieval of Use Case Diagrams From the input MDL file, part containing information about the Use Case View is extracted as shown below. It contains valuable information about the requirements specification for Course Registration System. (object Petal (object Design "Logical View" root_usecase_package (object Class_Category "Use Case View" (Object Class "Student" (Object Class "Professor" (object Use Case "register for courses" (object UseCase "view report card" (object UseCase "select courses to teach" (object UseCase "Submit Grades" Keywords Extracted by MDL File Reader: From this, the MDL File Reader extracts the keywords, which are input to the search engine. Keywords consist of Class name, Use Case name. Keywords in this example includes Class: Student, Professor Use Case: register for courses, view report card, select courses to teach, Submit Grades Now these keywords will be the input to search query fired to search engine. 4.2 Search Process There are two techniques that are to be applied for the search query. Case 1 MDL File Class Diagram Search: The input provided is the MDL file.we model a similar type of one class diagram of Library Management System in Rational Rose and store in MDL File. Now this MDL is set as input to the Search Engine. Extracted Part of MDL File (object Petal root_category (object Class_Category "Logical View" (object Class "Patron" (object ClassAttribute "libraryid" (object ClassAttribute "age" (object ClassAttribute "sex" (object ClassAttribute "username" (object ClassAttribute "password" (Object ClassAttribute "address" (object Class "Resource" (object ClassAttribute "name" (object ClassAttribute "title" (object ClassAttribute "accession no" (object ClassAttribute "price" (object Class Attribute "status" (object ClassAttribute "shelf no" (object Class "Student" (object ClassAttribute "roll no" (object ClassAttribute "department" (object ClassAttribute "branch" (object ClassAttribute "year" (object Class "Faculty" (object ClassAttribute "empid" (object ClassAttribute "department" (object ClassAttribute "post" (object Class "Book" (object ClassAttribute "accession no" (object ClassAttribute "author" (object ClassAttribute "publisher" (object ClassAttribute "subject" (object ClassAttribute "ISBN no" (object Class "Journal" (object ClassAttribute "title" (object ClassAttribute "publisher" (object ClassAttribute "accession no" (object ClassAttribute "subject" Keywords extracted by MDL File Reader Class: Patron Class Attributes: libraryid, age, sex, username, password, address Class: Resource Class Attributes: name, title, accession no, price, status, shelf no Class: Student Class Attributes: roll no, department, branch, year Class: Faculty Class Attributes: empid, department, post Class: Book Class Attributes: accession no, author, publisher, subject, ISBN no Class: Journal Class Attributes: title, publisher, accession no, subject As per the diagram modeled in MDL file. As we can see that the right components i.e. source codes and MDL files was selected and shown back along with their respective weights i.e. percentage match by the search engine. Case 2 MDL File Use Case Diagram Search: The input provided is the MDL file.we model a similar type of one use case diagram of Library Management System in Rational Rose and store in MDL File. Now this MDL is set as input to the Search Engine. Extracted Part of MDL File (object Petal root_usecase_package (object Class_Category "Use Case View" (object Class "Library Staff" (object Class "Resource" (object Class "Patron" (object UseCase "Check InResource" (object UseCase "Manage Resource" (object UseCase "Check OutResource Keywords extracted by MDL File Reader Class: Library Staff, Resource, Patron Use Case: Check InResource, Manage Resource, Check OutResource As we can see that the right components i.e. MDL files was selected and shown back along with their respective weights i.e. percentage match by the search engine. 5

6 5. RESULT ANALYSIS Precision: Precision is defined as the number of relevant components retrieved divided by the total number of components retrieved. Precision = Number of relevant components retrieved / Total number of components retrieved Recall: Recall is defined as the number of relevant components retrieved divided by the total number of relevant components in the index. Recall = Number of relevant component retrieved / Total number of relevant components in the index Case 1: MDL File Class Diagram Search: Total components in the repository = 120 Total number of components retrieved = 106 Total number of relevant components retrieved = 80 Total number of relevant components in the index = 115 Precision = 80 / 106 = Recall = 80/115 = 0.69 Case 2: MDL File Use Case Diagram Search: Total components in the repository = 3 Total number of components retrieved = 3 Total number of relevant components retrieved = 2 Total number of relevant components in the index = 2 Precision = 2/3 = Recall = 2/2 = 1 Attaining the precision of in Case 1 is considerably good which indicates that match is up to 75%. Recall value of 0.69 indicates that we would have been able to retrieve 69% relevant components, in Case 1. In Case 2, getting the recall value of 1 indicates that no relevant component is missed. 6. CONCLUSIONS Combining software reuse with UML is a new emerging trend in software development process. Combining these technologies helps the software development process by locating pre-existing components at the design time only, due to which the total effort of software development is decreased. To add to the tools for software reuse and UML, Component Retrieval Search Engine is proposed which helps for the diagram extraction and retrieval of the best-fit component from the repository. The diagrams of UML are modeled in Rational Rose and stored in MDL file format. These UML designs along with the source codes are stored in the Repository. After that a search engine is there to extract diagrams, as well as source codes from the repository according to requirements of the user, by search technique. The work presented can even produce more delicate reports depending upon the output required. UML and software reuse working together can be used to retrieve appropriate components. The search of diagrams as well as source codes from the repository can be made to search for compatible software/components available in the repository before designing and coding by providing the MDL file to the search engine. The search can be made on class diagrams to search according to structural description of the software. The search can be made on use case diagrams to search according to requirements specification of the software. The search results are displayed in decreasing order of percentage match. Hence the role of (re) user to find the best-fit component from the search result is much easier. It has been observed that there is considerable improvement in precision and recall. 7. REFERENCE [1] Arun Sharma, Rajesh Kumar and P.S. Grover, A Critical Survey of reusability aspects for componentbased systems, Proceedings of World Academy of Science, Engineering & Technology, Vol. 21, Jan [2] Boehm, B Managing software productivity and reuse, IEEE Computer 16(9), , [3] Burton, B. A., Aragon,R. W., Bailey,S.A., Koehler,K.D., and Mayer,L.A, The Reusable software library, IEEE Software 4, 4, 25-33, 1987 [4] Clifton, C. and W. S. Li, Classifying software components using design Characteristics, In proceedings of the 10th Knowledge-Based Software Engineering Conference, KBSE 95, IEEE Computer Society press, Los Alamitos, CA PP , 1995 [5] Daniel Lucredio, Antonio Francisico do Prado, Eduardo Santana de Almeida, A Survey on Software Components Search and Retrieval, euromicro, pp , 30th EUROMICRO Conference (EUROMICRO 04), 2004 [6] Frakes,W.B and Pole,T, An Empirical study of representation methods for reusable Software components, IEEE Trans. Soft Engg 20, 8, , 1994 [7] Fichman, R.G & Kemerer, C.E, Object Technology and Reuse: Lessons from early Adopters, IEEE Software 14(10), 47-59, 1997 [8] Grady Booch, James Rumbaugh, Ivar Jacobson, The Unified Modeling Language, User Guide (Pearson Education, 2005) [9] Grady Booch, James Rumbaugh, Ivar Jacobson, The Unified Modeling Language, Reference Manual (Pearson Education, 2005) [10] Hafedh Mili, Fatma Mili and Ali Mili, Reusing Software: Issues and research Directions, IEEE Transactions on Software Engineering, Vol. 21, No 6, 1995 [11] Henninger,S An Evolutionary Approach to constructing effective software reuse Repositories, ACM Transactions on software engineering and methodology 6(2), , 1997 [12] Isakowitz,T and R,J Kauffman, Supporting Search for Reusable Software Objects, IEEE Transactions on Software Engineering 22, 6, , 1996 [13] Jiang Guo, Lqui, A Survey of Software Reuse Repostories, ecbs, p-92, 7th IEEE International Conference and Workshop on the Engineering of Computer Based Systems, 2000 [14] Jilani L L, R.Mili, M Frappier, J.Desharnais and A.Mili, Retrieving Software Components that minimize adaptation effort, In Proceedings of the 12th IEEE International Automated Software Engineering Conference, ASE 97, IEEE Computer Society Press, Los Alamitos, CA pp , 1997a 6

7 [15] Jilani,L.L, R Mili and A Mili, Approximate Retrieval: An Academic Exercise or a Practical Concern, In Proceedings of the 8th Annual workshop on software Reuse (WISR-8), 1997b [16] Michel Ezran, Maurizio Morisio, Colin Tully, "A Survey of European Reuse Experiences: Initial Results," euromicro, p , 24 th. EUROMICRO Conference Volume 2 (EUROMICRO 98), 1998 [17] Mili.A, Yacoub.S, Addy.E, Hafedh.M, Toward an Engineering Discipline of Software Reuse, IEEE software 16(5), 22-31, 1999 [18] Michail,A. & Notkin,D., Assessing Software Libraries by Browsing similar classes, functions and relationships, In Proceedings of 21st International Conference on Software Engineering (ICSE 99), ACM Press, Los Angeles, CA, pp , 1999 [19] Mili A, Mili R and Mittermeir R.T, A Survey of Software Reuse Libraries, Annals of Software Engineering Vol. 5, , 1998 [20] Mili R, Mili A and Mittermeir R.T, Storing and Retrieving Software Components: A Refinement Based System, In Proceedings of 16th International Conference on Software Engineering, IEEE, pp , May 1994 [21] Mili and Edward Addy, Reuse Based Software Engineering (A Wiley-Interscience Publication, John Wiley and Sons, Inc.2002) [22] Peter Eisinga and Jos Trienckens, Software Components for the Industry, From testing of applications to evaluation of components. [23] Prieto-Diaz, Implementing Faceted Classification for Software Reuse, Communication of the ACM 34, 5, 88-97, 1991 [24] Prieto-Diaz, R. and Freeman.,P, Classifying Software for Reusability,IEEE Software.4, 1, 6-16, 1987 [25] Rajesh K Bhatia, Navneet Kaur, Information Retrieval from a composite based Repository using Genetics Algorithms IICAI 2005, page [26] Rajesh K Bhatia, Mayank Dave, R.C Joshi, Retrieval of most relevant reusable Component using genetic algorithms, Software Engineering Research and Practice 2006, [27] Rajesh K Bhatia, Mayank Dave, R.C Joshi, A Hybrid Technique for Searching a Reusable Component from Software Libraries, DESIDOC Bulletin of Information Technology, Vol.27, No.5, September 2007, pp [28] Rajesh K Bhatia, Mayank Dave, R.C Joshi, Ant Colony Based Rule Generation for Reusable Software Component Retrieval, Proceedings of the 1st Conference on India Software Engineering Conference, pp , Feb 19-22, 2008, Hyderabad, India [29] Rajiv D. Banker, Robert J Kauffman and Dani Zweig, Repository Evaluation of Software reuse, IEEE Transactions on Software Engineering, Vol. 19, No 4, April 1993 [30] Rym Mili, Ali Mili and R.T.Mittermeir, Storing and Retrieving Software Components: A Refinement Based System, IEEE Transactions on Software Engineering, Vol.23, No 7, July 1997 [31] S. Araban, A Two level Matching Mechanism for Object-Oriented Class libraries, Ada-Europe 1998: Uppsala, Sweden, pp , no.1, Jan 1993 [32] The MDL File Format [33] UML Summary, Rational Software Corporation, [34] W. Krueger, Software Reuse, ACM, June