653 656


Published on

Published in: Technology, Education
  • Be the first to comment

  • Be the first to like this

No Downloads
Total views
On SlideShare
From Embeds
Number of Embeds
Embeds 0
No embeds

No notes for slide

653 656

  1. 1. ISSN: 2278 – 1323 International Journal of Advanced Research in Computer Engineering & Technology Volume 1, Issue 4, June 2012 A Resolved Retrieval Technique For Software Components Swathy vodithala1, P.Niranjan Reddy2 and M.Preethi3 Abstract___Retrieval is a major area in softwarerepository because reuse is only effective if it is II. Related workeasier to locate than to write from scratch. In this A. Keyword Searchpaper we are describing the classical methods forretrieving the reusable components and also Keyword search requires assigning to eachproposed a retrieval mechanism which serves the object a number of relevant keywords or indicesbest. The advantages of signature and behaviouralalgorithms have been unified in order to retrieve [1].the best components. Ex: The following keywords are associated with. objects [5].Keywords: Repository, Retrieval technique,Signature matching, Behavioral matching OBJECT1 -KW1, KW2, KW3. OBJECT2 –KW3, KW4, KW5. I. Introduction If we search the objects based on KW3, then the The software engineering has been widely used result would be OBJECT1 & OBJECT2.Thisarea nowadays, since it accepts even the othertechnical areas into it which strengthens this involves a lot of irrelevant objects. The otherstream. There are three major areas in software disadvantages with the keyword method is theengineering which has to be focused when high cost associated with manual indexing,considering the components for software reuse. which requires skilled personnel and ambiguousThese are described as a) classifying the nature of keywords that can lead to substantialcomponents needed .b) describing the disagreement over the choice of keywordscomponents wanted .c) finding the appropriatecomponents. Finding the component includes amajor area of search techniques and retrieval B. Full-text Retrievaltechniques. In this paper we have surveyeddifferent types of retrieval mechanisms.However, finding and reusing appropriate The high cost of manual indexing has made itsoftware components is often very challenging, attractive to automate the indexing process. Theparticularly when faced with a large collection of simplest kind of automatic indexing is illustratedcomponents. A classified collection is not useful by fill-text retrieval systems. Such systems workif it does not provide the search-and-retrieval on the basis of a simple mechanism:mechanism to use it. At different times, different "Store the fill text of all documents in the collection in a computer so that every charactersoftware reuse classification and retrieval of every word in every sentence of every objecttechniques have been proposed and can be located by the machine. Then, when aimplemented. . Retrieval should allow users to person wants information from that storedformulate high-level queries about component collection, the computer is instructed to searchcapabilities and takes account of the context in for all documents containing certain specifiedwhich a query is performed to assist query words and word combinations, which the userformulation [10]. has specified "[1]. 653 All Rights Reserved © 2012 IJARCET
  2. 2. ISSN: 2278 – 1323 International Journal of Advanced Research in Computer Engineering & Technology Volume 1, Issue 4, June 2012C. Hypertext Search E. Attribute valueHypertext represents one of the newest forms of The attribute value classification scheme uses acomputer-based support for reading documents set of attributes to classify a component [8]. For[1]. Rather than being constrained to the linear example, a book has many attributes such as theorder of conventional documents, users are able author, the publisher, its ISBN number and it’sto move through a hypertext document by classification code in the Dewey Decimalfollowing links represented on the screen by system. These are only example of the possiblebuttons. The basic building blocks in hypertext attributes. Depending upon who wantsare nodes and links. Each node is associated with information about a book, the attributes could bea unit of information, and nodes can be of concerned with the number of pages, the size ofdifferent types. The node type depends on the paper used, the type of print face, thevarious criteria, for example, the class of data publishing date, etc. Clearly, the attributesstored (plain text, graphics, audio or an relating to a book can be:executable program), or the domain object it · Multidimensional. The book can be classifiedrepresents (diary entry, account, financial in different places using different attributesstatement). Links define an invoice to a detailed · Bulky. All possible variations of attributescustomer profile screen. Links are accessed from could run into many tens, which may not bethe source node and can be traversed to access known at the time of classification [9].the destination node. Current hypertext systemsprovide users with sophisticated user interfacetools that enable them to inspect node contents, F. Facetedand to navigate through the network by selectinga path to follow. Faceted classification schemes are attracting the most attention within the software reuse community [8, 9]. Like the attributeD. Enumerated classification classification method, various facets classify components, however, there are usually a lotEnumerated classification uses a set of mutually fewer facets than there are potential attributes (atexclusive classes, which are all within a most, 7).proposed a faceted scheme that uses sixhierarchy of a single dimension [8]. A prime facets.illustration of this is the Dewey Decimal system · The functional facets are: Function, Objects andused to classify books in a library. Each subject Mediumarea, e.g., Biology, Chemistry etc, has its own · The environmental facets are: System type,classifying code. As a sub code of this is a Functional area, setting.specialist subject area within the main subject.These codes can again be sub coded by author. G. Signature matching.This classification method has advantages anddisadvantages pivoted around the concepts of aunique classification for each item. The Signature matching compares names, parametersclassification scheme will allow a user to find and return types of component methods to themore than one item that is classified within the user query.same section/subsection assuming that if more Consider the signatures presented in Figures 1than one exists. For example, there may be more and 2 for a stack of integers and a queue ofthan one book concerning a given subject, each integers, respectively [7].written by a different author. This type ofclassification schemes is one, and will not allow Create: => Stackflexible classification of components into more Push: Stack x Integer =>Stackthan one place [9]. As such, enumerated Pop: Stack => Stackclassification by itself does not provide a good Top: Stack => Integerclassification scheme for reusable software Empty: Stack => Booleancomponents. Figure 1: Signature of a Stack 654 All Rights Reserved © 2012 IJARCET
  3. 3. ISSN: 2278 – 1323 International Journal of Advanced Research in Computer Engineering & Technology Volume 1, Issue 4, June 2012Create: =>Queue program determines a context in which theEnqueue: Queue x Integer:=>Queue behaviour of a component is exhibited. TheDequeue: Queue =>Queue behaviour of a component C is derived from theFront: Queue => Integer set of response sequences by removing thoseEmpty: Queue => Boolean responses which are proper extensions of other responses. Thus, the behaviour of a componentFigure 2: Signature of a Queue C is the set of guaranteed responses to a program p.These signatures are isomorphic up to renaming,and thus exemplify what we have come to refer P: Response  Behaviorto as the vocabulary problem. Software reusersimplicitly associate distinct semantics with In general, Behavioral matching executes eachparticular names, for example, pop and enqueue. component with input data in order to retrieveThus, by the choice of names, a component the components that present the expecteddeveloper can mislead reusers as to the semantics behavior.of components, or provide no means ofdiscriminating between components. Figure 3,for example, appears to be equally applicable as III. Proposed worka signature for both stack and queue, primarilydue to the neutral nature of the names used. The disadvantage of the behavioral algorithm is as follows: when sent a program, the componentCreate: Sequence will respond by executing each call in theInsert: Sequence x Integer =>Sequence program and producing a corresponding outputRemove: Sequence =>Sequence sequence called response. However, here weCurrent: Sequence => Integer have a set of predefined responses. If theEmpty: Sequence => Boolean program’s response is matched with the predefined response which is already defined,Figure 3: Signature of a sequence. then the component will be retrieved. But at times a call is rejected where execution stops andH. Behavioural matching. there are various reasons for this. The method name called may not be the name of an operation in the components interface, or the suppliedBehavioural retrieval works by exploiting the input may not satisfy the operationsexecutability of software components. Programs preconditions [4].are executed using components, and the The disadvantage of the signatureresponses of components are recorded. Retrieval matching is as follows: signature match isis achieved by selecting those components whose entirely based on function types, but not on theresponses (with respect to the program) are behaviour. If we consider the functions strcpy ( )closest to a pre-determined set of desired and strcat ( ), they have same signature butresponses. This idea was originally called opposite in nature.``behavioural matching. A component is (Most of the functions have same signature butrepresented as a relation between programs and different behaviours) [2].responses. This is because in general, a program With the description above, we can sayexecution can yield several responses (due to that alone signature matching and behaviouralnon-determinism) and a response may be evoked matching would not yield good results.by more than one program. Formally, a The proposed algorithm is thecomponent C can be declared as [4]: “combination of both the signature and behavioural matching”.If we consider theC: program  Response. behaviours along with the signatures the disadvantage of signature matching can be minimized. Similarly, if we consider theA program p belongs to program is modeled as a signatures along with the behaviours thesequence of calls on the components interface. A disadvantage of behavioural matching can beresponse is a sequence of values in minimized that is all the names of the methodscorrespondence with a program. In effect, each are properly filtered before having behavioural match. 655 All Rights Reserved © 2012 IJARCET
  4. 4. ISSN: 2278 – 1323 International Journal of Advanced Research in Computer Engineering & Technology Volume 1, Issue 4, June 2012 Stern School of Business Working Paper IS-93-EXAMPLE: COMPONENTS TO FIND 47ADDITION OF TWO NUMBERS [2] “Specification matching of software Components” Amy moormann zaremski andFirstly, we search for a component with Jeannette m. wingsignature [3] “Signature matching: A key to reuse”Amy int name (int ,int) moormann zaremski and Jeannette m. wing.i.e... a function which takes two integer [4] ”Examining behavioral matching”Stevenarguments and returns a integer value. Atkinson.Software Verification Research Centre.Department of Computer Science University of Queensland This may give a list of functions as [5] “Building Software Reuse Library with int sum (int ,int) Efficient Keyword based Search Mechanism” int add ( int ,int) Rajender Nath and Harish Kumar. technia – int sub (int, int) international journal of computing science and int mul (int, int) Communication Technologies, VOL. 2, NO. 1, int div (int ,int) July 2009. (ISSN 0974-3375) int avg (int, int) [6] “Facets of Software ComponentThe resulted list has functions with various Repository” Vaneet kaur and shivani goel. /behaviors. Then we apply the behavioural International Journal on Computer Science andmatching to find the addition of two numbers. Engineering (IJCSE)Here the response of a function is checked [7] “Component Classification and Retrievalagainst the predefined responses and the opposite Using Data Mining Techniques” Proceedings ofbehaviours are removed. The resulted list would National Conference on Challenges &be as Opportunities in Information Technology int sum (int,int) (COIT-2007) RIMT-IET, Mandi Gobindgarh. int add (int,int) March 23, 2007. Achala Sharma, Daman Deep int avg (int,int) Kaur [8] “An Integrated Classification Scheme forThe Expected behaviour is seen in all the above Efficient Retrieval of Components” Dr C.v.gurufunctions .Again here, the sum( ),add( ) comes Rao and P.Niranjan. Journal of Computerunder the “EXACT MATCH” and avg( ) comes Science 4 (10): 821-825, 2008 ISSN 1549-3636under the “RELAXED MATCH”. © 2008 Science Publications [9] “A mock-up tool for software component reusesRepository” P.Niranan andIV. Conclusion and future scope Dr.c.v.GuruRao. [10] ”Supporting reuse by task-relevant and In this paper we explained all the basic retrieval personalized information” Yunwen Ye andtechniques along with their strengths and Gerhard Fischer .ijcse’02.weakness. We also proposed a matchingtechnique which combines both the behaviouraland signature matching .This serves the best First Author SWATHY VODITHALAwhen compared with the individual algorithms. received her B.TechThe future scope involves the unification of basic degree in computer science andtechniques with the other technological engineering in 2005 fromdisciplines which are relevant to optimize the KITS, huzurabad (Jawaharlalsearch as well to retrieve the software Nehru Technologicalcomponents. University). And M.Tech degree in software engineering from KITS, Warangal (kakatiya university) in 2011.she worked with computer scienceReferences department for 2 years in KITS, huzurabad as an[1] “Supporting search for reusable software Assistant Professor. At present, she is working asobjects” Tomas isakowitz and Robert j. an Assistant Professor in computer scienceKauffman.Center for Digital Economy Research department at KITS, Warangal. 656 All Rights Reserved © 2012 IJARCET