J.-F. Girard, R. Koschke, and G. Schied. A metric-based approach to detect abstract data types and state encapsulations. Automated Software Engineering, 6(4):357--386, 1999.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....process. 5. Related work Existing research in clone detection is based on two major approaches: 1) using structural information about the code (e.g. metrics, AST, control data flow, slices, structure of the code expressions, etc. 3, 5, 20, 21, 23, 31] and 2) using string based matches [3, 9, 19, 36]. Each of these methods has its advantages and disadvantages. The methods that fall in the first category are obviously language dependent, thus a bit less flexible, while some of the methods in the second category can only deal with exact matches and can have scalability problems due to the large ....

....and disadvantages. The methods that fall in the first category are obviously language dependent, thus a bit less flexible, while some of the methods in the second category can only deal with exact matches and can have scalability problems due to the large number of comparisons needed. Johnson [19] has developed a method for the identification of exact duplications of sub strings in source code using fingerprints at file level granularity. Baker s tool, called DUP [3] finds exact matches and p matches based on parameters (i.e. replacing identifiers) The granularity is that of chunks of ....

Girard, J.-F., Koschke, R., and Schied, G., "A Metric-Based Approach to Detect Abstract Data Types and State Encapsluation", Journal Automated Software Engineering, vol. 6, no. 4, October 1999.

....order to do that it uses the structures implemented in the newsrc.c file. More than that, the list structure, indicated by the relationships in the second example, is in fact a list of news articles. 6. Related work Related research on similarity measures includes the work of Girard and Koschke [17, 19]. This work is also based on similarity metrics between sot tware components and defines similarity metrics that combine structural and semantic information. The structural information is defined using a resource flow graph representation of the source code and semantic information uses the work ....

Girard, J. F., Koschke, R., and Schied, G., "A Metric-Based Approach to Detect Abstract Data Types and State Encapsluation", Journal Automated Software Engineering, vol. 6, no. 4, October 1999.

....This precondition ensures that functions that operate on elements and their containers (insert element into hash tables) will not be coupled with the element types. This method can not group standalone variables into classes it works only on records (C structures) The clustering approach [8] identifies groups of entities (functions, user defined types and global variables) that should belong to the same class based on the level of similarity of the relationships they are involved in. The authors present an iterative algorithm that combines entities with similar relations into one ....

Girard J.F., Koshke R.,"A metric-based approach to detect abstract data type and abstract state encapsulation", Automated Software Engineering Conference, Nevada, 1997, pp. 82-89.

....function Connected Entities for each one that returns the base entities to which a given subprogram should be grouped as described by Table 1. Metric based approaches cluster entities based on a metric using an iterative clustering approach. Schwanke s [20] and our Similarity Clustering approach [8] and Typebased Cohesion [17] fall in this category. Schwanke s Similarity Clustering is aimed at finding related subprograms based on direct call relationships among the subprograms and common and distinct usages of non local names. Our Similarity Clustering approach distinguishes among different ....

....more iterations ahead. Considering this, the clustering algorithm can easily be modified to work incrementally. Only a preand a post processing phase is necessary. In the pre processing phase, the similarity relation is computed among all components and all free entities using a group similarity [8]. If there are nodes that are mutually exclusive, their similarity is set to 0. The clustering algorithm then clusters all components and free entities based on the similarity relation just computed. The results are clusters that may contain components and base entities. These clusters are then ....

J.F Girard and R. Koschke, "A metric-based approach to detect abstract data types and abstract state encapsulation", Conf. on Automated Software Engineering, Lake Tahoe, 1997.

....more ambiguous. 2.4. Clustering Based on Similarity Schwanke proposed a clustering of routines into modules based on a similarity metric [Schw91] His work was aimed at detecting subsystems and therefore, he only considered routines. This approach was extended to detect atomic components in [Gira97b]. It considers routines, types, global variables, and relationships among them. The clustering algorithm used in this approach works as follows: In each iteration, the most similar groups are combined. This group similarity is based on a similarity between simple entities, so it is explained ....

J.F Girard and R. Koschke. A metric-based approach to detect abstract data types and abstract state encapsulation. submitted for publication, 1997.

No context found.

J.-F. Girard, R. Koschke, and G. Schied. A metric-based approach to detect abstract data types and state encapsulations. Automated Software Engineering, 6(4):357--386, 1999.

No context found.

J.-F. Girard, R. Koschke, and G. Schied. A metric-based approach to detect abstract data types and state encapsulations. Automated Software Engineering, 6(4):357--386, 1999.

No context found.

J.-F. Girard, R. Koschke, and G. Schied. A metric-based approach to detect abstract data types and state encapsulations. Automated Software Engineering, 6(4):357--386, 1999.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC