Automated relationship traceability between software design artifacts

• US 7,735,068 B2
• Filed: 12/01/2005
• Issued: 06/08/2010
• Est. Priority Date: 12/01/2005
• Status: Active Grant

• Current Assignee: Infosys Technologies Limited (Infosys Limited)
• Original Assignee: Infosys Technologies Limited (Infosys Limited)
• Inventors: Agrawal, Salil, Siddaramappa, Nagaraja Nanjegowda, Thonse, Srinivas, Pooloth, Krishnakumar, Balakrishnan, Rajesh
• Primary Examiner(s): Zhen; Wei Y
• Assistant Examiner(s): SWIFT, CHARLES M
• Application Number:
  US11/292,895
• Publication Number:
  US 20070130561A1
• Time in Patent Office: 1,650 Days
• Field of Search: 717/106, 717/144, 717/127, 717/28, 717/102, 717/208, 717/209, 717/140, 717/101, 717/128, 715230-233, 707/9, 707/3, 707/5, 714/38
• US Class Current: 717/128
• CPC Class Codes: G06F 11/3604 Software analysis for verifying properties of programs byte-code verification G06F9/44589
• CPC Class Codes: G06F 8/10 Requirements analysis; Specification techniques

Abstract

Tools and methods are described herein that allows for measuring and using the relationship between artifacts of a software design, such as requirements, test plans, and so on. The relationship can be quantified by determining a relationship quotient for quantifying a similarity between components of software design artifacts and presenting the quantified relationships to a user, such as a software designer, so that he or she can account for the relationship between such components during design changes and so on. The relationship quotient is made more representative of substantive similarity by selecting the key terms that are to be submitted to a similarity analysis such that words that are too common in the English language, such as conjunctions, articles, etc., are not used. Ubiquity of certain key terms in an enterprise is accounted for by adding a term significance weight to the similarity analysis. The similarity analysis is made contextual, for instance, by the use of inputs from domain ontology including Entity Descriptions, and Entity Relationships.

Claims

1. A computer implemented method for determining relationships between concepts of a plurality of different types of software design artifacts related to a software application design, the method comprising:

receiving data representing a plurality of software design artifacts, wherein the plurality of software design artifacts comprise at least one requirements specification artifact, at least one test plan artifact, at least one use-case artifact, and at least one class specification artifact;

from the data representing the plurality of software design artifacts, extracting at least some concepts correspondingly associated with the plurality of the software design artifacts;

from the at least some extracted concepts, extracting at least some key terms correspondingly associated with the at least some extracted concepts;

determining term significance weights correspondingly associated with the at least some key terms;

at least based on the term significance weights, quantifying relationships between at least some of the extracted concepts of the plurality of the software design artifacts, the quantifying relationships comprising;

generating a matrix of relationship quotients that represent the quantified relationships between at least some of the extracted concepts of the plurality of the software design artifacts;

normalizing value of a relationship quotient with highest value in the matrix to a value of 1 and normalizing values of rest of the relationship quotients proportionally thereto to generate a normalized version of the matrix; and

based on the normalized values of the relationship quotients, classifying each relationship into one of the following three categories;

High level relationship, Medium level relationship, and Low level relationship; and

presenting the quantified relationships to a user via a user interface, including displaying the quantified relationships organized into their associated categories of High level relationship, Medium level relationship, and Low level relationship.

2. The method of claim 1 wherein, the concepts comprise at least one concept selected from a group of concepts consisting of test case concepts, individual use-cases, individual requirements, and individual class descriptions.

3. The method of claim 1 wherein, extracting at least some concepts correspondingly associated with the plurality of the software design artifacts comprises:

applying parsing methods to textual forms of the plurality of the software design artifacts to decompose at least some of the plurality of the software design artifacts into text-based document objects; and

based at least in part on different predefined templates corresponding to different concept types, classifying at least some of the text-based document objects as concept descriptions of the different concept types.

4. The method of claim 3 wherein, extracting the at least some key terms correspondingly associated with the at least some extracted concepts comprises:

parsing at least some of the extracted concepts to decompose the concept descriptions into words and phrases;

storing the decomposed words and phrases as key terms in a key terms database;

comparing the decomposed words and phrases to a stop list; and

deleting from the key terms database those key terms that correspond to words and phrases also present in the stop list.

5. The method of claim 4 further comprising:

applying stemming methods to the key terms of the key terms database to generate stemmed versions of key terms in the key terms data base; and

generating a list of stemmed key terms associated with the at least some of the extracted concepts of the plurality of the software design artifacts.

6. The method of claim 1 wherein, determining term significance weights correspondingly associated with the at least some key terms, comprises:

determining term frequencies of the key terms in the at least some of the concepts; and

determining inverse document frequencies of at least some of the key terms.

7. The method of claim 1 wherein, quantifying the relationships between at least some of the extracted concepts of the plurality of the software design artifacts comprises:

conducting a similarity analysis between the at least some of the extracted concepts of the plurality of the software design artifacts based on the at least some of the key terms.

8. The method of claim 7 wherein, the similarity analysis is based on a semantic meaning of the at least some of the key terms wherein the semantic meaning of the at least some of the key terms is determined by applying data from a domain ontology including entity relationships data and entity description data to the at least some of the key terms.

9. The method of claim 7 wherein, the similarity analysis is a Cosine Measure similarity analysis.

10. A computer system for tracing relationships between concepts of a plurality of different types of software design artifacts related to a software application design, the system comprising:

memory; and

a traceability engine comprising at least one processor that is communicative with the memory and, based on instructions stored in the memory, the traceability engine operable to;

receive data representing a plurality of software design artifacts;

from the data representing the plurality of software design artifacts, extract at least some key terms corresponding to concepts associated with the plurality of the software design artifacts, wherein the plurality of software design artifacts comprise at least one requirements specification artifact, at least one test plan artifact, at least one use-case artifact, and at least one class specification artifact;

determine corresponding term significance weights for at least some of the key terms corresponding to the concepts of the plurality of the software design artifacts;

at least based on the term significance weights, quantify the relationships between concepts associated with the plurality of the software design artifacts, the quantifying relationships comprising;

generating a matrix of relationship quotients that represent the quantified relationships between at least some concepts associated with the plurality of the software design artifacts;

normalizing value of a relationship quotient with highest value in the matrix to a value of 1 and normalizing values of rest of the relationship quotients proportionally thereto to generate a normalized version of the matrix; and

based on the normalized values of the relationship quotients, classifying each relationship into one of the following three categories;

High level relationship, Medium level relationship, and Low level relationship; and

present, to a user via a user interface, a display of the quantified relationships organized into their associated categories of High level relationship, Medium level relationship, and Low level relationship.

11. The system of claim 10 wherein the traceability engine is further operable to:

establish one or more relationships between at least one of the concepts associated with the quantified relationships and at least one other concept associated with explicitly defined relationships between other concepts of other software design artifacts.

12. The system of claim 10 wherein the traceability engine is further operable to performing an impact analysis based on the quantified relationships.

13. The system of claim 10 further comprising a key terms comparison engine operable to compare sets of key words correspondingly associated with a first selected concept and a second selected concept to generate a subset of key words of the first selected concept not found in the set of key words associated with the second selected concept to generate a missing key terms report.

14. The system of claim 10 wherein to extract at least some key terms corresponding to concepts associated with the plurality of the software design artifacts comprises to determine semantic meanings associated with the at least some of the key terms based at least in part on domain ontology associated with the software design artifacts.

15. A computer readable storage medium useful in connection with a computer which includes a processor and a memory, the computer readable storage medium storing computer executable instructions which are configured to cause the computer to process data related to software design artifacts for quantifying relationships between concepts of a plurality of different types of software design artifacts related to a software application design by performing the method steps of:

receiving data representing a plurality of software design artifacts, wherein the plurality of software design artifacts comprise at least one requirements specification artifact, at least one test plan artifact, at least one use-case artifact, and at least one class specification artifact;

from the data representing the plurality of software design artifacts, extracting at least some concepts correspondingly associated with the plurality of the software design artifacts;

from the at least some extracted concepts, extracting at least some key terms correspondingly associated with the at least some extracted concepts;

determining term significance weights correspondingly associated with the at least some key terms;

at least based on the term significance weights, quantifying the relationships between at least some of the extracted concepts of the plurality of the software design artifacts, the quantifying the relationships comprising;

generating a matrix of relationship quotients that represent the quantified relationships between at least some of the extracted concepts of the plurality of the software design artifacts;

normalizing value of a relationship quotient with highest value in the matrix to a value of 1 and normalizing values of rest of the relationship quotients proportionally thereto to generate a normalized version of the matrix; and

based on the normalized values of the relationship quotients, classifying each relationship into one of the following three categories;

High level relationship, Medium level relationship, and Low level relationship; and

presenting the quantified relationships to a user via a user interface, including displaying the quantified relationships organized into their associated categories of High level relationship, Medium level relationship, and Low level relationship.

16. The computer readable storage medium of claim 15 wherein, determining term significance weights correspondingly associated with the at least some key terms, comprises:

determining term frequencies of the key terms in the at least some of the concepts; and

determining inverse document frequencies of at least some of the key terms.

17. The computer readable storage medium of claim 15 wherein, quantifying the relationships between at least some of the extracted concepts of the plurality of the software design artifacts comprises:

conducting a similarity analysis between the at least some of the extracted concepts of the plurality of the software design artifacts based on the at least some of the key terms.

18. The computer readable storage medium of claim 15 wherein, the similarity analysis is based at least in part on a s semantic meaning of the at least some of the key terms.