JP2008117066A - Software development support method, software development support device, software development support program, and computer system - Google Patents

Abstract

When changing or adding specifications of an existing program, development knowledge accumulated in the past is efficiently utilized.
A software development support method for associating a development document as development knowledge with a predetermined program position included in program information, and generating a document characteristic of the development document by performing natural language processing on the input development document. Selecting a similarity point for associating the development document with the program information based on the generated document characteristics, selecting a program position of the program information based on the selected similarity point, and program of the program information The similarity between the information described in the position and the document characteristic is calculated, and when the calculated similarity exceeds a predetermined value, the development document is associated with the program position.
[Selection] Figure 1

Description

  The present invention relates to technology for supporting software design, development, and maintenance.

  In the field of software development, changing and adding specifications to existing programs are widely performed. At that time, even if sufficient reference materials related to the program are not attached, program analysis technology is disclosed to quickly understand the operation of the program and easily confirm that there are no harmful side effects due to the change. ing.

  For example, in the technique disclosed in Patent Document 1, first, the source program is parsed, and the syntax in the program is searched along the execution path, whereby a relationship such as substitution or comparison between variables, constants, and data files is performed. To extract. Then, for each extracted variable, find the location where the variable is set, and even if it is the same variable, search for the relationship separately for each location where the value is set. The analysis accuracy is not degraded in the impact analysis.

  Also, in quality control in semiconductor production, search past case information efficiently from past case database and analyze failure for the purpose of early launch of new products, early operation of new production lines, and maintenance of high yield in mass production. , There is a technology to be used for cause investigation and cause countermeasures.

For example, in the technique disclosed in Patent Document 2, the past result database is searched using a feature amount that is incidental information of text data, image data, and map data as a keyword, so that all the information is not inspected. Past case information having a high degree of coincidence with data can be acquired from the past case database.
JP 2001-142688 A JP-A-10-209230

  However, in Patent Document 1, there is a problem that development knowledge accumulated during past design, development, and maintenance cannot be used efficiently when changing or adding specifications of an existing program.

  Moreover, in patent document 2, there existed a problem that only the past case information regarding the cause of a defect was acquirable.

  The purpose of the present invention is to identify development documents that have a high degree of relevance to the program structure, functional similarity, development history, etc. that suggest various relationships such as the causal relationship of defects, the degree of module coupling, and the cooperative operation of hardware and software There is to do.

  Furthermore, by identifying the program location related to the program structure, functional similarity, development history, etc., and associating and outputting the program location and the development document, the development document is sent to the software and system developer. Is to efficiently refer to.

  In a typical embodiment of the present invention, in a software development support system, a software development support method for associating a development document as development knowledge with a predetermined program position of input program information, the software development system including a processor A memory connected to the processor; an input device that receives input of the program information and the development document; and a storage device that stores the input program information and the development document. Generating a document characteristic of the development document by subjecting the development document to natural language processing, and selecting a similarity viewpoint for associating the development document with the program information based on the generated document characteristic; Based on the selected similarity point of view, the program position of the program information is selected, the degree of similarity between the information described in the program position of the program information and the document characteristic is calculated, and the calculated degree of similarity When the value exceeds a predetermined value, the development document is associated with the program location.

  According to one aspect of the present invention, when a software and system developer refers to a program logical structure, it is also possible to refer to development knowledge highly related to the program logical structure.

  The object of the present invention is to specify a development document that is highly related to information such as a causal relationship and module coupling degree of defects, a program structure that suggests a cooperative operation between hardware and software, function similarity, development history, etc. Have. The present invention (1) selects an appropriate similarity point according to the document characteristics of the development document, (2) narrows down the program position where the similarity calculation with the development document is to be performed based on the similarity point, 3) The intended purpose is realized by calculating the similarity between the program comment at the program position and the development document. Hereinafter, specific embodiments will be described.

(First embodiment)
FIG. 1 is a configuration diagram of a software development support system 100 according to the first embodiment.

  The software development support system 100 includes a CPU 101, a memory 102, and an external interface 110. The CPU 101, the memory 102, and the external interface 110 are connected to each other via a bus 122. The external interface 110 is connected to the input device 111, the output device 112, the external storage device 113, and the communication device 121.

  The input device 111 is a device for a user to input information to the software development support system 100. The input device 111 has a function of receiving input of a source program and a development document, and a function of receiving input of an execution instruction or parameter of a processing program. The output device 112 displays, for example, a result of associating a source program with a development document, and is specifically a display.

  The external storage device 113 holds a program structure analysis unit 103, a program comment extraction unit 104, a development knowledge system generation unit 105, and a development knowledge system search unit 109 as processing programs. The development knowledge system generation unit 105 includes a natural language processing unit 106, a similarity viewpoint selection unit 107, and a similarity calculation unit 108. Hereinafter, it is assumed that all processing programs are read into the memory 102 and executed by the CPU 101 at the time of execution.

  The external storage device 113 includes a source program storage unit 114, a development document storage unit 117, a dictionary information storage unit 118, and a similarity viewpoint rule storage unit 119.

  The source program storage unit 114 stores a source program input from the outside via the input device 111. A source program is a design drawing of software written by a human using a programming language. The source program cannot be executed on the computer as it is, but is converted into a format that can be understood by a computer called object code (native code) by software such as a compiler and executed.

  The development document storage unit 117 stores a development document input from the outside via the input device 111. The development document is a product in the form of a document created accompanying the development of software and systems such as a defect report, a specification, and minutes. The development document includes not only an atypical document but also a structured document such as a template document in which chapters and items are templated, and an XML (eXtensible Markup Language) document. A specific example of the standard document is shown in FIG.

  FIG. 2 is a diagram illustrating an example of a standard document development document according to the first embodiment. The development document shown in FIG. 2 is a failure report that describes the phenomenon, cause, countermeasure, and the like of a failure that occurred in the past.

  The defect report shown in FIG. 2 is composed of items and contents corresponding to the items. Items include “Title”, “Phenomenon”, “Cause”, “Countermeasure”, and “Recurrence prevention”. In this way, the item and the keyword included in the content corresponding to the item are used to associate with the program information.

  The dictionary information storage unit 118 stores dictionary information input from the outside via the input device 111. The dictionary information is, for example, information obtained by digitizing a synonym dictionary, a syntactic dictionary, or a thesaurus (a dictionary in which words are organized by meaning instead of characters), and is a knowledge system for supporting natural language processing.

  The similarity viewpoint rule storage unit 119 stores a similarity viewpoint rule input from the outside via the input device 111.

  When associating program information with development documents, the relationship between the two is compared from various perspectives. The similarity viewpoint is a viewpoint focused on associating the program information with the development document. The program structure and functional similarity suggesting the cause-and-effect relationship of the defect, the degree of module coupling, and the cooperative operation of hardware and software. And development history. Specifically, if the development document describes a specific defect, a “program structure” that easily causes the defect can be set as a similarity viewpoint. In addition, if there is a process having a function similar to the process in which the defect occurs, the development document describing the specific defect is associated with the process having the similar function from the viewpoint of similarity of “function similarity”. be able to.

  The similarity is a numerical value of the degree of similarity between the contents described in the program information and the contents described in the development document from a certain similarity viewpoint. For example, if the similarity viewpoint is “causal relationship between defects”, the degree of similarity between the process structure defined in the program information and the process structure of the cause of failure described in the development document is quantified It becomes.

  The similarity viewpoint rule is an index for selecting a similarity viewpoint that is highly related to the development document. The similarity viewpoint rule is an association between the document characteristics of the development document (for example, keyword content, appearance position or appearance distribution) and the similarity viewpoint, for example, expressed by an If-Then rule. The If-Then rule is composed of a conditional part and a conclusion part. If the document characteristics of the development document satisfy the condition part, the similarity point of the conclusion part and the relevance of the development document are considered sufficiently high. Details of the similarity viewpoint rule will be described later with reference to FIG.

  Further, the external storage device 113 includes a program structure information storage unit 115 and a program comment storage unit 116. The program structure analysis unit 103 analyzes the structure of the source program read from the source program storage unit 114 and stores the result in the program structure information storage unit 115. Program structure information means the definition positions of functions and variables included in the program, the call relationship between functions, the assignment / reference relationship of variables, etc., and the software and system developers understand the program's logical structure. It is information for supporting.

  The program comment extraction unit 104 extracts a program comment from the source program read from the source program storage unit 114 and stores the program comment in the program comment storage unit 116 of the external storage device 113. The program comment is memo information described for the purpose of explaining the contents of the program separately from the syntax describing the processing to be executed in the source program. Program comments are generally identified by a programming language specific identifier. For example, in the Java (registered trademark) language, a period between “//” and a line feed and a period between “/ *” and “* /” are interpreted as program comments.

  The natural language processing unit 106 refers to the dictionary information read from the dictionary information storage unit 118 as necessary, and processes the program comments and development documents read from the program comment storage unit 116 and the development document storage unit 117 with natural language processing. To do. Natural language processing is a technique for causing a computer to process or understand natural language. For example, basic techniques for processing Japanese include morphological analysis, syntactic analysis, disambiguation of meaning, and anaphora analysis.

  The similarity viewpoint selection unit 107 applies the similarity viewpoint selection rule read from the similarity viewpoint rule storage unit 119 to the natural language processing result of the development document, and the degree of association with the development document exceeds a predetermined value. Select a similarity perspective.

  The similarity calculation unit 108 compares the selected similarity viewpoint with the program structure information read from the program structure information storage unit 115, and selects a program position that matches the selected similarity viewpoint. The program position is mainly a position on a source program where processing is described, and may be a position of program information (for example, a design document such as a flowchart) corresponding to the source program.

  For example, if the similarity viewpoint is a specific “global table sharing”, the program position suitable for the similarity viewpoint is all the positions sharing the global table. Similarly, depending on the selected similarity point of view, all the positions sharing the interface, the mutual definition positions of the parent and child functions, all the positions where the timing judgment conditions are similar, and the counter reset conditions are similar All the positions that have been changed, and all the positions whose specifications have been changed at the same time are the program positions that fit the similarity viewpoint.

  The similarity calculation unit 108 calculates the similarity between the program comment at the selected program position and the development document. Further, a development document whose similarity with a program comment exceeds a predetermined value is regarded as development knowledge at the program position where the program comment is described, and is written in the development knowledge system storage unit 120 of the external storage device 113.

  The development knowledge system search unit 109 searches the development knowledge system storage unit 120 for a development document corresponding to the development knowledge at the selected program position, and outputs the development document in association with the program position from the output device 112.

  Next, the operation of each part constituting the software development support system 100 will be described.

  FIG. 3 is a flowchart showing a procedure for associating program information such as the source program of the first embodiment with a development document. This process is executed at the back end of the software development support system 100.

  First, the software development support system 100 receives an input of a source program from the input device 111 (step 300). The input source program is stored in the source program storage unit 114. At this time, program structure information such as a program specification may be input instead of the source program. When the program structure information is input, the program structure analysis process described later need not be executed.

  Next, the program structure analysis unit 103 reads the source program stored in the source program storage unit 114, analyzes the structure of the program, and generates program structure information (step 301). The generated program structure information is stored in the program structure information storage unit 115.

  The program comment extraction unit 104 reads the source program stored in the source program storage unit 114 and extracts a program comment from the source program (step 302). The extracted program comment is stored in the program comment storage unit 116.

  Subsequently, the software development support system 100 accepts input of dictionary information from the input device 111 (step 303). The input dictionary information is stored in the dictionary information storage unit 118.

  Further, the natural language processing unit 106 reads out the program comments and dictionary information from the program comment storage unit 116 and the dictionary information storage unit 118, and performs natural language processing on the program comments while referring to the dictionary information as necessary (step 304). ). The program comment subjected to the natural language processing is stored in the program comment storage unit 116.

  The software development support system 100 receives an input of the similarity viewpoint rule from the input device 111 (step 305). The input similarity viewpoint rule is stored in the similarity viewpoint rule storage unit 119.

  The software development support system 100 receives an input of a development document from the input device 111 (step 306). The input development document is stored in the development document storage unit 117.

  The natural language processing unit 106 reads the development document and dictionary information from the development document storage unit 117 and the dictionary information storage unit 118, and performs natural language processing on the development document while referring to the dictionary information as necessary (step 307). The development document subjected to natural language processing is stored in the development document storage unit 117.

  Subsequently, the similarity viewpoint selecting unit 107 selects a similarity viewpoint having a high degree of association with the development document (step 308). Specifically, the similarity viewpoint selecting unit 107 first reads out the natural language processing result and similarity viewpoint rule of the development document from the development document storage unit 117 and the similarity viewpoint rule storage unit 119. Further, a rule that satisfies the condition part of the similarity viewpoint rule is extracted from the natural language processing result of the development document, and a conclusion part of the rule is selected as a similarity viewpoint that has a relevance degree with the development document exceeding a predetermined value. Then, the similarity viewpoint selecting unit 107 holds the selected similarity viewpoint.

  Here, the similarity viewpoint rule will be specifically described with reference to FIG. 4, and a procedure for specifically selecting the similarity viewpoint will be described.

  FIG. 4 is a table showing a specific example of the similarity viewpoint rule of the first embodiment. The If-Then rule described above is adopted as the similarity viewpoint rule of the first embodiment, and is configured by the condition unit 500 and the conclusion unit 501.

  The condition part 500 includes data / item 502 and a keyword 503, and stores the document characteristics of the development document. The development document is subjected to natural language processing to extract keywords corresponding to data or items. Therefore, the similarity viewpoint selecting unit 107 compares the data or items and keywords extracted from the development document with the data / items 502 and keywords 503 included in the condition unit 500, and selects the conclusion unit 501.

  The conclusion unit 501 includes a major classification 504 and a minor classification 505, and stores similarity viewpoints. The conclusion unit 501 classifies and stores similarity viewpoints related to information stored in the data / item 502 and the keyword 503 into a large classification 504 and a small classification 505.

  Specifically, referring to the similarity viewpoint rule record 506, “not specified” is registered in the data / item 502, and “configuration information and recovery failure” is registered in the keyword 503. Since a value is not specified in the data / item 502, if the keywords “configuration information” and “recovery failure” are included in the entire development document subjected to natural language processing, the condition part is satisfied. In this case, the similarity degree selected is “program structure” for the major classification 504 and “shared global table” for the minor classification 505. That is, when the keywords “configuration information” and “recovery failure” are included in the development document, it can be seen that attention should be paid to processing related to “program structure” and “global table sharing”.

  Now, the description returns to the flowchart of FIG.

  The similarity calculation unit 108 selects a program position corresponding to the program structure information that matches the similarity viewpoint selected by the similarity viewpoint selection unit 107 in the process of step 308 (step 309). Specifically, the similarity calculation unit 108 first reads out the similarity viewpoint selected by the similarity viewpoint selection unit 107 in the process of step 308. Further, the program structure information is read from the program structure information storage unit 115. Then, the similarity viewpoint selected in the process of step 308 is compared with the program structure information, a program position that matches the similarity viewpoint is selected, and the result is held.

  Subsequently, the similarity calculation unit 108 calculates the similarity between the contents of the program position selected in the process of step 309 and the development document (step 310). Specifically, the similarity calculation unit 108 first reads the natural language processing result of the program comment from the program comment storage unit 116, and further reads the natural language processing result of the development document from the development document storage unit 117. Then, the degree of similarity between the contents of the program position selected in the processing of step 309 and the development document is calculated and held in the degree of similarity calculation unit 108.

  As a method of calculating the similarity between the program comment and the development document, for example, there is a method based on the number of keywords that match between the program comment and the development document. At this time, when the number of matching keywords is large, weighting may be performed for each keyword. It may also be based on the ratio of keywords that match between program comments and development documents. The ratio of matching keywords can be, for example, the ratio of the number of matching keywords to the total number of keywords included in one or both of the program comment and the development document.

  The similarity between the program structure information and the development document can be calculated, for example, based on the described syntax and variables.

  The similarity calculation unit 108 stores the combination information of the program comment and the development document in the development knowledge system storage unit 120 based on the similarity between the program comment and the development document calculated by the process in step 310 (step 311). Specifically, when the similarity between the program comment and the development document exceeds a predetermined value, the development document is regarded as development knowledge at the program position where the program comment is described.

  The software development support system 100 repeats the processing from step 307 to step 311 for the input development document (step 312). When the processing is completed for all the input documents (the result of step 312 is “NO”), this processing ends (step 313).

  The above is the description of the process of associating the program information with the development document and recording it as development knowledge. Hereinafter, a process of outputting the accumulated development knowledge in association with program information will be described.

  FIG. 5 is a flowchart illustrating a procedure for displaying a related development document by designating a program position according to the first embodiment.

  The software development support system 100 accepts input of program position information by the input device 111 (step 400). The input program position is recorded in the development knowledge system search unit 109.

  The development knowledge system search unit 109 searches the development knowledge system storage unit 120 for the combination information of the program position and the development document, and acquires the development document corresponding to the input program position (step 401).

  The development knowledge system search unit 109 outputs the development document acquired in the process of step 401 from the output device 112 in association with the corresponding program position (step 402), and then ends the process (step 403).

  Further, as a method of outputting a development document in association with a program position, for example, there is a method of linking a development document from within a source program displayed in an editor. A specific example of a method for outputting a development document in association with a program position is shown in FIG.

  FIG. 6 is a diagram illustrating a specific example in which a development document is output in association with a program position according to the first embodiment. The left side of the figure is an editor screen for displaying and editing a source program, and the right side shows related documents.

  The left side of the figure displays a program written in the JAVA language. A comment line starting with “//” describes “exclusive control logic D”. In the development knowledge system storage unit 120, combination information of the program position corresponding to the comment line and the “defect report 00X” displayed on the right side of the drawing is recorded.

  Therefore, when the development knowledge system search unit 109 selects a comment line on the editor, the processing shown in FIG. 5 is executed, and “defect report 00X” is searched as a related development document, as shown in FIG. indicate.

  Since “exclusion control logic D” is included in “countermeasure method” in “defect report 00X”, “exclusive control logic D” is extracted as a keyword when “defect report 00X” is processed in natural language. Is done. Therefore, if “exclusive control logic D” is described as a comment in the source program, “defect report 00X” is retrieved as a development document having a high similarity to the program position.

  According to the first embodiment, when a software and system developer confirms a program logical structure, it is also possible to confirm development knowledge related to the program logical structure. As a result, the relationship between each combination for each software part corresponding to the cause and effect of the defect that is obvious or can be estimated from the program logical structure, the specification change and influence, the specification provider and the use destination It is possible to associate development documents that deal with problems and precautions caused by sex as development knowledge.

  In addition, according to the first embodiment, during software and system design, development, and maintenance, development knowledge that refers to problems and cautions resulting from the relationship between the cause of the failure and the affected party Can be provided.

  Furthermore, according to the first embodiment, development knowledge that refers to problems and cautions resulting from the relationship between the specification change source and the influence destination or the specification supply source and the use destination is provided. It becomes possible to do.

  In addition, according to the first embodiment, development knowledge that refers to problems and cautions resulting from the relationship between the two is provided to software parts that are susceptible to changes in hardware specifications and defects. Is possible.

(Second Embodiment)
In the first embodiment, the similarity viewpoint rule input from the input device 111 is used. In the second embodiment, the similarity viewpoint rule is automatically generated based on the development document and the source program to which the similarity viewpoint rule is input. To generate. Further, before the similarity viewpoint selecting unit 107 reads out the natural language processing result and similarity viewpoint rule of the development document from the development document storage unit 117 and the similarity viewpoint rule storage unit 119, the similarity degree is internally set in the software development support system 100. Generate viewpoint rules. Hereinafter, description of processes common to the first embodiment will be omitted, and only differences will be described.

  FIG. 7 is a configuration diagram of the software development support system 100 according to the second embodiment.

  In addition to the configuration of the first embodiment, the software development support system 100 of the second embodiment includes a similarity viewpoint rule generation unit 123, a document characteristic information storage unit 124, and a document group information storage unit 125 in the external storage device 113. A common program characteristic information storage unit 126 and a table combination result storage unit 127.

  The similarity viewpoint rule generation unit 123 is a program that defines a procedure for generating a similarity viewpoint rule. The similarity viewpoint rule generation unit 123 generates information to be stored in the document characteristic information storage unit 124, the document group information storage unit 125, the common program characteristic information storage unit 126, and the table combination result storage unit 127. A specific procedure for generating the similarity viewpoint rule will be described with reference to FIG.

  The document characteristic information storage unit 124 stores document characteristic information such as keywords extracted by processing a development document with natural language. The document characteristic information is, for example, keyword content, appearance position, or appearance distribution. A specific example of document characteristic information will be described with reference to FIG.

  The document group information storage unit 125 stores information obtained by grouping development documents based on the document characteristic information stored in the document characteristic information storage unit 124. A specific example of development document grouping information will be described with reference to FIG.

  Based on the grouping information stored in the document / group information storage unit 125, the common program characteristic information storage unit 126 analyzes the program information at the program location explicitly associated with the development document belonging to each group. Program characteristic information is extracted and stored as program characteristic information highly related to the development document group. A specific example of program characteristic information having a high degree of association with the development document group will be described with reference to FIG.

  The table combination result storage unit 127 combines document characteristic information of development documents, grouping information, and program characteristic information having a high degree of association with the grouped development documents, and information on a combination of document characteristic information and program characteristic information Is stored. A specific example of information of a combination of document characteristic information and program characteristic information will be described with reference to FIG.

  FIG. 8 is a diagram illustrating a specific example of document characteristic information stored in the document characteristic information storage unit 124 according to the second embodiment. Document characteristic information includes data / items, keywords, and application frequency. Specifically, referring to FIG. 8, it is understood that the keywords “configuration information” and “recovery failure” appear five times in the “defect report 001” as the entire document.

  FIG. 9 is a diagram illustrating a specific example of the grouping information of the development document according to the second embodiment. The document grouping information includes the name of the development document group and the names of the development documents constituting the development document group. Specifically, referring to FIG. 9, “Group A” includes “Defect Report 001”, “Defect Report 005”, and “Defect Report 036”.

  FIG. 10 is a diagram illustrating a specific example of correspondence information between the development document group and the program characteristic information having a high degree of association according to the second embodiment. Program characteristic information is composed of a large classification and a small classification. Specifically, referring to FIG. 10, “group A” corresponds to major classification “program structure” and minor classification “shared global table” as program characteristic information.

  FIG. 11 is a diagram illustrating a specific example of combination information of document characteristic information and program characteristic information according to the second embodiment. The combination information of the document characteristic information and the program characteristic information is generated by combining the information shown in FIGS. Specifically, the “development document” in the document characteristic information in FIG. 8 is associated with the “development document” in the grouping information in the development document in FIG. Further, the “development document group” of the development document grouping information in FIG. 9 and the “development document group” of the correspondence information between the development document group and the program characteristic information in FIG. As a result, for example, the development document “defect report 001” corresponds to the program characteristic information of the major classification “program structure” and the minor classification “global table sharing”.

  FIG. 12 is a flowchart illustrating a procedure for generating a similarity viewpoint rule according to the second embodiment. The procedure for associating the program information such as the source program with the development document is the same as the procedure of the first embodiment shown in FIG. In this process, as described above, instead of reading the input similarity viewpoint rule, the similarity viewpoint rule is generated before the process of step 308 in FIG. 3 is executed. Hereinafter, a procedure for generating the similarity viewpoint rule will be described.

  First, the similarity viewpoint rule generation unit 123 reads a natural language processing result for one development document from the development document storage unit 117, and generates document characteristic information of the development document (step 1200). The extracted document characteristic information is stored in the document characteristic information storage unit 124 of the external storage device 113.

  As a method for generating document characteristic information, for example, keywords that are high in appearance frequency in a single development document but low in average appearance frequency in the entire development document are extracted, or only within a specific item in the entire development document. Extract and generate keywords with high appearance frequency.

  The similarity viewpoint rule generation unit 123 determines whether there is a development document for which no document characteristic information has been generated among all the development documents stored in the development document storage unit 117 (step 1201). When there is an unprocessed development document in the development document storage unit 117 (the result of step 1201 is “Yes”), the similarity viewpoint rule generation unit 123 repeats the process of step 1200. When document characteristic information is generated for all the development documents (the result of step 1201 is “No”), the processing of step 1202 is executed.

  The similarity viewpoint rule generation unit 123 groups the development documents (step 1202). Specifically, the document characteristic information of all the development documents is read from the document characteristic information storage unit 124, and the similarity between the development documents is calculated using the document characteristic information. Then, development documents whose similarity exceeds a predetermined value are extracted and grouped, and stored in the document group information storage unit 125 as grouping information of the development document.

  Here, as a method of calculating the degree of similarity between the developed documents, for example, the number or ratio of keywords (can be weighted) that match in both document characteristic information (the number of matching keywords with respect to the total number of keywords included in both document characteristic information) For example). When grouping development documents, a single development document may belong to a plurality of groups.

  The similarity viewpoint rule generation unit 123 generates program characteristic information having a high degree of association with the development document group (step 1203). Specifically, the grouping information for one case is read from the document / group information storage unit 125. It collects information such as program structure information, program comments, and development history at all program locations that are explicitly associated with one or more development documents that make up the group, and program characteristics common to the collected information Information is extracted and written into the common program characteristic information storage unit 126 as program characteristic information having a high degree of association with the development document group.

  Here, examples of the program location information explicitly associated with the development document include a failure occurrence location described in the failure report or a mounting location described in the specification. As a method of extracting program characteristic information common to many program positions, for example, there is a method of extracting program characteristics common to program positions exceeding a predetermined number or ratio.

  The similarity viewpoint rule generation unit 123 determines whether or not program characteristic information having a high degree of relevance has been extracted for all development document groups stored in the document group information storage unit 125 (step 1204). If there is an unprocessed development document group (the result of step 1204 is “YES”), the process of step 1203 is repeated. If program characteristic information having a high degree of relevance is extracted for all development document groups, the processing in step 1205 is executed.

  The similarity viewpoint rule generation unit 123 generates combination information of document characteristics and program characteristics information (step 1205). Specifically, the document characteristic information of the development document is related from the document characteristic information storage unit 124, the grouping information of the development document is stored from the document group information storage unit 125, and the development document group is related from the common program characteristic information storage unit 126. Read high-quality program characteristic information. Then, using the grouping information of the development document as a join key, the document characteristic information of the table-type development document and the program characteristic information highly related to the table-type development document group are combined, and the document characteristic and the program characteristic information The combination information is written in the table combination result storage unit 127.

  Finally, the similarity viewpoint rule generation unit 123 selects a representative combination from the combination information of the document characteristic and the program characteristic information stored in the table combination result storage unit 127 and regards it as a similarity viewpoint rule. Write to the viewpoint rule storage unit 119 (step 1206). As a method for selecting representative combination information, for example, there is a method for selecting combination information exceeding a predetermined number or ratio.

  Thereafter, the similarity viewpoint rule generation unit 123 ends the similarity viewpoint rule generation processing (step 1207). Then, the software development support system 100 executes the processing after step 308 in FIG. 3 to construct a development knowledge system.

  According to the second embodiment, it is possible to reduce the labor for creating and inputting the similarity viewpoint rule by automatically generating the similarity viewpoint rule based on the input development document.

  In addition, according to the second embodiment, the accuracy increases as the amount of development documents and source programs that become input data at the time of similarity viewpoint rule generation increases, and the quality of the similarity viewpoint rule improves. I can expect that. Furthermore, even when the characteristics of the development document and the source program change over time, the changed contents can be automatically reflected in the similarity viewpoint rule.

(Third embodiment)
In the first and second embodiments, the development document searched by the development knowledge system search unit 109 is output from the output device 112 in association with the program position. In the third embodiment, the development document is related to the program structure information. Output the development document.

  Specifically, in the process of step 402 in the flowchart of FIG. 5, the development knowledge system searching unit 109 reads program structure information corresponding to the corresponding program position from the program structure information storage unit 115. Then, the development document searched in the process of step 401 is output from the output device 112 in association with the program structure information at the corresponding program position.

  As a method of outputting a development document in association with program structure information, for example, when displaying a function list, a variable list, a call relationship between functions, a variable assignment / reference relationship, etc. as program structure information, There is a method to link to the development document from the display part of the variable.

  According to the third embodiment, when a software and system developer confirms a program logical structure, it is also possible to confirm development knowledge having a high degree of association with the program logical structure. Specifically, for example, for each software part corresponding to the cause and influence of a defect that is obvious or can be estimated from the program logic structure, the specification change and influence, the specification provider and the use destination, It becomes possible to associate development documents dealing with problems and cautions resulting from the relationship of the combinations as development knowledge.

(Fourth embodiment)
In the first to third embodiments, the development document is output from the output device 112 by the development knowledge system search unit 109 in association with the corresponding program position or program structure information.

  In the fourth embodiment, in the process of step 311 in FIG. 3, the similarity calculation unit 108 uses the development knowledge system to combine the combination information of the similarity with the program position and the development document having a similarity exceeding a predetermined value. Store in the storage unit 120.

  Further, the development knowledge system search unit 109 reads combination information of the program position, the development document, and the similarity from the development knowledge system storage unit 120 in the process of step 401 in FIG. Further, the degree of similarity with the development document corresponding to the development knowledge of the program position equal to the input program position information is searched.

  Further, the development knowledge system search unit 109 associates the development document searched in the process of step 401 with the corresponding program position or program structure information in the process of step 402, and outputs it from the output device 112 in descending order of similarity. To do.

  According to the fourth embodiment, when the number of development documents associated with program location or program structure information is large, the importance of each development document as development knowledge is replaced by similarity, for example, similarity It is possible to prioritize development documents, such as confirming only development documents that exceed a predetermined value.

(Fifth embodiment)
In the first to fourth embodiments, only the software development support system is a single configuration. In the fifth embodiment, in addition to the software development support system 100, a host computer and a storage subsystem connected via a network are used. Includes system.

  FIG. 13 is a configuration diagram of a software development support system 100 according to the fifth embodiment.

  The system of the fifth embodiment includes a software development support system 100, a host 209, and a storage subsystem 210. The software development support system 100, the host 209, and the storage subsystem 210 are connected to each other via a network 221.

  The software development support system 100 has the same configuration as that of the first embodiment shown in FIG. 1 and has equivalent functions. The external storage device 113 includes a hardware / software (H / S) log storage unit 128 that stores a log of processing executed by the storage subsystem 210.

  The host 209 is connected to the software development support system 100 and the storage subsystem 210 via the network 221. The host 209 accesses resources stored in the storage subsystem 210 and executes predetermined business processing.

  The storage subsystem 210 includes a CPU 211, a memory 212, and an external interface 216. The CPU 211, the memory 212, and the external interface 216 are connected to each other via the bus 220. An external storage device 217, a hardware / software log storage device 218, and a communication device 219 are connected to the external interface 216.

  The memory 212 includes a system control unit 213, a cache 214, and a hardware / software log generation unit 215.

  The system control unit 213 controls data input / output from the host 209 to the external storage device 217. The cache 214 is a temporary storage area for speeding up data input / output from the host 209 to the external storage device 217.

  The hardware / software log generation unit 215 generates a log of normal processing and abnormal processing regarding both hardware and software constituting the storage subsystem, and records the log in the hardware / software log storage device 218.

  The software development support system 100 reads the hardware / software log from the hardware / software log storage device 218 of the storage subsystem 210 via the network 221 and writes it to the hardware / software log storage unit 128 of the external storage device 113. .

  In the first to fourth embodiments, the natural language processing unit 106 performs natural language processing on the program comment and writes the result in the program comment storage unit 116 (step 304 in FIG. 3). Further, the similarity calculation unit 108 calculates the similarity between the program comment explaining the program content at the selected program position and the development document (step 310 in FIG. 3).

  On the other hand, in the fifth embodiment, the natural language processing unit 106 reads a program comment from the program comment storage unit 116 and reads a hardware / software log from the hardware / software log storage unit 128. The natural language processing unit 106 performs natural language processing on the read program comments and hardware / software log, and writes the results in the program comment storage unit 116 and the hardware / software log storage unit 128 (step 304 in FIG. 3). ).

  Further, the selected program position held by the similarity calculation unit 108 is read, and the program comment explaining the program contents at the program position and the similarity between the hardware / software log and the development document are calculated (step 310 in FIG. 3). ).

  As a method of calculating the similarity between the program comment and the hardware / software log and the development document, for example, first, the similarity between the program comment and the development document is set as the first similarity. Further, the similarity between the hardware / software log explicitly associated with the program processing at the selected program position and the development document is set as the second similarity. Then, there is a method of calculating a weighted sum of the first similarity and the second similarity and regarding the result as the similarity between the program comment and the hardware / software log and the development document.

  According to the fifth embodiment, not only can a development knowledge system related to software be constructed from program structure information and program comments, but also a hardware / software log, so that a development knowledge system related to both hardware and software is targeted. Can be built. Specifically, for example, a development document that deals with problems and cautions resulting from the relationship between both can be associated as development knowledge with a software part that is susceptible to changes in hardware specifications and defects. .

  The present invention is applicable to hardware design, development, and maintenance support applications that involve development document generation, as well as software and system design, development, and maintenance support applications. The first to fifth embodiments can be used by being stored in a storage medium as a computer-readable program. Further, it can be distributed through a network as a computer-readable program.

The lineblock diagram of the software development support system of a 1st embodiment. The figure which shows an example of the development document of the fixed form document of 1st Embodiment. 5 is a flowchart showing a procedure for associating program information such as a source program according to the first embodiment with a development document. The figure which shows the specific example of the similarity viewpoint rule of 1st Embodiment. The flowchart which shows the procedure which designates the program position of 1st Embodiment, and displays a related development document. The figure which shows the specific example which outputs a development document linked | related with the program position of 1st Embodiment. The block diagram of the software development assistance system of 2nd Embodiment. The figure which shows the specific example of the document characteristic information stored in the document characteristic information storage part of 2nd Embodiment. The figure which shows the specific example of the grouping information of the development document of 2nd Embodiment. The figure which shows the specific example of the corresponding | compatible information with the program document information with high relevance degree with the development document group of 2nd Embodiment. The figure which shows the specific example of the combination information of the document characteristic information and program characteristic information of 2nd Embodiment. The flowchart explaining the procedure which produces | generates the similarity viewpoint rule of 2nd Embodiment. The block diagram of the software development support system of 5th Embodiment.

Explanation of symbols

100 Software Development Support System 101 CPU
DESCRIPTION OF SYMBOLS 102 Memory 103 Program structure analysis part 104 Program comment extraction part 105 Development knowledge system generation part 106 Natural language processing part 107 Similarity viewpoint selection part 108 Similarity degree calculation part 109 Development knowledge system search part 110 External interface 111 Input device 112 Output device 113 External storage device 114 Source program storage unit 115 Program structure information storage unit 116 Program comment storage unit 117 Development document storage unit 118 Dictionary information storage unit 119 Similarity viewpoint rule storage unit 120 Development knowledge system storage unit 121 Communication device 122 Bus 209 Host 210 Storage subsystem 211 CPU
212 Memory 213 System Control Unit 214 Cache 215 Hardware / Software Log Generation Unit 216 External Interface 217 External Storage Device 218 Hardware / Software Log Storage Device 219 Communication Device 220 Bus 221 Network

Claims (15)

In a software development support system, a software development support method for associating a development document as development knowledge with a predetermined program position included in program information,
The software development system includes a processor, a memory connected to the processor, and an input device that receives input of the program information and the development document.
The processor is
Generating a document characteristic of the development document by subjecting the development document to natural language processing;
Based on the generated document characteristics, select a similarity viewpoint for associating the development document with the program information,
Based on the selected similarity point of view, select a program position of the program information,
Calculating the similarity between the information described in the program location of the program information and the document characteristics;
A software development support method comprising associating the development document with the program position when the calculated similarity exceeds a predetermined value. The software development system further includes an output device,
2. The software development support method according to claim 1, wherein the output device outputs the development document in association with the program location. The input device receives an input of a predetermined program position of the program information;
3. The software development support method according to claim 2, wherein the output device outputs the input program position in association with the development document.   The software development support method according to claim 1, wherein the program information is a source program. The processor is
Analyzing the source program to generate program structure information,
Selecting the program structure information that satisfies the selected similarity point of view,
5. The software development support method according to claim 4, wherein a position corresponding to the selected program structure information is set as the program position. The processor is
Extracting a program comment described in the program location;
Natural language processing of the extracted program comments;
Calculating a similarity between the natural language processed program comment and the document characteristics;
6. The software development support method according to claim 5, wherein when the calculated similarity exceeds a predetermined value, the position where the program comment is described is associated with the development document.   The similarity is included in the number of keywords that match between the natural language processed program comment and the document characteristics, or in either or both of the natural language processed program comment and the development document. The software development support method according to claim 6, wherein the software development support method is calculated based on a ratio of the number of matching keywords to the total number of keywords.   The software development support method according to claim 1, wherein the similarity viewpoint is selected based on a rule input from the outside by the input device. The input device accepts an input of a program position explicitly associated with the development document from the outside,
The processor is
Group the development documents based on the document characteristics;
Analyzing the program information of the program location explicitly associated with the grouped development document to extract program characteristics,
A rule for selecting the similarity viewpoint based on the document characteristics by associating the document characteristics of the grouped development documents with the extracted program characteristics and considering the program characteristics as the similarity viewpoint. Generate
The software development support method according to claim 1, wherein the similarity viewpoint is selected based on the generated rule.   The software development according to claim 1, wherein the similarity viewpoint is selected based on a rule configured by a condition part corresponding to the document characteristic and a conclusion part corresponding to the similarity viewpoint. Support method.   The software development support method according to claim 1, wherein the processor outputs development documents associated with the program positions in descending order of the similarity. The software development support system includes:
It has an interface to connect to the network,
Connected to the storage subsystem via the network, and obtains log information of hardware or software processing executed in the storage subsystem;
The processor is
Natural language processing is performed on the information described in the program location of the program information and the log information generated when the program at the program location is executed,
Calculating the similarity between the natural language processed program comment and the log information and the document characteristics;
The software development support method according to claim 1, wherein when the calculated similarity exceeds a predetermined value, the development document is associated with the program position. A software development support apparatus that associates a development document as development knowledge with a predetermined program position included in program information,
A processor, a memory connected to the processor, and an input device that receives input of the program information and the development document,
The processor is
Generating a document characteristic of the development document by subjecting the development document to natural language processing;
Based on the generated document characteristics, select a similarity viewpoint for associating the development document with the program information,
Based on the selected similarity point of view, select a program position of the program information,
Calculating the similarity between the information described in the program location of the program information and the document characteristics;
A software development support apparatus, wherein when the calculated similarity exceeds a predetermined value, the development document is associated with the program position. A software development support program for causing a computer to execute a process of associating a development document as development knowledge with a predetermined program position included in program information,
Generating document characteristics of the development document by subjecting the development document to natural language processing;
Selecting a similarity perspective for associating the development document with the program information based on the generated document characteristics;
Selecting a program location of the program information based on the selected similarity point of view;
Calculating the similarity between the information described in the program location of the program information and the document characteristics;
A software development support program causing a computer to execute a procedure of associating the development document with the program position when the calculated similarity exceeds a predetermined value. A host computer, a storage device for storing data read / written by the host computer via a network, and a predetermined program location included in a program connected to the storage device via the network and executed by the host computer A software development support apparatus for supporting development by associating a development document with development knowledge, and a computer system comprising:
The storage device includes a first interface connected to the network, a first processor connected to the first interface, and a first memory connected to the first processor,
The first processor records log information of processing executed based on a request from the host computer,
The software development support apparatus includes: a second interface connected to the network; a second processor connected to the second interface; a second memory connected to the second processor; An input device for receiving input of program information and the development document,
The second processor is
Obtain log information of hardware or software processing recorded in the storage device,
Generating a document characteristic of the development document by subjecting the development document to natural language processing;
Based on the generated document characteristics, select a similarity viewpoint for associating the development document with the program information or the log information,
Based on the selected similarity point of view, select a program position of the program information,
Calculating the similarity between the information described in the program location of the program information, the log information generated when the program at the program location is executed, and the document characteristics;
A computer system characterized in that, when the calculated similarity exceeds a predetermined value, the development document is associated with a program position of the program information.

Images