WO2000050988A1 - Apparatus for preparing program specifications, method for preparing program specifications, and recording medium on which program for preparing program specifications is recorded - Google Patents

Abstract

A separator for identifying various comments inserted in a source program is set up, and a keyword for extracting the comment item corresponding to the program specifications is set up for every comment. Therefore, it is made clear that the comment item belongs to which comment, and the range of the comment is correctly and reliably recognized. By extracting the comment item from the range of each comment, manual correction of the extracted comment item is dispensed with, and the efficiency of preparing program specifications is improved.

Description

 Description Program specification creation device, program specification creation method, and recording medium on which program specification creation program is recorded

Technical field

 The present invention relates to a technique for creating a program specification from comments inserted into a source program in the field of re-engineering and reverse engineering, and more particularly to a technique for improving the efficiency of creating a program specification.

Background art

 Conventionally, a comment describing the processing contents of a program is generally inserted into a source program. There are comments on files and comments on symbols. Here, “symbol” means that if the source program is written in C ++ language, it has one property in common, such as class, member function, external variable, etc. A collection of such elements. In general, comments about files are inserted at the beginning of the file, and comments about symbols are inserted immediately before the symbol.

 Therefore, at the beginning of the file, the comment about the file and the comment about the symbol are adjacent to each other, which has the following problems. That is, each comment includes a comment item corresponding to the program specification (hereinafter referred to as “item”). For this reason, when the same item is included in each comment, it is difficult to determine which item belongs to which comment, and it has been impossible to extract the item, or the item has been extracted redundantly. Due to such a defect, even if an attempt was made to create a program specification based on the extracted items, the items could not be used as they were, and manual correction work was indispensable.

In view of the foregoing, an object of the present invention is to clarify the comment to which each item belongs, thereby improving the extraction accuracy of the item and improving the efficiency of creating a program specification. Disclosure of the invention

 In order to achieve the above object, as a first means, a separator setting means for setting a separator for identifying various comments inserted in the source program, and a separator set in the program specification for each separator set by the separator setting means. Key code setting means for setting an extraction key of a corresponding comment item; and comment range extracting means for extracting a range of each comment of the source program based on the separator set by the separator setting means. Comment item extracting means for extracting a comment item from a source program based on the extraction key word set by the key word setting means for each comment range extracted by the comment range extracting means; Based on the comment items extracted by the comment item extraction means, Characterized by being configured to Purodaramu specification creation device includes a Purodaramu specification generator means for generating a beam specifications, the.

 According to such a configuration, to efficiently create a program specification from the comments entered in the source program, the following may be performed. That is, as a work for describing the source program, a separator for identifying the comment type is inserted into the source program. Then, when creating a program specification, if a separator is set by the separator setting means, the range of each comment is extracted based on the set separator. On the other hand, a comment item is extracted from the source program for each range of the extracted comment based on the extraction key set by the keyword setting means. Then, a program specification is generated based on the extracted comment items.

At this time, even if comments of different types are adjacent to each other at the beginning of the file, each comment range can be accurately and reliably identified by the separator for identifying both. Therefore, it is prevented that comment items are not extracted or comment items are redundantly extracted as in the past, and program specifications should be created using the extracted comment items as they are. Can be. As a result, the accuracy of extracting comment items is improved, and the efficiency of creating program specifications can be improved. Further, the separator is configured to specify a start line and an end line of various comments. You may.

 According to such a configuration, the start line and the end line of various comments are reliably identified by the separator. For this reason, the extraction accuracy of the comment range is improved, and the efficiency of creating the program specification can be further improved.

 Furthermore, the separator setting means may be configured to set at least a separator for identifying a comment on a file and a comment on a symbol.

 According to such a configuration, it is possible to reliably identify an adjacent comment at the head of the file. That is, in general, a comment on a file and a comment on a symbol are inserted adjacent to each other at the beginning of the file. Therefore, by setting at least a separator that identifies both, the comment range at the beginning of the file can be reliably extracted.

 In addition, the comment range extracting means extracts, for a comment for which a separator is not set by the separator setting means, a comment range based on a separator of another comment or an executable statement. The configuration may be such that Also, the comment range extracting means inserts between two comments when two comments are inserted adjacently at the beginning of a file and a separator is not set by the separator setting means. A configuration may be adopted in which the range of comments is extracted based on the executed statement.

 With such a configuration, it is possible to cope with various comment description methods. Further, the comment item extracting means may be configured to extract a comment item from which a comment identifier corresponding to a programming language in which a source program is described is removed.

According to such a configuration, an unnecessary comment identifier is removed when a program specification is generated, so that an easy-to-read program specification can be created. As a second means, a separator setting step for setting a separator for identifying various comments input to the source program, and a comment corresponding to the program specification for each separator set in the separator setting step A keyword setting step of setting an extraction keyword of the item; A comment range extracting step of extracting a range of each comment of the source program based on the W separator; and an extraction set in the keyword setting step for each comment range extracted in the comment range extracting step. A comment item extracting step of extracting a comment item from the source program based on the keyword; and a program specification generating step of generating a program specification based on the comment item extracted in the comment item extracting step. The feature is that a method for creating a program specification has been configured.

 According to this configuration, to efficiently create a program specification from the comments inserted in the source program, the following can be performed. That is, as a work when describing the source program, a separator for identifying the comment type is inserted into the source program. Then, when creating a program specification, if a separator is set in the separator setting step, the range of each comment is extracted based on the set separator. On the other hand, comment items are extracted from the source program based on the extraction key set in the keyword setting step for each range of the extracted comments. Then, a program specification is generated based on the extracted comment items.

At this time, even if comments of different types are adjacent to each other at the beginning of the file, each comment range can be accurately and reliably identified by the separator for identifying both. Therefore, it is prevented that comment items are not extracted or comment items are redundantly extracted as in the past, and program specifications should be created using the extracted comment items as they are. Can be. As a result, the accuracy of extracting comment items is improved, and the efficiency of creating a program specification can be improved. As a third means, a separator setting function for setting a separator for identifying various comments inserted in the source program, and extraction of comment items corresponding to the program specifications for each separator set by the separator setting function A keyword setting function for setting a keyword; a comment range extraction function for extracting a range of each comment of the source program based on the separator set by the separator setting function; and a comment range extraction function for extracting each comment range by the comment range extraction function. For each comment range, based on the extracted keywords set by the keyword setting function, —A program for realizing a comment item extracting function for extracting comment items from a source program, and a program specification generating function for generating a program specification based on the comment items extracted by the comment item extracting function. Preparation of specifications The program was recorded on a recording medium.

 Here, the “recording medium” refers to a medium that can reliably record various kinds of information and can reliably retrieve it as necessary. Magnetic tapes, magnetic disks, magnetic drums, IC cards, CD ROM, etc. Of portable recording media.

 According to this configuration, a program specification creation program for realizing the separator setting function, keyword setting function, comment range extraction function, comment item extraction function, and program specification generation function is recorded on the recording medium. . Therefore, if such a recording medium is provided, a general-purpose computer can be used to easily construct the program specification creating device according to the present invention. BRIEF DESCRIPTION OF THE FIGURES

 FIG. 1 is an overall configuration diagram of a specification creation device according to the present invention.

 FIG. 2 is a diagram illustrating a first example of a source program to be processed.

 FIG. 3 is a diagram illustrating a second example of a source program to be processed.

 FIG. 4 is a diagram illustrating a third example of a source program to be processed.

 FIG. 5 is a diagram illustrating a fourth example of a source program to be processed.

 FIG. 6 is a diagram showing an example of a generated program specification.

 Fig. 7 is a flowchart showing the flow of the specification creation procedure.

 FIG. 8 is a diagram showing an example of an extraction keyword setting screen.

 FIG. 9 is a diagram showing an example of an extraction keyword setting screen.

 Fig. 10 shows the set extracted keywords, (A) shows comment information on the file, and (B) shows comment information on the class.

 FIG. 11 is a flowchart of the main routine for extracting comments.

 FIG. 12 is a flowchart of a subroutine for extracting comments on a file.

Figure 13 is a flowchart of a subroutine that extracts comments on symbols. It is.

 FIG. 14 is a flowchart of a subroutine for obtaining a comment extraction range for a file.

 FIG. 15 is a flowchart of a subroutine for obtaining a comment extraction range for a symbol.

 FIG. 16 is a flowchart of a subroutine for extracting comment items necessary for the specification. BEST MODE FOR CARRYING OUT THE INVENTION

 The present invention will be described in more detail with reference to the accompanying drawings. 2 to 5 show an example of a source program to be processed by the program specification creator according to the present invention (hereinafter referred to as “specification creator”). These source programs are written in the C ++ language as the programming language, with comments beginning with “/ *” and ending with “* /”, and the part after “〃” on each line. The number at the beginning of each line in the source program indicates the line number of the program.

 The configuration of the source program shown in FIG. 2 will be described as follows. That is, the first to sixth lines are comments about the file, and the seventh to 16th lines are classes.

Comment on (symbol), line 17 and beyond is the class definition statement. The comment about the file and the comment about the class are provided with separators indicating the start and end of the comment, respectively. In comments about the file, “************” in the first and sixth comment lines is Seno. In comments related to classes, “” included in the comment lines on lines 7 and 16 functions as a separator.

The source program shown in Fig. 3 is an example in which no comment is provided for the class and no separator is provided. In the source program shown in Fig. 4, comments regarding files and classes are not provided with a separator, and an executable statement "#include <stdio.h>" that functions as a separator is inserted between the comments. It is an example. The source program shown in Figure 5 is an example without a comment on the class. Figure 1 shows the overall configuration of the specification creation device.

 The specification creator includes a display device 10, an input device 20, an external storage device 30, an arithmetic device 40, and a printing device 60. The display device 10 is configured by a display such as a CRT or an LCD, and provides a GUI (Graphics 1 User Interface) function to a user. The input device 20 is constituted by a pointing device such as a keyboard and a mouse, and provides a data input function to the GUI. The external storage device 30 is configured by a hard disk or the like, and provides a function of saving a source program. The arithmetic unit 40 is configured to include at least a central processing unit (CPU) and a memory, and controls the entire specification creating apparatus by a program loaded in the memory. The printing device 60 is configured by a printer or the like, and provides a printing function of the created program specification.

 In addition, the arithmetic unit 40 includes an extraction keyword setting unit 42, a comment range extraction unit 44, a comment item extraction unit 46, a specification generation unit 48, and a specification printing unit 50. Implemented programmatically.

 In the extraction keyword setting section 42, GUI is provided in cooperation with the display device 10 and the input device 20, and an extraction keyword corresponding to the type of comment is set. Here, "comment type" refers to an identifier such as a comment on a file or a comment on a class when the source program is described in the C ++ language. The extracted keyword setting unit 42 functions as a separator setting means, a separator setting step, a separator setting function, a keyword setting means, a keyword setting step, and a keyword setting function.

The comment range extraction unit 44 reads the source program file 32 from the external storage device 30, and extracts the comment extraction range over the entire file based on the set extraction keyword. Taking the source program shown in Fig. 2 as an example, lines 1 to 6 are extracted as comments for files, and lines 7 to 16 are extracted as comments for classes. It goes without saying that the range of extracted comments includes not only the beginning of the file but also comments after that. The comment range extracting section 44 functions as a comment range extracting means, a comment range extracting step, and a comment range extracting function. The comment item extraction section 46 extracts comment items included in each comment based on the set extraction keyword. Taking the comments on the source program file shown in Fig. 2 as an example, based on the extraction keywords "1. Overview", the keywords "Car class declaration" and "2. "Koji Sasaki" is extracted. The comment item extractor 46 functions as a comment item extracting means, a comment item extracting step, and a comment item extracting function.

 The specification generator 48 generates a program specification as shown in FIG. 6 based on the extracted comment items. The specification generation unit 48 operates as a program specification generation means, a program specification generation process, and a program specification generation function.

 In the specification printing section 50, the generated program specification data is transmitted to the printing device 60, and the printing device 60 prints the program specification.

 Next, the operation of the specification creating apparatus having the above configuration will be described.

 Figure 7 shows a flowchart of the procedure for creating a program specification.

 In step 1 (abbreviated as "S1" in the figure; the same applies hereinafter), extraction keywords for extracting comments from the source program are set. That is, when the “set extraction keyword” command is executed, the extraction keyword setting section 42 is selectively activated, and an extraction keyword setting screen as shown in FIGS. 8 and 9 (hereinafter referred to as “setting screen”). "100" is displayed. Then, on this setting screen 100, an extraction keyword is set for each type of comment. The extracted keywords include separator information indicating the start and end of the comment and specification information corresponding to the program specification.

As a procedure for setting the extracted keyword, first, the type of comment is selected from the type selection unit 102. Then, by appropriately operating the add button 104, the delete button 106, and the change button 108, the user inputs a separator and an extraction key word relating to a comment item. In addition, when setting an extraction keyword related to another comment type continuously, select another comment type from the type selection unit 102 and repeat the same operation. After completing the extraction keyword input, operate the OK button 110. By doing, the extracted keyword is determined. On the other hand, if you want to cancel the setting of the extracted keyword, press the cancel button 1 1 2 at any time to return the extracted keyword to the previous state.

 Note that FIG. 8 shows a state in which separator information and specification information “1. Outline” and “2. Creator” are set as extracted keywords related to file information (FIG. 9 is also the same). Then, as a result, an extraction keyword as shown in FIG. 10 is set.

 In step 2, comments are extracted from the source program for each type of comment. That is, when the “comment extraction” command is executed, the comment range extraction unit 44 and the comment item extraction unit 46 are selectively activated, and the comments are extracted according to the flowcharts of FIGS. 11 to 16 described later. You.

 In step 3, a program specification as shown in Fig. 6 is generated based on the extracted comments. That is, when the “create program specification” command is executed, the specification generator 48 is selectively activated, and a program specification is generated by its function.

 In step 4, the generated program specification is output from the printing device 60. That is, when the “print program specification” command is executed, the specification printing unit 50 is selectively activated, and the printing device 60 prints the program specification.

 In the above procedure, creation and printing of the program specifications can be arbitrarily selected.For example, if comments are extracted, printing of the program specifications may be performed automatically. ,.

 FIG. 11 shows a main routine of processing for extracting a comment from a source program.

In step 10, the source program file 32 is read from the external storage device 30, and the symbol information in the source program is extracted. Here, “symbol information” refers to line information of a symbol definition statement and symbol type information. Taking the source program shown in FIG. 2 as an example, the line definition is “line 17” and the type information is “class” because there is a class definition statement on line 17. The extracted symbol information is held in ascending order using, for example, row information as a key. In step 11, a subroutine is called to extract comments about the file. The extracted comments are stored in a table format using, for example, the comment type and the comment item for the specification as a key (the same applies hereinafter).

 In step 12, the next symbol information is read. That is, the first symbol information is read in the first process, and the second, third, and other symbol information is read in the subsequent processes.

 In step 13, a subroutine for extracting a comment on the symbol is called using the symbol information as a parameter.

 In step 14, it is determined whether or not the comment extraction for the symbol has been performed up to the end of the source program file 32, that is, whether or not the comment extraction has been completed. If the extraction of the comment is completed, the process is terminated (Y e s), and if the extraction of the comment is not completed, the process returns to step 12 (No). FIG. 12 shows a subroutine called in step 11, that is, a subroutine for extracting a comment on a file.

 In step 20, a subroutine is called to determine the extent of comment extraction for the file.

 In step 21, a subroutine for extracting a comment item corresponding to the program specification from the comment extraction range obtained in step 20 is called.

 FIG. 13 shows a subroutine called in step 13, that is, a subroutine for extracting a comment on a symbol.

 In step 30, a subroutine is called for finding the comment extraction range for the symbol.

 In step 31, a subroutine for extracting a comment item corresponding to the program specification from the comment extraction range obtained in step 30 is called. The subroutine called here is the same as the subroutine called in the previous step 21.

FIG. 14 shows a subroutine called in step 20, that is, a subroutine for obtaining a comment extraction range of a file. In step 40, the contents of the first line of the source program are extracted.

 In step 41, it is determined whether or not the content of the extracted first line is an executable statement. Here, “executable statement” refers to a statement other than a comment line. For example, if the source program is written in C ++ language, a statement to be processed by the preprocessor such as a “#define” statement, Definition statements, class definition statements, etc. correspond. If the content of the first line is an executable statement, it is determined that there is no comment on the file in the source program, and the process proceeds to step 47 (Yes). On the other hand, if the content of the first line is not an executable statement (that is, a comment line), the process proceeds to step 42 (No).

 In step 42, it is determined whether or not the content of the first line is a symbol separator. Here, the symbol separator is set on the extraction keyword setting screen 100 shown in FIG. 9 as a reference (the same applies hereinafter). If it is a symbol separator, it can be determined that the first line of the source program is a comment relating to the symbol. Therefore, it is determined that there is no comment relating to the file, and the process proceeds to step 47 (Yes). On the other hand, if it is not a symbol separator, it is determined to be the start line of the comment on the file, and the process proceeds to step 43 (No).

 In step 43, the contents of the next line (hereinafter referred to as "line contents") are extracted. In step 44, it is determined whether or not the line content is an executable statement. If the content of the line is an executable statement, the format is that there is no separator indicating the end of the comment on the file (see the source program in Fig. 4). Proceed to (Yes). On the other hand, if the content of the line is not an executable statement, it is determined that the comment on the file continues, and the process proceeds to step 45 (No).

 In step 45, it is determined whether or not the line content is a file separator. If it is a file separator, it is determined that the comment is the end line of the file, and the process proceeds to step 47 (Yes). On the other hand, if it is not a file separator, it is determined that comments regarding the file are continued, and the process proceeds to step 46 (No).

In step 46, it is determined whether or not the line content is a symbol separator. If it is a symbol separator, the comment on the file ends on the previous line. Then, the process proceeds to step 47 (Yes). On the other hand, if it is not a symbol separator, it is determined that comments about the file are continuing and

4 Return to 3 (No). Here, such determination is made, for example, in consideration of the case where there is no separator indicating the end of the comment on the file.

 In step 47, the extraction range of the comment on the file is calculated. If it is determined in step 41 or 42 that there is no comment on the file, information indicating that there is no comment line on the file, for example, `` 0 '' is set in the start line and end line .

 FIG. 15 shows a subroutine called in step 30, that is, a subroutine for obtaining a comment extraction range for a symbol.

 In step 50, the contents one line before the symbol definition line are extracted. Here, the symbol definition line is specified based on the contents of the parameter (symbol information) at the time of calling the subroutine for the symbol (see Fig. 13) in step 13. For example, in the case of the source program in Figure 2, the definition statement of the symbol (class) is on line 17, so the contents of line 16 are extracted.

 In step 51, it is determined whether the content one line before the symbol definition line is an executable statement. If the content of the line before the symbol definition statement is an executable statement, it is determined that there is no comment on the symbol, and the process proceeds to step 56 (Yes). On the other hand, if the content one line before the symbol definition statement is not an executable statement, the process proceeds to step 52 (No). As shown in Fig. 5, when the comment immediately before the symbol definition statement is a comment of a file, it is desirable to refer to the comment extraction range of the file in order not to extract the overlapping range.

 In step 52, it is determined whether the content of the previous line of the symbol definition statement is another separator, that is, a separator between a file and another symbol. If it is another separator, it judges that there is no comment about the symbol, and

Proceed to 5 6 (Y e s). On the other hand, if it is not another separator, it is determined to be the end line of the comment on the symbol, and the process proceeds to step 53 (No).

In step 53, the contents of the previous line are retrieved. That is, In this way, file analysis is performed to search for the start line of the comment on the symbol. In step 54, it is determined whether or not the line content is an executable statement. If the content of the line is an executable statement, the format is such that there is no separator indicating the start of a comment on the symbol (see the source program in Figure 3), and the comment on the symbol starts on the next line. Proceed to 6 (Yes). On the other hand, if the content of the line is not an executable statement, it is determined that the comment on the symbol is continued, and the process proceeds to step 55 (No).

 In step 55, it is determined whether or not the line content is another separator. If the content of the line is another separator, it is determined that the comment related to the symbol starts from the next line, with no separator (see the source program in Figure 3) indicating the start of the comment related to the symbol. Then, go to step 56 (Yes). On the other hand, if the line content is not another separator, it is determined that the comment on the symbol is continued, and the process returns to step 53 (No).

 In step 56, the extraction range of comments on the file is calculated. If it is determined in step 51 or 52 that there is no comment on the file, the same information as in step 47 is set.

 FIG. 16 shows a subroutine called in steps 21 and 31, that is, a subroutine for extracting a comment item necessary for a program specification.

 In step 60, the contents of the first comment line of the file or symbol are extracted. Here, the content to be extracted is a comment identifier indicating a comment line. For example, in the comment on the source program file shown in FIG. 2, “*” and “water /” are indicated by “* = N == M «M« I «M« M ”is output.

 In step 61, it is determined whether or not the content of the comment line is specification information. If it is the specification information, the process proceeds to step 62 (Yes), and if it is not the specification information, the process proceeds to step 64 (No).

 In step 62, the element number, pointer, etc. of the array storing the comment items according to the specification information are changed.

In step 63, it is determined whether or not the next line is the extraction range of the comment item. If it is within the comment item extraction range, the process returns to step 60 (Y es), If it is not within the comment item extraction range, the process ends (No).

 In step 64, it is determined whether or not the storage location of the comment item has been set. That is, it is determined whether or not the storage destination has been changed (set) in the process of step 62. To make such a determination, if the storage destination has not been set, “NULL” may be set as the storage destination. If the storage destination has been set, the process proceeds to step 65 (Yes), and if the storage destination has not been set, the process proceeds to step 63 (No). Here, an example of a case where the storage destination is not set is a case where the first line of the comment line is being processed and the line is a separator.

 In step 65, the comment item is stored in a predetermined storage location.

 According to the processing described above, in order to efficiently create a program specification from comments entered in a source program, a separator for identifying a comment type is inserted when describing the source program. Then, when creating the program specification, if the inserted separator information is set, the range of each comment is extracted based on the separator information. At this time, even if the comment about the file and the comment about the symbol are adjacent to each other at the beginning of the file, the range is accurately and reliably identified by the separator that identifies the two. Then, within the extracted range, each comment item corresponding to each comment is extracted, and a program specification is generated.

 Therefore, confusion of comment types at the beginning of the file can be prevented, and a program specification can be created using the extracted comment items as they are. For this reason, the efficiency of creating program specifications is greatly improved, and, for example, the program development period can be shortened, and development costs can be reduced.

 In the above-described specification creator, the source program described in the C ++ language is to be processed as a programming language. However, other programming languages, for example, BAS IC, COBOL, FORTRAN, Java, etc. May be targeted for processing.

If a program for realizing such a function is recorded on a recording medium such as a magnetic tape, a magnetic disk, a magnetic drum, an IC card, a CD-ROM, etc., the program specification creating program according to the present invention can be marketed Can be distributed. Soshi Thus, a person who has obtained such a recording medium can easily construct the program specification creator according to the present invention using, for example, a general computer. Industrial applicability

 As described above, the program specification creating apparatus and the program specification creating method according to the present invention are extremely useful because the accuracy of extracting comment items is improved. Further, the recording medium on which the program specification creating program according to the present invention is recorded is extremely useful because a program specification creating device can be easily constructed using a general computer.

Claims

The scope of the claims 1-separator setting means for setting a separator for identifying various comments inserted in the source program,  A keyword setting means for setting a keyword for extracting comment items corresponding to the program specifications for each of the separators set by the separator setting means; and a comment for each source program based on the separator set by the separator setting means. Comment range extracting means for extracting a range of  Comment item extracting means for extracting a comment item from a source program based on the extracted keyword set by the key word setting means for each comment range extracted by the comment range extracting means;  Program specification generating means for generating a program specification based on the comment item extracted by the comment item extracting means;  A program specification creation device characterized by comprising:  2. The program specification creating device according to claim 1, wherein the separator is configured to specify a start line and an end line of various comments.  3. The program specification creating device according to claim 1, wherein the separator setting means is configured to set at least a separator for identifying a comment on a file and a comment on a symbol.  4. The comment range extracting means is configured to, for a comment for which a separator is not set by the separator setting means, extract a comment range based on a separator of another comment or an executable statement. The program specification creation device described in paragraph 1.  5.The comment range extracting means inserts two comments adjacent to each other at the beginning of the file, and inserts between the two comments when no separator is set by the separator setting means. 3. The program specification creating device according to claim 1, wherein the device is configured to extract a comment range based on an executable statement. 6. The comment item extracting means is a programming language that describes a source program. 2. The program specification creating device according to claim 1, wherein the comment item is obtained by removing a comment identifier corresponding to the comment item.  7. A separator setting step of setting a separator for identifying various comments entered in the source program;  For each separator set in the separator setting step, a key code setting step for setting an extraction keyword of a comment item corresponding to the program specification, and a source program based on the separator set in the separator setting step. A comment range extracting step of extracting a range of each comment;  A comment item extracting step of extracting a comment item from a source program based on the extracted keyword set in the key word setting step for each comment range extracted in the comment range extracting step;  A program specification generation step of generating a program specification based on the comment item extracted in the comment item extraction step;  A method for preparing a program specification, comprising:  8. Separator setting function to set a separator to identify various comments inserted in the source program,  For each separator set by the separator setting function, a keyword setting function for setting an extraction keyword of a comment item corresponding to the program specification, and a source program of the source program based on the separator set by the separator setting function. A comment range extraction function for extracting the range of each comment,  A comment item extraction function for extracting a comment item from a source program based on the extraction keyword set by the keypad setting function for each comment range extracted by the comment range extraction function;  A program specification generation function for generating a program specification based on the comment item extracted by the comment item extraction function;  Recording medium storing a program specification creation program for realizing the above.