JP2915678B2 - Module structure creation device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an apparatus for creating a module structure used in software development, and more particularly, to an apparatus for automatically creating and outputting a module structure diagram from a data flow diagram. The present invention relates to an apparatus for creating a modular structure.

[0002]

2. Description of the Related Art In software development, a data flow diagram is provided and converted into a module structure diagram. As a conversion method (design method) of the module structure at that time, in a conventionally proposed structured design method, a module structure is designed based on a required specification defined using a given data flow diagram.

In such a conventional structured design method, first, a process included in the data flow diagram is classified into input, conversion, and output processes by focusing on a change in the contents of input / output data in the data flow diagram. I do. Next, the conversion process is defined as a parent module, and a process adjacent to the conversion process (parent module) in the data flow is defined as a child module.

When this child module is adjacent to another process, a hierarchical module structure diagram is created (converted) stepwise by using the child module as a parent and the adjacent process as its child module.

[0005]

There have been proposed CASE tools that support the above structured design techniques. However, according to these tools, the height of the generated module structure diagram is limited by the data flow. Since the length is proportional to the length of the path, it tends to have a deep structure. On the other hand, a module structure manually created from a data flow diagram often has a flat structure.

Therefore, there is a problem that the module structure generated by these CASE tools is inconvenient because it is separated from the module structure manually created. Further, when a plurality of processes have a complicated input / output relationship, the module structure generated from the data flow diagram is no longer a tree structure, has a complicated calling relationship, and is disadvantageous.

Further, since the process type is estimated on the basis of the meaning of the input data name and the output data name of the process, a large-scale thesaurus is required, and there is a problem that work efficiency is poor and not practical. .

The present invention solves the above-mentioned drawbacks of the prior art, efficiently estimates the process type of each process in a given data flow diagram, and automatically creates a module structure diagram as natural as possible. It is an object of the present invention to provide a module structure creating apparatus capable of performing the following.

[0009]

According to the present invention, in order to achieve the above object, a module structure creating apparatus includes a data store, an external entity, and a process as nodes, and a data flow connecting the nodes with a data flow as a parent. A module structure creating apparatus that converts and outputs a module structure diagram including modules and child modules, wherein the data store and the external entity
A first mark to the node of the tee, and the first marked non node exists whether processes category estimation completion determining process, and all and all input sources is the first marked A process whose output destination is not the first marked process is an input process, a process whose output destinations are the first marked process and all input sources are not the first marked process are an output process, Performs an estimation process for determining a process type in which the first marked and all output destinations have the first marked process as a conversion process, and is connected to the first marked node by a data flow. Performing a process of making the first mark on a node of a first non-marked process, until no process has not been subjected to the first mark process; Serial process category estimation completion determination processing,
Estimation processing for the process type determination and the first
A process type estimating unit for performing a process of repeating a process of marking a process type, a process type confirmation panel unit for visually presenting the estimated and determined process type in a table format and enabling interactive correction, and a process type. And data flow combination
Matching pattern 1 consists of the first process and the second process.
Data flow, and input the first process type
In the process, if the second process type is conversion process or output
A pattern that is a force process, wherein
Pattern 2 converts the third process into a fourth process.
Connected by tap flow, the third process type is conversion process
And the fourth process is the conversion process, the input process or
Is a pattern that is an output process,
Set 3 is a data store for the fifth process.
Or a putter that connects to an external entity via a data flow
And the combination of process type and data flow
Predetermined for each pattern consisting of
The module structure of creating rules, the parent of the new process
Module creation rule 1, new process parent module
Create rule 2 and output data
Including the creation rule 3 as a module, the creation rule 1 includes:
Pattern consisting of a combination of process type and data flow
If the pattern matches the pattern 1, the first process
Create a sixth process superior to the second process,
Process as a parent module, the first process and the second process
Process as a child module.
Rule 2 is a combination of process type and data flow
If the pattern consisting of
Process and the seventh process that is higher than the fourth process
Create the seventh process as a parent module and create a third process.
The process and the fourth process are child modules.
The creation rule 3 is based on the process type and the data flow.
Pattern consisting of a combination of
Then, the fifth process is the parent module, and the data is
The feature is that output is made a child module.
A module structure creation rule management unit which manages the serial creation rule, to pattern consisting of a combination of the process type and data flow after the implementation of the interactive modifications in the process type confirmation panel portion, of the data flow diagram A second mark is applied to the data store and the node of the external entity , and a fifth process connected to the second marked node by a data flow is set as a conversion candidate to the module structure. Perform module structure conversion completion determination processing of the existence or not, and perform module structure conversion processing of applying the module structure creation rules to the pattern of the conversion candidate to convert to the module structure, from the conversion candidate A module structure conversion completion process for deleting the fifth process is performed, and the second marked node is processed. And performing a conversion candidate addition process with the process connected by the data flow as a conversion candidate, and until there is no more unprocessed process, the module structure conversion completion determination process, the module structure conversion process, the module structure conversion completion process, and And a module structure generating unit that converts the data flow diagram into a module structure diagram and outputs the result by repeating the conversion candidate addition process.

[0010]

The module structure creation device is provided with the data flow diagram, converts the data flow diagram into the module structure diagram, and outputs it. At this time, the process type estimating unit performs the data store and external error processing in the given data flow diagram.
Make a first mark on the node of the entity
The process type estimation completion determination process is performed to determine whether or not there is a node that has not been marked.
The processes which are marked and all the output destinations are the first non-marked processes are set as input processes, and all the output destinations output the processes which are the first marked and all the input sources are not the first marked processes. And performing an estimation process for determining a process type in which all input sources are the first marked process and all output destinations are the first marked process as a conversion process.
Performing the first mark on the nodes of the first non-marked process connected in a data flow with the marked nodes of the first and second processes until the processes without the first mark processing are exhausted. Estimating and determining the type of the process by estimating and determining the type of the process by performing a process of repeating a type estimation completion determination process, an estimation process for the process type determination, and a process of making the first mark. . The process type confirmation panel unit visually presents the estimated and determined process type in a table format to enable interactive correction. if,
If necessary, the designer can adjust the process type on the confirmation panel.

Combination of process type and data flow
Pattern 1 of the first example is to convert the first process into the second process.
Data flow, and the first process type is the input process.
And the second process type is conversion process or output process.
A pattern that is a set,
2 transfers the third process to the fourth process.
Connection, and the third process type is a conversion process,
Process 4 is conversion process, input process or output
A pattern that is a process,
Turn 3 describes the fifth process as a data store or
Is a pattern that connects to the
Is composed of a combination of process type and data flow.
For each pattern,
The rules for creating a Joule structure allow a new process to be
Rule 1, new process parent module
Output the creation rule 2 and data
Includes a creation rule 3 which is a process rule.
Pattern consisting of a combination of
If the pattern 1 matches, the first process and the second process
Create a sixth process above the process and create a sixth process.
Process as a parent module, the first process and the second process
A child module of the process.
2 is a combination of process type and data flow.
If the pattern matches the pattern 2, the third
Process and create a seventh process above the fourth process
And the seventh process as a parent module and the third process
And a fourth module as a child module
However, the creation rule 3 is based on the process type and the data flow.
The pattern consisting of the combination matches the pattern 3
If the fifth process is the parent module,
It is characterized in that the
A rule structure creation rule management unit prepares and manages the creation rules .

[0012] The module structure generation unit is configured to, for the pattern composed of the combination of the process type and the data flow after the interactive correction in the process type confirmation panel unit, perform the data storage of the data flow diagram. And making a second mark on the node of the external entity , making a fifth process connected to the second marked node by a data flow a conversion candidate to the module structure, and whether the conversion candidate exists A module structure conversion completion determination process of determining whether or not a module structure is to be converted to the module structure by applying the module structure creation rule to the pattern of the conversion candidate; A module structure conversion completion process for deleting a process is performed, and the second marked node and the The process connected to the process flow is performed as a conversion candidate, and the module structure conversion completion determination process, the module structure conversion process, the module structure conversion completion process, and the conversion are performed until there is no process that has not been converted yet. By repeating the candidate addition process, the data flow diagram is converted into the module structure diagram and output, whereby the data flow diagram is converted into the module structure diagram and output. This makes it possible to automatically and easily create the module structure.

[0013]

Next, an embodiment of the present invention will be described with reference to the drawings. FIG. 1 is a block diagram showing a system configuration according to an embodiment of the present invention. In the figure, 1 is a process type estimation unit, 2 is a confirmation panel unit, 3 is a module structure generation unit, 4
Denotes a module structure creation rule management unit, 5 denotes a module structure diagram editor, and 6 denotes a data flow diagram editor.

The module structure creating apparatus according to the present invention comprises the above-mentioned elements 1 to 4, and the module structure diagram editor 5 and the data flow diagram editor 6 are accessories. That is, given a data flow diagram,
On the other hand, if necessary, a human (designer) makes a correction (edit) using the data flow diagram editor 6, and then gives it to the process type estimating unit 1.

The process type estimating section 1 estimates and determines the type of a process based on the adjacency between processes related to a data flow in a given data flow diagram. The confirmation panel unit 2 visually presents the estimated and determined process type in the form of a table to enable interactive modification with humans. If necessary, the designer can adjust the process type on the confirmation panel 2.

The module structure creation rule management section 4 prepares and manages a module structure creation rule that is predetermined in correspondence with each pattern composed of a combination of a process type and a data flow.

The module structure generation unit 3 responds to the combination pattern of the process type and the data flow after the interactive modification in the confirmation panel unit 2 by the correspondence prepared in the module structure creation rule management unit 4. By applying the rules described above, the combination pattern is converted into a module structure, and as a result, a given data flow diagram is converted into a module structure diagram and output.

A human (designer) can use the module structure diagram editor 5 to modify (edit) the module structure diagram output from the module structure generation unit 3 as necessary.

FIG. 2 is an explanatory diagram showing an example of a data flow diagram. In the figure, p1 to p7 are processes, d1 to d9 are data (data flows), respectively.
e1 and e2 are external entities (for example, operation
, Users, other systems, communication channels, etc.), f
1 is a data store (or sometimes a file),
It is. In the figures and in the text that follows, outside refers to outside
Means a department entity.

FIG. 2 is a data flow diagram,
The algorithm is also shown. That is, steps 1, 2, and 3 are algorithm steps. The outline will be described below. In step 1, when the algorithm is processed, first, for the processes P1, P6, and P5, a process type such as an input process, an output process, and a conversion process is determined. Become. In other words, the first step is to determine the process type.
This is the purpose of the algorithm.

In the method described in detail later, in step 1, the process P1 is determined as an input process, the process P6 is determined as an input process, and the process P5 is determined as an output process. Briefly, when the direction of the arrow goes from the outside to the inside, the input is made, and the opposite case is made the output.

At the stage 2, (data flow) d2 is observed, and since it is directed from outside to inside, the process P2
Looks at the input process, similarly (data flow) d8, and since this is going from outside to inside, process P7
Is an input process, and (data flow) d5 is determined to be an output process because the process P4 is directed from inside to outside.

In step 3, for the process P3, (data flow) d3 enters and (data flow) d4
Comes out, and d3 and d4 intersect, so
The process P3 is determined as a conversion process. With such a concept, a process of labeling each process P as an input process, an output process, and a conversion process is performed.

FIG. 3 is an explanatory diagram showing a module structure diagram created according to the present invention from the data flow diagram shown in FIG. In FIG. 3, P and P1 to P7 are processes, d1 to d9 are data (data flows), respectively.
It is.

In FIG. 2, a process of labeling each process P as an input process, an output process, and a conversion process is performed, and FIG. 3 is referred to. In step 1 of FIG. 3, since the type of the process P1 is an input process, a child module “input d1” is created, and a module corresponding to the process P1 is referred to as “P1”.
1 ”.

Similarly, P6 is made. Since P5 is an output process, d6 is created so as to output data from top to bottom, and a child module "output d6" is created.

In step 2 of FIG. 3, the relationship between the module P2 and the module P1 is created in the manner just described. So is the relationship between P7 and P6,
The same goes for the relationship between 4 and P5.

Next, in step 3 of FIG. 3, since P3 is a conversion process, a parent module P (there is always a central module, so that it is automatically generated without permission) is provided and directly below it. Put P3. P2, which directly exchanges data with P3, is also placed at the same level as P3.

Similarly, P4 is also placed at the same level, and P7
Also put on the same level. Since P4 and P7 exchange data directly, if P4 is at the same level, P7 is also placed at the same level, so that the desired module structure (tree structure) is completed.

The outline of the processing of the process type estimation unit 1 and the processing of the module structure generation unit 3 in FIG. 1 has been described above with reference to FIGS. The process of the confirmation panel unit 2 is merely a process of visually presenting the process type estimated and determined by the process type estimating unit 1 in the form of a table so as to enable interactive modification with humans. No need to explain.

FIG. 4 is a processing flowchart showing the processing of the process type estimating section 1 in FIG. 1 in detail. This will be described below with reference to FIG. That is, the process type estimating unit 1 performs the following processing based on the data flow extracted from the data flow diagram of FIG. 2, the process, the outside, and the data store according to the processing flow shown in FIG. Estimate the process type of the process in the data flow diagram.

[Step 0] e1, e2 and f1 are marked. [Step 1] The output destination of e1 is P1 and the output destination of P1 is P2
Is not already marked, the process type of P1 is set to “input”. The output destination of f1 is P6, and the output destination of P6 is P
Since 7 is not already marked, the process type of P6 is set to “input”. The output destination of P5 is f1 and the input source P of P5
Since No. 4 has not been marked, the process type of P5 is set to “output”. Mark P1, P5, P6.

[Step 2] The output destination of P1 is P2 and P2
Since the output destination P3 has not been marked, the process type of P2 is set to “input”. The output destination of P6 is P7 and P7
Since the output destination P4 has not been marked, the process type of P7 is set to “input”. The input source of P5 is P4 and P4
Is not marked, the process type of P4 is set to "output". Mark P2, P4 and P7.

[Step 3] The output destination of P2 is P3 and P3
Since the output destination P4 has already been marked, the process type of P3 is set to "conversion". [Step 4] The process ends because all nodes have been searched.

FIG. 5 is a processing flow chart showing the processing of the module structure generator 3 in FIG. 1 in detail again. This will be described below with reference to FIG. That is, in the module structure diagram creation unit 3, the process type and the data flow diagram (FIG. 2) confirmed by the confirmation panel unit 2
According to the processing flow shown in FIG. 5 on the basis of the data flow and the process extracted from the above, based on the module structure creation rule (FIG. 6) managed by the external, data store and module structure creation rule management unit 4, FIG.
The module structure diagram shown in FIG. 3 can be created from the data flow diagram of FIG.

Before that, the rules for creating the module structure shown in FIG. 6 will be briefly described. In FIG. 6, there are a rule column, a data flow column, and a module structure column. This means that if there is a pattern shown in the column of data flow, it is a rule to create a module shown in the column of module structure.

In the data flow column corresponding to the rule 1, the row a) is shown in the module structure column when the type of the process P1 is input and the type of the process P2 is conversion or output. Indicates that a module is to be created. Similarly, the row b) indicates that when the type of the process P1 is output and the type of the process P2 is input or conversion, a module shown in the module structure column is created.

Similarly, in line d), the type of the process P1 is input, the type of the process P2 is also input, and
If the hierarchy level of 1, P2 is 1, it indicates that the module shown in the column of the module structure is created. Hereinafter, the same applies to rules 2 to 5.

After the description of the module structure creation rules has been completed, it will be described that the module structure diagram shown in FIG. 3 can be created from the data flow diagram of FIG. 2 in the following manner in accordance with the processing flow shown in FIG.

[Step 1] P1, P5 and P6 are conversion candidates. Mark P1 to remove it from the conversion candidates and apply Rule 4 to the set <e1, P1>. Since the output destination P2 of P1 is not marked, P2 is added as a conversion candidate.

Mark P6 to remove it from the conversion candidates and apply Rule 4 to the set <f1, P6>. Since the output destination P7 of P6 is not marked, P7 is added as a conversion candidate.
Mark P5 to exclude from conversion candidates and set rule 5 as <P
5, e2>. Since the input source P4 of P5 is not marked, P4 is added as a conversion candidate.

[Step 2] P2, P7 and P4 are conversion candidates. Since the input source P1 of P2 is marked, P2
To be excluded from conversion candidates, and set rule 2 as a set <P1, P
2> Since the output destination P3 of P2 is not marked, P3 is added as a conversion candidate.

Since the input source P6 of P7 is marked, P7 is marked and excluded from conversion candidates, and Rule 2 is set as <
P6, P7>. Since the output destination P4 of P7 is not marked, P4 is added as a conversion candidate. Since the output destination P5 of P4 is marked, P4 is marked and excluded from conversion candidates, and Rule 3 is applied to the set <P4, P5>.

At this time, since there are a plurality of input sources of P4,
The level of P4 is set to 1. That is, P4 is a junction of the outputs. Since the input source P7 of P4 is marked, c) of Rule 1 is applied to the set <P7, P4>. However, since another input source P3 of P4 is not marked, P3 is added as a conversion candidate.

[Step 3] The conversion process P3 is a conversion candidate. Mark P3 and remove from conversion candidates. Since the input source P2 of P3 is marked, set a) of rule 1
2, P3>. Also, since the output destination P4 of P3 is marked, c) of Rule 1 is applied to the set <P3, P4>.

[Step 4] The process ends because there are no more conversion candidates.

[0047]

As described above, according to the present invention,
Since a natural module structure can be automatically generated from a data flow diagram, there is an advantage that the work efficiency of module structure creation can be greatly improved as compared with the related art. Further, there is an effect that the design method of the module structure can be unified, and standardization of the design can be facilitated.

[Brief description of the drawings]

FIG. 1 is a block diagram illustrating a system configuration according to an embodiment of the present invention.

FIG. 2 is an explanatory diagram showing an example of a data flow diagram.

FIG. 3 is an explanatory diagram showing a module structure diagram created according to the present invention from the data flow diagram shown in FIG. 2;

FIG. 4 is a processing flow chart showing a process of a process type estimating unit 1 in FIG. 1;

FIG. 5 is a processing flowchart showing the processing of the module structure generation unit 3 in FIG. 1;

FIG. 6 is an explanatory diagram showing a module structure creation rule.

[Explanation of symbols]

DESCRIPTION OF SYMBOLS 1 ... Process type estimation part, 2 ... Confirmation panel part, 3 ... Module structure generation part, 4 ... Module structure creation rule management part, 5 ... Module structure diagram editor, 6 ... Data flow diagram editor.

──────────────────────────────────────────────────続 き Continuation of the front page (56) References "IPSJ Research Report" Vol. 85 (SE-81) p. 9-16 (1991.10.4) "Information Processing Society of Japan" Report, Vol. 66 (SE-80) p. 43-501 (July 18, 1991) (58) Field surveyed (Int. Cl. ⁶ , DB name) G06F 9/06

Claims (1)

(57) [Claims] 1. A data flow diagram in which a data store, an external entity, and a process are used as nodes and the nodes are connected by a data flow are converted into a module structure diagram including a parent module and a child module. In the module structure creating apparatus, the first node is marked on the data store and the external entity , and a process type estimation completion determination process is performed as to whether or not the first unmarked node exists. A process in which all input sources are the first marked and all output destinations are not the first marked, and all output destinations are the first marked and all input sources are the first marked. Let the first unmarked process be the output process, all sources being said first marked and all outputs The first non-marked process node connected to the first marked node in a data flow with a presumed process for determining a process type using the first marked process as a conversion process The first mark processing is performed, and the process type estimation completion determination processing, the estimation processing for the process type determination, and the first mark are performed until there is no process in which the first mark processing has not been performed. A process type estimating unit that repeats the process of performing the process, a process type confirmation panel unit that visually presents the estimated and determined process type in a table format and enables interactive correction, a process type and data Flow combination pattern 1
Converts the first process into a second process with a data flow
Connect, the first process type is input process, the second
If the process type is conversion process or output process,
Turn, wherein pattern 2 of the combination describes the third process as a fourth process
Connected to the process with the data flow, the third process type
Another is the conversion process, the fourth is the conversion process,
If the pattern is an input or output process
Te, pattern 3 of the combination, the fifth process
Connect to datastore or external entity with dataflow
Pattern that consists of a combination of process type and data flow.
Each over emissions, modules predetermined in corresponds respectively
The rules for creating a file structure specify that a new process
Creation rule 1, new process parent module
Outputting rule 2 and data is a child module.
The creation rule 1 includes a combination of a process type and a data flow.
If the pattern consisting of matching matches the pattern 1
For example, the sixth process above the first process and the second process
Process and create a sixth process as a parent module,
Child module of the first process and the second process
The creation rule 2 is a combination of a process type and a data flow.
If the pattern consisting of matching matches the pattern 2
For example, the seventh process above the third process and the fourth process
Process, the seventh process is the parent module,
Child module of the third process and the fourth process
The creation rule 3 is a combination of a process type and a data flow.
If the pattern consisting of matching matches the pattern 3
If the fifth process is the parent module, output the data
It was characterized to be a child modules to a module structure creation rule management unit that manages the creation rule, the process type and data flow after the implementation of the interactive modifications in the process type confirmation panel portion For the pattern of combinations, mark the data store of the data flow diagram and the node of the external entity with a second mark, and execute a fifth process connected with the second marked node by a data flow. A module structure conversion candidate is determined, a module structure conversion completion determination process is performed as to whether or not the conversion candidate exists, and the module structure is converted to the module structure by applying the module structure creation rule to the pattern of the conversion candidate. Module for performing module structure conversion processing and deleting the fifth process from the conversion candidates Performing a structure conversion completion process, performing a conversion candidate addition process using a process connected to the second marked node by a data flow as a conversion candidate, and performing the module structure conversion completion determination process until there is no more unprocessed process. A module structure generation unit that converts a data flow diagram into a module structure diagram and outputs the result by repeating the module structure conversion process, the module structure conversion completion process, and the conversion candidate addition process. Characterized module structure creation device