JPS6068435A - How to search for software parts - Google Patents

Abstract

(57) [Summary] This bulletin contains application data before electronic filing, so abstract data is not recorded.

Description

[Detailed Description of the Invention] [Field of Application of the Invention] The present invention relates to software design and development. In particular, the present invention relates to a search method such as a similar search for software components based on knowledge, which is suitable for storing standard software components in a component database and creating component specifications and source programs that meet required system specifications.

[Background of the invention]

Conventional software component search methods used information retrieval systems that did not require knowledge, so it was not easy to establish a complete keyword system for searching, accumulate components, and then change the keyword system. In fact, if the desired search results are not obtained due to insufficient information required for a search, users do not know what information to add. Another drawback is that no information can be obtained if there are no parts that meet the required specifications.

[Purpose of the invention]

The purpose of the present invention is to utilize the entire team of software design experts and others in order to search for software components that meet required specifications from a component database that stores software components for software design and development. Therefore, it is an object of the present invention to provide a method that can infer a similar software component as the next best solution and search it from a component database as efficiently as possible even if a matching software component is not found.

[Summary of the invention]

In order to achieve these objectives, the present invention automatically searches for software components, which is performed manually by software design experts, when designing and developing software using software components.
This method stores the search know-how of experts and others in a knowledge base, and uses this to enable accurate searches by non-experts. The knowledge for component searching stored in the knowledge base is written using all production rules to describe what kind of external environment, functions, characteristics, etc., are useful for processing when creating software components.

In addition, with this method, the number of parts that can be stored in the parts database is limited, so it is likely that there will be no parts that completely match the required specifications. Money is also created, and although it is within a limited range, it also provides a system game whose parts can be used.

[Embodiments of the invention]

Hereinafter, one embodiment of the present invention will be explained with reference to FIGS. 2, 3, and 4. First, the overall configuration of this embodiment will be described in the first section.
This is explained in the figure. The knowledge base 2 stores knowledge for searching software component specifications. The knowledge base 2 is composed of a standard parts search knowledge source 3 and a similar standard parts search knowledge source 4. 3 and 4 are the know-how learned from software design experts.
All knowledge expressions in the form of production rules were input into a computer. 3 and 4 are constructed in the knowledge base 2 of the computer as knowledge sources (groups of knowledge expressed in the form of production rules). Each knowledge source is
It is created with the following structure (represented by 8 expressions in LISP language).

((Knowledge source startup preconditions) (Knowledge source common information) ((Production rule 1) (Rule 1 (IF (Prerequisite clause) (Prerequisite clause)) (THEN (PROPO8E (Action) (SEL
ECT-INF Prerequisite] (PROPO8E...) (Production Rule n))) The knowledge source consists of knowledge source startup preconditions, knowledge source common information, and production rule group.

The blackboard 5 is a storage device that stores an in-progress state. The inference control information area 6 is used to decide which knowledge source to interbrid with next.

The terminal device 7 is used to input the required specifications of the system to be created, to receive a temporary response from the inference, to display and modify dialogue of the component specifications retrieved from the component database 8, and to modify the source code of the retrieved component. used for The component database 8 is a database that can be searched by software component name as a set of software components, component specifications, and component source codes.

The software components handled in this example consist of a set of component specifications and source code, and parts that can be commonly used when creating standard components are written in source code format, and processing is task-specific. The parts are made of perforated blocks so that standard parts users or users can create their own units.

An example of a software component is shown in FIG.

Next, the contents of knowledge base 2 will be explained.

Standard parts search knowledge source 3 is knowledge used for ρ inference in order to search for standard parts that meet the required specifications.
The environment in which the parts are used and the external functions are defined as knowledge that is associated with the part names in the parts database 8, and is expressed and stored in the form of a knowledge source. As a result of the inference using 3, the name of a standard part that can be used (can only be used in own coding of the holed part) is output. Similar standard parts search knowledge source 4 is a standard part that can be used by modifying a hole when a standard part that meets the required specifications cannot be found. The knowledge for inferring the above is created by reducing the number of prerequisite phrases in the IP part of the production rule in the knowledge source in step 3, or by spelling out all the conditions, and is stored in the form of a knowledge source.

Next, the operation of this embodiment will be explained with reference to FIGS. 3 and 4. The basic control of inference, which is the core of the operation, is shown in FIG. First, the required specifications inputted by the user-defined initial processing function shown in the figure are written into the blackboard 5.15゜Is this user-defined initial processing function activated for the first time at the end? Specify the knowledge source to apply 15° In this example, standard parts search knowledge source 3
Configure all settings. Next, 16°1 enters interpreter processing.
The IP part of the production rule in the knowledge source specified in 5 or 18 is matched with the event written on the blackboard 5, and the T'HEN part of the rule where the matching is established is executed.

17 inference termination conditions (search for available standard parts,
If a similar standard part is searched, or by inference, it is written in a predetermined node on the blackboard that the standard part is not found in the search, it is fully determined, and the interbrid process is repeated for 19 days until it is established.

The selection 18 of the knowledge source to be interbrid is determined by using the control information in the production rules applied up to the relevant time point stacked in the inference control area 9.

Next, the operation of searching and utilizing the specifications of necessary software components from the component database based on the component specifications that can be requested by inference will be explained with reference to FIG. First, the reasoning for searching for standard parts will be explained. The result of inputting the required specifications from the input device 7 is written on the blackboard 5, and inference is started. For example, in the case of a component having a matching function, the states of the blackboard 5 at the start of inference are: Matching Object name arrangement Yes Attribute name Value input Intermediate file 2 Output Intermediate file 3 Reference mask Supplier mask. On the other hand, the component name is inferred using the standard component search knowledge source 3, and the component specification and source program are searched from the component database 8.

An example of the knowledge at this time is (R20-10 (IF (with matching arrangement) (without matching parts) (number of breaks in matching file matching 11)) ('lI'HEN (PROPO8E (matching part MATCHI) (S
ELECT-INF Standard parts search knowledge)]]) [Explanation] Rule number R, 20-10 If there is a matching arrangement.

If the collation parts are not determined and the number of breaks in file matching for collation is J:1, then (if there is no file matching break count attribute for collation object on the blackboard, the input/output device will The matching component is MATCHI.

As information for selecting the next knowledge source to launch,
Standard parts search knowledge sounds are used.

It is. If standard parts can be searched in this way, the searched part specifications and source programs may be partially modified interactively by a human to meet the requested part specifications, if necessary. Complete the book and source program.

If a standard part cannot be found by inference, use all similar standard part search knowledge 4 to find a standard part that will serve as a starting point for creating a part that meets the required specifications by inference, and search from the parts database 8. do. An example of the knowledge at this time is (R, 3O-10 (IF (verification placement) (verification part not present) (verification file matching break number present)) (THEN + PROPO8E (verification similar standard part MATCHI)
) (5ELECT -INF similar standard parts search knowledge)))) (R30-11 (IF' (verification placement available) (verification component absent) (verification file matching break number included)) (THEN (PROPO8E (verification similar standard) Parts MAT CH
N) (SELECT-INF Similar standard parts search knowledge)))) [Explanation] Rule number R30-10 If there is a collation arrangement, the collation parts have not been determined, and there is a break number for file matching of the collation. Then, if there is no file matching break for (matching object on the blackboard), move from the input/output device to the user base) A similar standard component for matching is MATCHI.

As information for selecting the next knowledge source to launch,
Use similar standard parts search knowledge.

Rule number R30-11 If there is a collation arrangement, the parts of the collation have not been determined, and there is a break number for file matching of the collation, (the file matching break number attribute of the collation object on the blackboard is (If not, it will be transferred from the input/output device to the user.) A similar standard component for verification is MATCHN.

Similar standard parts search knowledge is used as information for selecting the knowledge source to be activated next.

It is. Rule number R, 30-10 is a loosened condition of the third prerequisite phrase of the IF portion of rule number R20-10 in the standard parts search knowledge source described above. Similar standard parts MATCHNI'i with rule number 'R30-11,
Originally, it is a standard part when the number of breaks in file matching is 1:N, but it can also be used as a starting point when the number of breaks in file matching for the requested part specification is 1:2. It shows.

If similar standard parts can be searched for in this way, the system user can modify the searched standard part specifications and source programs to complete the desired parts specifications and source programs (holes in standard parts). It is also necessary to modify areas other than empty areas). If a similar standard part cannot be retrieved, a message indicating that inference could not be made is displayed on the input/output device 7.

According to this embodiment, when a standard part completely conforms to the required specifications, it is possible to not only create correct software by only modifying the holed part as intended when creating the standard part. Even if there are only similar parts depending on the suitability of the search, it may be necessary to modify parts other than the holes, but since it can be used as a starting point, the accumulated parts can be used effectively.

〔Effect of the invention〕

According to the present invention, in software design e-development using software components, even non-experts in software design can use the entire computer by storing knowledge for searching for components such as software design extract heart in a knowledge base. You can search for software components accurately, even if they do not necessarily meet the required specifications.
Since similar parts can be searched, the accumulated parts database can be effectively utilized.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing the overall structure of an embodiment of the present invention;
FIG. 2 is an explanatory diagram showing an example of the configuration of standard parts in the present invention, FIG. 3 is a flowchart showing the overall inference control procedure in the present invention, and FIG. 4 is a flowchart showing the inference processing procedure for parts search in the present invention. It is. 1... Reasoning processing device, 2... Knowledge base, 3...
Standard parts search knowledge source, 4... Similar standard parts ice inspection knowledge source, 5... Blackboard, 6... Inference control information area, 7... Input/output device, 8... Parts database, 9
... Standard parts configuration example, 10... Standard part name, 11
...Parts specifications, 12...Parts source code, 13
・・・Parts Specification No. 3, Figure 4

Claims (1)

[Claims] 1. In a search method that has a knowledge pace that has accumulated knowledge that associates the external functions and usage environment of software components with required specifications, and a component database that has accumulated software components, input the characteristics of the software to be developed and match it with the knowledge. When a usable standard part is not found, matching is performed using knowledge that partially deletes the environment and external function conditions required for the part to find a usable part. 1. A search method for software components, comprising the steps of inferring a name, searching for similar standard components from a component database, and creating a required component by referring to the standard components.