JPS6386030A - Link system between program languages - Google Patents

Abstract

PURPOSE:To increase the executing speed with the titled link system by using a procedure type language corresponding module, a knowledge processing language corresponding module and an access module to perform independent processes respectively. CONSTITUTION:A CPU 201 includes a C program module 202, a prolog interpreter module 203 and a fast node access module 204 as function parts respectively. The CPU 201 also contains a common data area 205 together with a prolog stack 207, a large quantity data base 208 and a knowledge base 209 secured in a prescribed memory space respectively. The data transfer control is carried out among those modules through the common data area 205 and two exclusive stacks. Thus the executing programs are processed in both knowledge processing and procedure type languages independently of each other. The transfer of a variable is omitted since an access is possible to the area 205 via a fact node from the knowledge processing language side.

Description

[Detailed Description of the Invention] [Purpose of the Invention (Industrial Application Field) The present invention aims to improve the execution speed of a program written in a knowledge processing language by using a procedural language to perform procedural processing included in the program. Concerning linking methods between programming languages for execution.

(Prior Art) As is well known, when a knowledge processing language such as prot09 is used, the program execution speed is slow, and therefore, when processing a large amount of original data, the processing time becomes excessive.

Therefore, it has been proposed to first write the procedural processing of a program in a procedural language such as FORTRAN or C language.

In this conventional case, as shown in the functional block diagram shown in Figure 6, when inference is performed while exchanging data between the inference execution means and the knowledge base in Prolog, the procedure is A type language is called, and at the time of this call, variables are received and passed to exchange information with prolog and the procedural language.

For example, as shown in FIG. 7, in Pr0100, aa
a (A, B): -bbb (A, C)#
When ccc (C, B), in C language, #ccc (C, B) int C, B B-C+1.

However, in such a conventional method, Pr
It is only possible to call procedural languages from olog, and there are specific limitations when that call is made.

(Problem to be solved by the invention) That is, the conventional method has the following problems.

(b) Only a format in which a procedural language is called from Prolog is available; the reverse format in which Prolog is called from a procedural language has not been realized.

(b) There are restrictions on the variables that can be passed.

(c) Procedural control must be executed in Pr0IOi7, making it difficult to improve execution speed.

The present invention has been made in view of the above-mentioned problems, and its purpose is to enable a programming language for knowledge processing @Yeki and a procedural language to call each other individually from both sides, and to set variables. Our goal is to propose a linking method between programming languages that eliminates the need for passing.

[Structure of the Invention] (Means for Solving the Problems) In order to achieve the above object, the linking method between programming languages of the present invention can be accessed by both a knowledge processing language and a procedural language. A procedure that defines a common data area in a predetermined memory space, and stores the execution state of a procedural language program in its dedicated stack when a knowledge processing language is called during program processing in a procedural language. It is possible to access the common data area defined by the procedural language compatible module and the procedural language compatible module, and when the procedural language 5n is called during the processing of a program using the knowledge processing language, A language compatible module for knowledge processing stores the execution state of the program in its dedicated stack, and when the common data area is accessed by the language compatible module for knowledge processing, the program fact (fact) fE1 is generated by the knowledge processing S g. Similarly, the common data area is provided with an access module that accesses the common data area using a fact clause, and the execution program can be written in a knowledge processing language and by hand by controlling transfer between the modules using the common data area and the two dedicated stacks. It is characterized by sharing processing with the continuation type language.

(Function) In the present invention, the functional block is configured as shown in FIG. 1, and includes a procedural language support module a, a knowledge processing language support module b, and an access module C. are doing.

The procedural language support module a defines a common data area in a predetermined memory space that can be accessed by both the knowledge processing language and the procedural language. When a user language is called, the execution state of a procedural language program is stored in its dedicated stack.

Knowledge processing language support module b accesses the common data area defined by the procedural language support module iq
Ru. On the other hand, the execution state of the program using the knowledge processing language is stored in its dedicated stack.

When the access module C accesses the upper common data area by the language compatible module for knowledge processing,
Program facts (fac) by KH for knowledge processing
Similar to clause t), the common data area can be accessed using the fact clause.

Transfer control between these modules is performed using the common data area at the top and the two dedicated stacks mentioned above, so the execution program is divided between the knowledge processing language and the procedural language. .

Furthermore, since the upper common data area can be accessed from the knowledge processing language side using the fact clause, there is no need to pass variables.

(Example) Examples of the present invention will be described in detail below.

FIG. 2 is a functional block diagram schematically showing the main configuration of an embodiment to which the present invention is applied.

In this embodiment, in order to make the CPU 201 function as the control center of the entire system, the functional units include a C program module 202, a prolo (fink printer module 203), and a fact clause access module 204. Equipped with.

The C program module 202 defines a common data area 205 that can be accessed by both Pr010g and C language in a predetermined memory space. At times, the execution state of a program written in C language is stored in the C stack 206.

On the other hand, the Prolog interpreter module 203 accesses the common data area 205 and reads p
When the C language is called during the processing of a program by roloi;l, the execution state of the program by prolog is stored in the prolog stack 20.

In addition, the fact “Access module 204 is Pro
When accessing the common database 205 by the log inkblitter module 203, the pr.

The common data area 205 is accessed using the fact clause in the same way as the fact clause of the program using log.

Therefore, the common data area 205 defined by the C program module 202 is divided into a header and a data area as shown in FIG. 3, and the data area is accessed via the header.

Furthermore, the database of the C program module 202 is a large database 208 that stores primitive data of the human world, and this large database 208 is secured in a predetermined memory space.

In addition, the database of the prolog interpreter module 203 uses a knowledge base 209 that stores rules for inference by Proto(l), and this knowledge base 209 stores rules for inference by Proto(l).
09 is also secured in a predetermined memory space.

Next, the operation of this embodiment will be explained.

First, the case where Pr010(] is called from a C program will be described.

In this case, the C program module 202 refers to the footfall database 208 while reading Prot.

Proto9 in 9 ink printer module 203
The data (status) necessary to execute is stored in the common data area 205. Thereafter, in order to perform inference using the Prolog interpreter module 203, the C program module 202 starts the Prolog inkblitter module 203, and stores the execution state of the C program at the time of starting in the C stack 206.

The prolog interpreter module 203 thus started by the C program module 202 uses the knowledge base 209 and the fact clause access module 204.
The inference is executed while referring to the common data area 205 through and through the common data area 205 respectively. When the execution of the inference is completed, the execution result of the inference is stored in the common data area 205 through the fact clause access module 204, and control is transferred to the C program module 202. At that time, the C program module 202 is returned to a bidirectional execution state by the C stack 206. however. The common data area 205 is excluded.

Conversely, can you call C language from Prolog? In the case of 1', the result is as follows.

When calling C language from this prolog, p
The rolog interpreter module 203 executes inference while referring to the knowledge base 209, and the data necessary for executing the C program in the C program module 202 is transferred to the common data area 205 in the form of a fact clause via the fact clause access module 204. Store in. Thereafter, in order for the C program module 202 to perform procedural processing, the proIog interpreter module 203 starts the C program module 202,
The execution state of the prolo(] program at the time of startup is pr.

Pay 20 ft of scale stack for log.

Thus, Pr0100 interpreter module 203
The C program 202 started up executes procedural processing while referring to the great circle database 208 and the common data area 205. Then, once the execution of the procedural process is completed, the execution result of the procedural process is stored in the common data area 205, and control is transferred to the Prolog interpreter module 203. At this time, the proloo interpreter module 203 returns to the previous execution state of the proloOg stack 20. However, the common data area 205 is excluded.

As mentioned above, when calling ProtoO from a C program, or conversely when calling a C program from proIOg, the common data area 205 shown in FIG. 3 is used. An example of the declaration of the common data area 205 is shown in FIG. 4, and how the common data area 205 is accessed will be described below.

First, when accessing data in the common data area 205 using the C language, the common data area is directly accessed.

On the other hand, when prolog accesses data in the common data 205, the fact clause access module 20
accessed indirectly via 4.

That is, when declaring a common data area by the prolog interpreter module 203, Prolo0
As shown in FIG. 5, when the 22 mm clause appears during execution of the prolog program, the interpreter module 203 accesses not only the knowledge base 209 but also the common data area via the fact clause access module 204.

Therefore, the fact clause access module 204
When viewed from the rolOg ink printer module 203, the contents of the common data area 205 are pr.

It will be accessed in the same way as the fact clause of the program by log.

For this reason, with the system configuration of this embodiment, (a) C language can be called from the proIOg side, and conversely, prolog can be called from the C language side.

(b) Prolog fact clauses can be inserted from C language.

(c) By writing the main routine in C language,
There is no need to write procedural control in prolog, and execution speed is improved.

You can get benefits such as:

In one embodiment, Prol is used as the eight knowledge words.

Although the present invention has been described using other knowledge engineering languages and tools, the present invention can be implemented using other knowledge engineering languages and tools. Further, although the C language has been described as a procedural language, FORTRAN etc. can also be applied.

[Effects of the Invention] As explained above, the linking method for programming languages to which the present invention is applied provides a common data area for knowledge processing languages and procedural languages, and two stacks dedicated to each language. By using , the execution program can be divided and processed between the knowledge processing language and the procedural language by controlling the transfer between the procedural language support module, the knowledge processing language support module, and the access module. .

Therefore, according to the present invention, between a knowledge processing language and a procedural language, it is possible to individually call the other from both sides, and there is no need to pass variables, which speeds up execution. Improved.

[Brief explanation of the drawing]

FIG. 1 is a functional block diagram outlining the linking method between programming languages of the present invention, FIG. 2 is a functional block diagram outlining the main configuration of an embodiment to which the present invention is applied, and FIG. 3 is a diagram showing the structure of the common data area, Figure 4 is a diagram showing an example of a declaration that uses the common data area, Figure 5 is a diagram showing an example of the operation of the fact clause access module, and Figure 6 is an overview of the conventional method. Another side block diagram shown in FIG. 7 is a diagram showing an example of information exchange for a programming language in the conventional method. a... Procedural language support module b... Knowledge processing language support module C... Access module 201... CPU 202... C program module 203... Prolog interpreter module 20
4...Fact part access module 205...Common data area 206...Stack for C 207-Prostack stack 208 for rolOg Foot database 209...Knowledge base 7j, l Figure 6th rolog C word 2 gccc (C+ B) intc, [3 B=C+1 Figure 7

Claims (1)

[Claims] (1) A common data area that can be accessed by both the knowledge processing language and the procedural language is defined in a predetermined memory space, and the knowledge processing language is called during program processing using the procedural language. A procedural language compatible module that sometimes stores the execution state of a program written in a procedural language on its dedicated stack; and a procedural language compatible module that can access a common data area defined by the procedural language compatible module, a knowledge processing language compatible module that stores the execution state of a program in a knowledge processing language in its dedicated stack when a procedural language is called during processing; and a knowledge processing language compatible module that accesses the common data area. and an access module that allows the common data area to be accessed by the fact clause in the same way as the fact clause of the program written in the knowledge processing language, and each of the above using the common data area and the two dedicated stacks A method for linking between programming languages, characterized in that an execution program is dividedly processed between a knowledge processing language and a procedural language through transfer control between modules.