JPS62274343A - Type conversion and selection system - Google Patents

Abstract

PURPOSE:To improve the execution efficiency of a program by developing a program by an automatic converting function, executing the developed program in a mode where automatic conversion is not performed, and performing error processing and correcting a different type. CONSTITUTION:An instruction inputted from a terminal 7 is interpreted by an interpreter part 8 and executed. In such case, an instruction interpreter part 23 interprets the instruction, converts it to an easy-to-execute form, and outputs the instruction to an intermediate table 24. The instruction is executed by an instruction execution part 25 by using information in the table 24 and when the inputted instruction is a macroinstruction, a macroinstruction definition file 9 is referred to execute instructions described in the macroinstruction one after another. Further, when the instruction is an instruction for referring to data in a data base 1, the data are referred to through a data base system 10. A programmer develops a program without considering the coincidence nor dissidence of data because of this automatic converting function. When, however, the program is executed, the type of data in the data base 11 and data on variables are changed to improve the execution efficiency.

Description

[Detailed description of the invention] [Industrial application field] The present invention relates to a calculation processing system for an electronic computer.

In particular, it relates to type conversion options suitable for improving program development efficiency and execution efficiency.

[Conventional technology]

In programming languages such as Fortran, for example, as described in "Introduction to rFoRTa*N77" published by Baifukan, for operations between numbers such as integer quantities and numeric types, there are Mixed operations are allowed, such as adding numbers. This is because the compiler automatically creates type conversion instructions.

[Problem that the invention seeks to solve]

When developing a program, since the above-mentioned mixed operations are allowed, the programmer does not have to consider each "type", which improves the efficiency of program development. However, when executing a program, there is a problem in that execution efficiency decreases by the amount of type conversion required.

In the case of programs that are used only a few times, it is important to improve development efficiency rather than execution efficiency.] One important issue is that for programs that are used repeatedly and require long calculation times, execution efficiency is more important than execution efficiency. Increasing efficiency becomes a more important issue. However, since the type conversion instructions issued by the compiler are done purposefully, programmers tend to overlook differences in types, making it difficult to create programs with high execution efficiency.

An object of the present invention is to provide a type conversion selection method that increases the execution efficiency of a completed program without reducing the program development efficiency, particularly in an interpreted language processing system that executes a single-source program. 0 [Means for solving the problem] The above purpose is to use the automatic type conversion function when developing a program, then run the developed program in a mode that does not perform automatic type conversion, and perform error handling. This is achieved by correcting the different "types" and improving the program.

[Effect]

Program development is my own! Because this is done using @ type conversion,
Development efficiency is high. After development, the program is run in a mode that does not perform self-gJ type conversion, and different "types" are corrected, so the program execution efficiency is increased.

〔Example〕 Hereinafter, one embodiment of the present invention will be described with reference to the drawings.

FIG. 2 is an overview diagram of a counter system used in an embodiment of the type conversion selection method according to the present invention.

Instructions (commands) input from the electric bed 7 are interpreted and executed by the interpreter Be. If the input next instruction is a macro command name, the macro definition file 9 is referred to and the instructions described in the macro are executed one after another.

Furthermore, if an instruction to refer to data in the database 11 is issued, the data is referenced by activating the database system 10.

An example of the command is shown below.

(Re SET MODE=, TRUL.

(2) WHILE, NOT, EOF (3)
FETCHX, Y (4) 5ETX=Y bone α1 (5) E N D W HI I, Iff command (
In this step, the logical 1 block, TRUB, is assigned to the variable μODD that holds the type conversion selection mode. Command (2) a, completes the data in the database and executes EOF (End at
Repeat commands (3) and (4) until File )9
This is a command to return. In command (3), select X from the database.
and Y, and in instruction (4) replaces the value of X with the value of yxα1.

FIG. 3 is a system variable table, and FIG. 4 is a user variable table. Each table has columns that hold variable names, types, values, etc. regarding variables. When command (1) is executed, the variable MOD in the system variable table in Figure 6
FfO (TRIJFf is entered in the i column. Also,
If the value of Y of X and Y retrieved in command (3) is the character string '+00', the value of X in the user variable table of FIG. 4 is entered as an integer 10.

FIG. 5 is a schematic diagram of the interpreter section of FIG. 2. In the instruction interpretation section 923, the instruction is interpreted, converted into an easily executable form, and outputted to the intermediate table 24.

The instruction execution unit 25 uses the information in the intermediate table 24 to execute instructions (su~(5), etc.).

FIG. 1 shows the processing procedure when the instruction execution unit 25 executes the instructions (1) to (5).

First, check the consistency of the stars using Nuteb 1.

In the case of instruction (4), since Y is a character type and fll is a real number type, the types are inconsistent, so branch at step 2 and test the automatic conversion mode at step 3.
Since the value of the variable MODE that holds the automatic type conversion mode is (, TRUE, in Figure 3), Y is selected in step 5.
After converting the character string °100° held by into a real value,
Execute the calculation on the right side of instruction (4) in step 6.〇SelfIg! In the Type 7 conversion mode, calculations are completed normally even when the data types are different, so the programmer can develop a program without being aware of the coincidence or mismatch of data presentation.

However, in the self-llIh type conversion mode, more time is required to convert the character string ``100° to the real number [100, 0'' than when executing the same instruction (4) as the Yt-real number type. Therefore, in order to improve execution efficiency, it is necessary to redefine Y to a real number type.

So, after developing the program for instructions (1) to (5).

Out of this program, execute the command (I) E=, FA
LSIL and execute the procedure shown in Figure 1. Then,
Upon execution of instruction (4), the determination in step 3 becomes negative (N) and the program branches to step 4, where the programmer learns where the data types do not match and requires type conversion.

Based on this result, the data types in the database and the data types of variables are changed to obtain a program with high execution efficiency.

In the error handling in step 4, an error message is displayed on the terminal 7, and as post-processing, (1) the instruction is skipped and the next instruction is executed, or (2) the execution of the instruction is stopped. There are cases when waiting for the next instruction to be input from the terminal, and cases where (3) type conversion is performed and execution of the instruction is continued. [Effects of the Invention] As explained in detail above, the present invention has advantages. According to the method, it performs the automatic type conversion mode and handles the function of cutting off inappropriate behavior, so it has the following effects.

(In the automatic MJ type conversion execution mode, there is no need to consider data type mismatches, so 70gram development efficiency can be improved. (2) In the automatic Km non-execution mode, the language processing system detects data type mismatches. However, since the location can be easily known, it is easy to improve the program to be more efficient.

[Brief explanation of drawings]

FIG. 1 is a language processing flowchart according to an embodiment of the present invention, and FIG. 2 is a configuration diagram of a computer system. FIG. 6 is a system variable table diagram, FIG. 4 is a user variable table diagram, and FIG. 5 is an explanatory diagram of the interpreter shown in FIG. 2. 07...Sea floor, 8...Interpreter section. °),

Claims (1)

[Claims] 1. In a computer system that has a language processing system that interprets and executes a programming language that describes the operation of the computer system, an automatic type conversion execution mode that performs type conversion when the data types that are the targets of the operation do not match, and an operation A type conversion selection method characterized by providing an automatic type conversion non-execution mode for outputting mismatch information when the target data types do not match, and selecting and specifying one of the modes.