JPS6150315B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to an access control method for a kanji pattern memory used in a kanji information processing device that handles kanji pattern information, such as a Japanese language processing system.

Conventionally, in processing devices that handle kanji pattern information, a magnetic disk device is used as a kanji pattern memory for storing kanji pattern information, and each time a character is selected, access to the magnetic disk device is controlled and the required kanji pattern is stored. I was extracting information. However, since such conventional means for retrieving kanji pattern information uses a magnetic disk device for kanji pattern memory, access time is long, and the more characters handled, the more time it takes to search for characters. Therefore, it takes a lot of time to extract the necessary kanji pattern information, which has the disadvantage that the kanji processing speed is significantly reduced.

This invention was made in view of the above circumstances.
The purpose of the present invention is to provide a kanji pattern memory access control method that can quickly and efficiently retrieve any desired kanji pattern information from the kanji pattern memory using JIS kanji codes (JIS-C-6226) as they are. shall be.

An embodiment of the present invention will be described below with reference to the drawings. Figure 1 shows the configuration of the entire device, and 1 in the figure is a processing unit (hereinafter referred to as
2 is this CPU's DMA (Direct
3 is the main memory (hereinafter referred to as MM) which is connected to the DMA-BUS 9, stores software programs, and stores various tables for character pattern position searches. A large capacity file memory (hereinafter referred to as FM) 4 is connected to the DMA bus 9 and stores all character patterns including kanji handled by the device, as well as dictionaries, etc., and is provided with a table for searching character positions. A kanji pattern memory using a high-speed random access memory that is connected to the I/O bus (I/O-BUS) 10 of the CPU 1 and stores a predetermined number of character patterns necessary for processing from the FM 3 when the device is in operation. (Hereinafter referred to as KPM
5 is a display control unit (hereinafter referred to as CRT-C) connected to the I/O bus 10;
6 is a display refresh memory (hereinafter referred to as RM), 7 is a CRT display section, and 8 is a KPM bus (KPM-BUS) for direct access to KPM4.
It is. Then, before starting the system, the CPU 1 reads a predetermined number of kanji patterns (for example, 1650 characters) from the FM 3 and converts them into KPM.
After the predetermined kanji pattern is registered in KPM4, the JIS kanji code (JIS-C −
6226), determine the position of the kanji pattern registered in KPM4, read the kanji pattern from FM3 as necessary, and register it in a predetermined area (variable area described later) of KPM4. In this case, the various tables shown in Figure 3 a to e located in MM2 and the pattern position information table shown in Figure 5 located in FM3 are used to extract the kanji pattern. . In addition, the FM3 has the storage format shown in Figure 2 a and b, and stores all the kanji patterns handled by the device in advance.
Kanji pattern layout in the sector, b in the same figure is 1
The kanji pattern layout in the track is shown for each. In addition, the size (capacity) of the KPM4 mentioned above is determined by the system, but usually, for example,
The configuration is 128KB (kilobyte), 64KB, etc. The kanji pattern for one character handled here consists of 72 bytes, and the KPM4 is provided with a fixed area and a variable area, and the fixed area stores frequently used kanji patterns for, for example, 1650 characters. For example, kanji patterns for up to 50 characters are stored in the variable area as needed. KPM at this time
The memory configuration of No. 4 is shown in FIG. The predetermined kanji pattern (1650 characters) stored in FM3 is written in the fixed area under the control of CPU1 at system startup, and the variable area is written in during operation. Also CPU if necessary
1, the kanji patterns read from the FM 3 are written. If the number of characters registered in the variable area exceeds its capacity, the oldest character patterns are sequentially erased.

The effect will be explained here. First, when explaining the detailed operation of this invention, the general flow of data will be explained. The input to the CPU 1 is a 2-byte JIS Kanji code (JIS-C-6226),
By operating the proposed logic,
2-byte information indicating the position of the kanji pattern in KPM 4 is created by CPU 1 as output. This output is given to the CRT-C5 via the I/O bus 10 and used for displaying kanji. As the above logic operates, data, ie, Kanji patterns, are transferred between devices in the manner described below.

(1) When the kanji pattern corresponding to the given JIS kanji code is in the fixed area or variable area in KPM4. At this time, the data is
By controlling CRT-C5, KPM4 is changed to KPM
It is moved to RM6 via bus 8.

(2) When the kanji pattern corresponding to the given JIS kanji code is not included in KPM4. At this time, data is extracted from FM3 and moved to KPM4 by CPU1. Further, the CPU 1 provides position information to the CRT-C5. CRT-C5 moves data from KPM4 to KPM6 via KPM bus 8.

Next, we will explain in detail the process of extracting the corresponding kanji pattern from a given JIS kanji code. First, CPU1 determines whether the given JIS Kanji code is in KPM4.
It is carried out by. For this judgment, CPU1
uses the table shown in FIG. 3a, which is stored in MM2. This table corresponds to approximately 10,000 characters in the JIS Kanji code, and is hereinafter referred to as EXIST. One bit in this EXIST corresponds to one JIS kanji code, and
The kanji pattern corresponding to the kanji code is KPM4.
It is used to indicate whether it is inside or not. In other words, when a 16-bit (2-byte) JIS Kanji code is input, the code/address conversion logic of CPU 1 operates, extracts the bit information corresponding to the given JIS Kanji code, and changes the state of this bit. It will be judged. If the bit is "0" here, it indicates that the required kanji pattern is not in the fixed area of KPM4. Also, the bit is “1”
If so, it indicates that the required kanji pattern is in the fixed area of KPM4. Furthermore, in MM2,
As for the JIS Kanji code, a table (referred to as NONEXT) shown in FIG. 3b is also stored in order to detect non-existing characters as errors. As shown in FIG. 3b, this table consists of a code string with the beginning and end of a non-existing code (the beginning and end are sometimes the same). By referring to this table one by one, it is determined whether the input JIS Kanji code is valid or not. If it does not exist, it is determined as an error.

Next, we will explain the case where it is not in the fixed area of KPM4, but exists as a JIS Kanji code.
In this case, if it is determined by the above check that it is not in the fixed area of KPM4 but exists as a JIS kanji code, then it is checked whether the corresponding kanji pattern exists in the variable area of KPM4. . For this check, the table shown in FIG. 3c (called VARTBL) stored in MM2 is used. Nothing is registered in this table at the start of operation. By checking this table, the process branches to the following two steps a and b.

(a) When the corresponding kanji pattern is included in the variable area of KPM4. The position of KPM4 and the position of table (VARTBL) are always made to correspond one-to-one. That is, in the first entry of the table, a corresponding pattern is stored from the first position VARPT of the variable area in KPM4. The second entry stores the corresponding pattern starting from VARPT+72. Therefore, once the entry position is known, ADRS shown below = {(VARTBL position) - 1} x 72
......The position of KPM4 (ADRS) can be easily determined using the calculation formula (2).

(b) When the corresponding kanji pattern is not included in the variable area of KPM4. Extract the corresponding kanji pattern from FM3 and write it into the variable area of KPM4. Approximately 10,000 kanji patterns are stored in the FM3, and in order to determine their location information, a number of location information tables (DISLOC) corresponding to the wards of the JIS kanji code are stored, as shown in Figure 5. ). The location information table is indexed from the ward information of the JIS kanji code, and the location of the kanji pattern can be known. This table is large because it is referenced infrequently, and
It is stored in FM3, which takes time to access. Advance VARTBL position and KPM to that position
In addition to writing the position information of FM 4, the corresponding kanji pattern extracted from FM 3 is also written in KPM 4. When VARTBL becomes full, the oldest data is deleted. That is,
VARTBL is used for cycling. In order to know the usage status of this table, two pieces of information are held: the table's use start position and use end position. Naturally, this information is updated every time it is registered in this table.

Next, if the kanji pattern corresponding to the given JIS kanji code exists in the fixed area of KPM4,
If it is determined, an explanation will be provided. In this case, the two tables NBRTBL and PTEXTN shown in FIG. 3d and e, which are stored in MM2, are used.
The first item in the NBRTBL entry indicates the number of codes in ward "1" whose kanji pattern is registered in the fixed area of KPM4. The second term indicates the number of codes in ward "2" whose pattern is registered in the fixed area of KPM4. For the third and subsequent areas, in the following corresponding wards,
Indicates the number of registered patterns in KPM4. Since a certain JIS Kanji code is given, based on the value of its first byte (corresponding to ward),
Add all the numbers registered from ward "1" in NBRTBL to the previous ward of the given code. Next, the corresponding code is given for each ward.
An information table (PTEXTN) for checking the position in KPM4 is indexed.
One entry in PTEXTN consists of two parts. In other words, within each district, the corresponding pattern continues from the part indicating the beginning of the code in KPM4 and from that code.
There are two parts that indicate the number of codes that contain the corresponding pattern. From this table, it is possible to find out what number, starting from the point "1" of a given JIS kanji code, the corresponding pattern is located in KPM4. The above operation is shown in the flowchart of FIG.

As mentioned above, in order to determine whether the kanji pattern corresponding to a given JIS kanji code is in the fixed area of KPM4, a table corresponding to about 10,000 characters of JIS kanji code is used, and
Once it is determined that the pattern is in the fixed area of KPM4, a table corresponding to the wards and points of the JIS Kanji code is used to calculate the position information of the pattern, which significantly reduces the time required for calculation. Ru. Furthermore, for kanji patterns that are used infrequently, a variable area is provided in KPM4, and a table is used that allows checking the existence of the pattern from the JIS kanji code.
can be used efficiently.

As detailed above, according to the present invention, there is provided an access control method for a kanji pattern memory that can quickly and efficiently retrieve any required kanji pattern information from a kanji pattern memory using JIS kanji codes as they are. Can be provided.

[Brief explanation of the drawing]

The figures are for explaining one embodiment of the present invention, and FIG. 1 is a block diagram, FIGS. 2 a and b are diagrams showing the storage format of kanji patterns in a large-capacity file memory, and FIGS. 3 a to e is a diagram showing various tables provided in the main memory, FIG. 4 is a diagram for explaining the storage area configuration of the kanji pattern memory, and FIG. 5 is a diagram showing the tables provided in the large-capacity file memory. , FIG. 6 is a flowchart showing the operating state. 1... Processing unit (CPU), 2... Main memory (MM), 3... Large capacity file memory (FM), 4
... Kanji pattern memory (KPM), 5 ... Display control unit (CRT-C), 6 ... Display refresh memory (RM), 8 ... KPM bus.

Claims (1)

[Claims] 1. A large-capacity file memory in which all the kanji pattern information to be handled is stored, a kanji pattern memory having first and second areas in which part of the kanji pattern information of this large-capacity file memory is stored, and the large-capacity file a first table in which position information of the kanji pattern stored in the memory is stored;
A second area that is addressed by a JIS kanji code and indicates whether or not the corresponding kanji pattern information is stored in the first area of the kanji pattern memory.
table, a third table in which JIS kanji codes for all of the kanji pattern information stored in the second area of the kanji pattern memory are registered, and a third table in which character codes that do not exist as JIS kanji codes are stored. and a fourth table for determining whether or not the input JIS kanji code is valid by indexing the second table. If the kanji pattern information indicates the existence of the kanji pattern, extract the kanji pattern information from the first area of the kanji pattern memory, and if the second table does not indicate the existence, the fourth table is indexed and the validity of the input JIS kanji code is determined. After that, the third table is indexed, and if the presence is indicated here, the kanji pattern information is retrieved from the second area of the kanji pattern memory, and if the presence is not indicated, the kanji pattern information is retrieved based on the first table. A kanji character that is characterized in that the corresponding kanji pattern information is read from a large-capacity file memory, this kanji pattern information is stored in a second area of the kanji pattern memory, and then the corresponding kanji pattern information is retrieved from the second area. Pattern memory access control method.