JPH0523458B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION <Field of Industrial Application> The present invention relates to electronic dictionaries, and particularly to improvements to user dictionaries therein.

<Prior Art> Computer systems, word processors, electronic typewriters, etc. that handle word information are equipped with electronic dictionaries to check whether there are any errors in the spelling of input words. This electronic dictionary consists of a main dictionary and a user dictionary.
The main dictionary holds word information supplied in advance by the system, and the user dictionary holds word information registered by the user.

As shown in Figure 6, normal spell check searches to see if the main dictionary has the same information as the word information, and if the same information is not found, searches to see if the user dictionary has the same word information, and then searches the main dictionary to see if the same word information exists. Or, if the same word information exists in the user dictionary, but the spelling is not correct, it is determined that the spelling is incorrect.

<Problems to be Solved by the Invention> In the spell check described above, even if the main dictionary has a good data structure and processing is fast, if processing of the user dictionary takes time, the overall processing will be slow. Unlike the main dictionary, the user dictionary is not provided by the system, but allows the user to freely register words, so it must have a certain degree of compatibility. That is, it is necessary to be able to realize high-speed processing like the main dictionary, and to be able to freely output a list of word information registered in the user dictionary. Since the area of the user dictionary is limited, when the data area expires, it is impossible to determine which word information to delete without the list.

FIG. 7 shows the structure of a user dictionary using only the Hash method. In the Hash method, the word information registered by the user is registered by combining numerical data and Hash, that is, meaningless information for making the data block length constant. In this user dictionary, index data is allocated on average by Hash, and the data search speed is fast. However, when the data section expires or when you want to check the currently registered words, you can It is difficult to reproduce the words. Another problem is what to do when the words registered in the user dictionary happen to have the same Hash value. same
If it is used as a hash value, etc., the next time you delete it, you don't know which one to delete or which one to keep.

Figure 8 is an example of using character strings without using Hash values. When just character strings are arranged in this way, the characters are compared according to the length of the word, but while it is possible to see what kind of words are registered in the user area, when using spell check, it is difficult to compare the characters by word length. Since it is a comparison, the processing will inevitably be slow.
To cover this, a method is sometimes taken to arrange the words in the user area in order of frequency or in the order of most recent use, but if a misspelled word is actually entered, it will still be displayed until the end of the user area. Searching takes time. Furthermore, as shown in Figure 9, even if an index that does not use Hash, for example, an index is constructed from the initial letter of a word or the length of a word, data is not allocated on average like an index that uses Hash (c There are many 8-letter words starting with ``or'', but there are almost no 8-letter words starting with ``z'' or ``x.'' Therefore, depending on the word, it takes longer to search than when using the Hash method.

<Means for Solving the Problems> The electronic dictionary according to the present invention includes a main dictionary that holds word information supplied in advance by the system, and a user dictionary that holds word information registered by the user. The electronic dictionary has a function of checking the spelling of a word by referring to both of the dictionaries mentioned above, and the user dictionary has a first area where the words registered by the operator are stored as numerical data using a hashing method. and a second area for storing the word as character data, the first area being used for spell check reference, and the second area being used for operator confirmation. It is characterized by It is characterized by

<Example> FIG. 1 shows the structure of a user dictionary according to this example.
This user dictionary 1 includes an index section 2, a data section 3 consisting of an area for storing words registered by the user as numerical data using the Hash method, and a data section 4 consisting of an area for storing the above words as character data. configured.

FIG. 2 shows the configuration of a computer system using this user dictionary 1. The input device 5 is a device for inputting character and word information into the storage device 6,
For example, it consists of a keyboard device, a tablet device, an OCR device, a magnetic tape device, etc. The storage device 6 is an area that stores information on characters and words inputted from the input device 5, and includes, for example, a core memory, an IC memory, a magnetic disk, and the like. The output device 7 is the storage device 6
It is a device that outputs the information saved and edited in the computer, and includes, for example, a printer, a display device, a magnetic tape device, a magnetic disk device, etc. The spell check dictionary device 8 is a dictionary area for supplying valid information as appropriate in response to inquiries about the spelling information of characters and words stored in the storage device 6, and includes a main dictionary (not shown) and the above-mentioned spelling dictionary. It consists of a user dictionary 1. This spell check dictionary device 8 includes, for example, a core memory, an IC memory,
Consists of ROM, magnetic disk, etc. Control device 9
consists of a computer that controls the exchange of signals and data between the above components.

The effect will be explained below.

When performing a spell check, the Hash value of the word is calculated using the Hash method, and the data is searched at high speed by the index section 2.

The data section 4 is mainly used to register, delete, and output a list of word information in the user dictionary 1.

When registering a word in the user dictionary 1, as shown in Figure 3, first search to see if the same word information exists in the main dictionary (step #11). Search for information (Step #12). If the same word information exists in the main dictionary or data section 4, an error will occur. When there is no same word information,
A Hash value is created for the word information (Step #13), and a search is made to see if the same Hash value as the created Hash value exists in the data section 3 (Step #14).
If the same Hash value exists, turn on the double flag (Step #15). The hash value is represented by 27 bits, and the first 5 bits constitute the index, so the data part is represented by 22 bits. However, in reality, the data part 3 of the user dictionary 1
is stored in 3 bytes, so there is a margin of 2 bits. Therefore, one of the two bits is ON.
If it is, it means that the hash value is being used twice. After the above processing, the index is counted up, the hash value is set, and character data is added to the data section 4, thereby completing the registration of word information (step #16).

To delete a word from user dictionary 1, use the fourth
As shown in the figure, first, the user dictionary 1 is searched to see if there is character data for the word in question (step #21), and if character data is found, this character data is deleted from the data section 4 (step #22). Next, a Hash value is created for the word (step #23), and the same Hash value in data section 3 is flagged as double.
Search to see if it is ON (Step #24).
If the double flag is not ON, delete this Hash value (Step #25) and count down the index (Step #29). On the other hand, if the double flag is ON, the Hash value is calculated using the Hash method for all words in User Dictionary 1, and the same value is calculated using the Hash method.
Check whether there are more than one Hash value (Steps #26, #27), and if there are more than one, count down the index (Step #29),
If there is only one, turn off the double flag (step #28) and count down the index (step #29).

When outputting a list of word information registered in the user dictionary 1, as shown in FIG. If there is, this character data is sent to the output device 7 (step #32). Since the data section 4 holds word information in alphabetical order of initial letters, the list can be output in an easy-to-read manner.

<Effects of the Invention> As explained above, in the present invention, the first area stores the words registered in the user dictionary as numerical data using the hashing method, and the second area stores the words as character data. Consisting of
The first area is used for spell checking reference, and the second area is used for operator confirmation, so when spell checking a user dictionary, Hash values can be used for high-speed processing. In addition, processing that does not require high speed, such as registration, deletion, and list output, can be handled by using character data, and the functionality of the electronic dictionary can be improved.

[Brief explanation of the drawing]

FIG. 1 is a diagram showing the structure of a user dictionary according to an embodiment of the present invention, FIG. 2 is a block diagram showing the configuration of a computer system to which the present invention is applied, and FIGS. 3, 4, and 5 are diagrams showing the implementation of the present invention. FIG. 6 is a flowchart showing the processing procedure of a spell check, and FIGS. 7, 8, and 9 are diagrams showing the structure of a conventional user dictionary. 1...User dictionary, 2...Index, 3,
4...Data section.

Claims (1)

[Claims] 1 Equipped with a main dictionary that holds word information supplied in advance by the system and a user dictionary that holds word information registered by the user, the spelling of words input by the operator is checked by referring to both dictionaries. The user dictionary has a first area that stores words registered by the operator as numerical data using the hashing method, and a second area that stores the words as character data. An electronic dictionary characterized in that the first area is used for spell check reference, and the second area is used for operator confirmation.