JPS5868140A - Character processor - Google Patents

Abstract

PURPOSE:To obtain a device having a dictonary with superior expansibility while the use efficiency of memories is improved, by providing the device having a dictionary using pieces of phonetical information as indexes with memories stored with plural divided phonetical reading elements. CONSTITUTION:A CPU1 processes input data from a keyboad 5 and the processed data are displayed on a CRT8. An RAM14 is a random access memory, and the memory 14-1 includes 16 kinds of partial dictionaries to classify words. The 1st starting character of reading information is indicated in a dictionary index table, so it is not stored. In a memory 14-2, a pointer for indicating what position of a partial dictional a word starting with optional KANA (Japanese syllabary) is stored at in the dictionary index table is set up. A memory 14-3 stores a partial dictonary starting code table, and a memory 14-4 has a partial dictionary starting address table. A memory 14-5 is a register for storing inputted character data strings, and a memory 14-6 stores word codes and grammatical information.

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to a character processing device having a dictionary that prepares one persimmon of reading data as an index in a separate area and does not register it in the reading field of the dictionary, for example in a kana-kanji conversion method. be.

Conventionally, when inputting kanji, the kana-kanji conversion method inputs kana from the keyboard and converts it into kanji. There was a dictionary on top, but
The reading field of the dictionary contained all the readings of that word.

The order in which words are stored in a dictionary is usually in a so-called lexicographical arrangement, for example, in a lexicographical arrangement, so words with the same first letter are registered in a certain location.

For example, the single RFt ``6kikai'' starts with the first letter ``ki'', and the words registered in the surrounding VCs of ``1kikai'' are words that start with ``1kion1kisetsu'lCdo1ki''.' 9Fi ',! For example, if there are 1000 candy words, ``ki'' will be registered 1000 times in the ??t field, and a memory area of 100 bytes will be used.

Thus, conventional dictionaries had the disadvantage that long characters were registered in the kan and mi fields.

In view of the above-mentioned points, the present invention provides a character processing device that divides reading information as an index into a plurality of pieces and stores them in six separate memories in association with each other.

In order to achieve the above-mentioned object, the present invention does not register the entire pronunciation of a word in the pronunciation field of a dictionary, but instead registers the pronunciation minus the first character -1. Then, an index table is created by collecting only the first characters in a separate location, and a pointer to the position in the dictionary starting with the first character is stored. The index table is a table consisting only of pointers, and the reading of the first character is not registered. For example, when searching for a word starting with ``6'', the dictionary can be searched from the position indicated by the seventh pointer in the index table.

In the pre-self dictionary, the first character is not registered.

For example, the readings 8kikai'°, ° ``kion'', and 1kisetsu'' are registered as ``kai'', ``on'', and ``se1tsu'', respectively, so each requires only a 2-byte reading field. .Also, in terms of search time, for a 6-character word with a total of 1 words, the comparison is 1 word with 2 words.
Since the search is performed for the number of characters, and the dictionary to be searched is limited and narrowed by the index, the time is shortened.

This makes it possible to create a dictionary that saves memory space and reduces search time.

-2 Z finish: Shun S: b admonition I cut: Me-; [:” J-i force 7-self-
+(+: Please refer to 1i(11, 1! II.

First, the structure of the dictionary containing W+(1) in the memory of the character blocking device 1' will be described below.

As shown in Figure 1, read and read 11 kanji for each year.

Contains grammar information. l, l' is candy (y■,
EN-(-KLEN -1-1) tM f at r7-de
k, will be done. For example, 1 read is 16 hits (2 bytes)
Consists of. The first YT, l x 2 bytes of this 110-component single file are invitation information1), and then the next KI following the last YI, 'l.:N words.
By inserting i; The dictionaries a to p are constructed.Hereinafter, the partial dictionary with the reading field length YLBN and the kanji field length KL'BN will be referred to as the (YLEN, KLEN) dictionary.For example, the (1, 2) dictionary is a partial dictionary consisting of one word in the reading field and two words in the kanji field.Since each partial dictionary has a fixed length format inside its partial word W, the lp word is stored from the word code. It is easy to calculate addresses (hereinafter, a dictionary with this structure will be referred to as a semi-fixed length dictionary).
An embodiment of the present invention including a dictionary as described above will be further explained.

FIG. 3 is a block diagram showing one embodiment according to the present invention.

In the illustrated configuration, 1 is a microprocessor, 2
is an address bus, which specifies the address of each person's output memory. In addition, 3 is a data bus, which is a bidirectional bus used for transferring various data, and 4 is a flight control bus, which is used for reading, writing, and accepting data from each memory, etc. There is one piece to transmit the timing etc. Furthermore, it is a 5c1 Gee board controller. The keyboard controller 6 encodes the data inputted through the keyboard 5 and sends an inter-raft signal to the control bus 4. Next, 7
is 0TIT controller A; I, CIIT device 8,
Character generator 9. Refresh memory 10.
The reflex memory controller 1] is controlled. Further, 8 is an o+t'i' device, which can perform a display similar to a normal television receiver, and 9 is a character generator, which inputs codes such as character codes and line addresses of character patterns. 10 is a refresh memory, which stores the character code to be displayed in alt7"41i18, and outputs the pattern of that line out of the character pattern sequentially. II is a refresh memory controller that displays 0ItT in response to instructions from the CRT controller 7.(1) A series of refresh cycles and data from the microprocessor I Carrying the
Furthermore, 12 is a read-only memory, that is, the so-called 1 (OM), which performs the control procedures shown in FIGS. In addition, 13 is a controller for the ROM 12, and 14 is a random access memory (Lj).It is a so-called RAM, and is used to store the dictionary and each gentleman's table, as well as temporary variables for names. Used for commemoration.

14-1 is a semi-fixed length dictionary and includes 16 types of partial death. Each single death is classified into one of these 16 types and registered. (yr, nN.

KL 箕N) Dictionary has reading YLBN x 2 characters, kanji K
T, 'EN letter words and rules (yLnNX 2 +
1) Words of characters and Kanji KLBN characters are registered. For example, in the (3,2) dictionary, a word with 6 pronunciations and 2 kanji characters and a word HjH with 7 pronunciations and 2 kanji characters are registered.

(0,1) Only words consisting of one character shima-mi+i'X+one character are registered in the dictionary. H4? The first character of ``mi'' is shown in the dictionary index table (mentioned NT) described later.
'ii'.

Zeng (has stored talent).

14-2 is the term νN index 7' with the index of the partial total V)-\. Get this
Knee+J-1:11) This is a dictionary index table memory added to the BV bottle to shorten 1. There is a pointer in memory indicating in which section of each partial dictionary the word starting with (() is stored in the first part of the dictionary for Renhui's kana. It has been desecrated.I
' When searching for "H:"+, use this pointer σ]
It is necessary to search from ``t %'' and then search the entire dictionary. For example, the search is started by following the pointer to the dictionary (1, 2).

You can find the machine 1.

14-3 is a partial dictionary start code table memory Ji・0] indicating the opening #i jl" ifi+ code of the partial dictionary.
As shown in FIG. 8, the start code of each partial dictionary is stored for each partial dictionary. Word codes are assigned to every word in each partial dictionary.

The pointer PTH of the fRSr 'M reference table memory JTND is set to the partial dictionary start code table memory JTOD.
If you add the starting single item code found in , you will get the search starting word code. Also, if you add the start address to the pointer P'l'H based on the required storage capacity per candy for that partial dictionary, you will get the dictionary address at which to start the search.From the three-word code, you will find the information that the word has. When you want to know the partial M book start code table memory JTOD, look up the partial dictionary that includes the word, and set the start address to the word code minus the start code multiplied by the storage capacity required per word in the dictionary. By adding the word, you can get the address where that word is not registered.

14-4 is a partial dictionary start address table J'I'AD indicating the start address of a partial dictionary, and FIG. 8 is a partial dictionary start address table J'I'AI) necessary for searching a semi-fixed length dictionary. (1c, showing the formation of -4).

For each partial measure, the starting address of that partial dictionary is stored. The single f, 76 code is numbered sequentially in each partial dictionary.
If we add the start word code found in 'I' 0 ]), we get -S - CH start simple food. Also pointer I”
One unit of that partial dictionary to rR. Adding the starting address along with the required storage capacity per word yields the dictionary address at which to start the search.

When you want to know the information of the word d11 from the word code, look up the partial dictionary that includes that word in the dictionary lookup table memory JTO'D, and subtract the start code from word:1-d to 11'L of that dictionary. By multiplying the required storage capacity per candy and adding it to the Iti address,
The address where the single H:i is registered can be obtained.

Reference numeral 14-5 denotes a search word register 5WII which stores a character data string entered manually, and as shown in the 91st line j, this register stores the input character data string.

The size of register 8WB 14-5 is N words,
You can change the size as needed. As shown in the figure, regarding the input, the input start code II (u is entered first, then the input character data string is entered, and finally the input end code 9) is entered.

14-6 stores word codes and grammar information corresponding to each reading data - Sarna word table memory 8WTB
At L, as shown in FIG. 10, the reading data OE, the word code 'I'0 of the word retrieved from the dictionary, and the grammar information BI are stored for each of the above-mentioned 芊read elements. For one on-yomi element, one word of reading data and one word code
word.

It consists of three words in total, one word of grammatical information. The total number of words can be determined as required, and 18 is a controller for a 70-twin drive.

Further, 19 is a printer, and 20 is a controller for the printer.

The '4-invention character processing device comprising the above-mentioned configurations and elements is operated by input from the keyboard 5, and when an input is made from the keyboard 5, an interrupt signal is sent to the microprocessor 1. I was told, T
10M12 performs various control operations that are safe for character processing in accordance with control procedures stored in the memory.

Next, refer to flowchart 1 shown in FIG. 4 +tt+L,
The control operations of the character processing device 6' of the present invention will now be explained. First, in step 4-1, a series of data output from the keyboard controller 6 is stored in the RAM 1.
Write to the input data register SWB in 4.

Next, 4-2 Kana Kanji Hen PiI! Determine whether the beginning is specified or kutsu, and if it is not the beginning of the kanji conversion, step 4-
Moving on to step 3, various manual processes such as inputting ordinary English characters, kana characters, inputting character codes, etc., are performed. Such input processing is similar to input processing performed by a normal word processor, etc., and its explanation will be omitted. On the other hand, if it is the beginning of kana-kanji conversion, the process moves to step 8-4.

In step 4-4, the input character data string in the register BWB is divided into on-yomi elements νI as shown in FIG.

In step 4-5, a dictionary search is performed for the pronunciation obtained in step 4-4.

In step 4-6, the word d to be searched as a result of the dictionary search.
It is determined whether or not l) has occurred, and if not, the process moves to step 4-8. If it exists, move to step 4-7,
Reading data, word food, and grammar information are registered in SW'l'BL as shown in FIG.

In step 4-8, it is determined whether or not the dictionary search has ended. If not, the process moves to step 4-4 and the next on-yomi element is created.

In step 4-9, kana-kanji conversion is performed using the memory 8WTBL.

Next, FIG. 5 will be explained. This part 1 explains step 4-5 in FIG. 4 in detail.

In step 5-1, the reading length of the on-yomi element to be searched for is determined and stored in the register WLBN.

In step 5-2, the registers WT and EN are divided by 2, the decimal part is rounded down, and the result is set in the register YIJN indicating the length of the I-stop.

In step 5-3, a register KLIiN indicating the number of backward characters is set to 1.

In step 5-4 (y■, vh, KT, BN
) Search the dictionary.

In step 5-5, as a result of the (YLI state, KLBN) dictionary search, it is determined whether or not the word of the desired on-yomi element is found. If it is found, a flag register indicating that it has been found is set to 1, and the dictionary search is performed. It goes through a series of processes.

In step 5-6, it is determined whether or not the value of the KIN register exceeds the upper limit, and if M1 is exceeded, the flag and register are reset to 0, and a series of processing called dictionary search is completed.

In step 5-7, the register KLEN is counted up by 1, and the process moves to step 5-4.

Such Stella 75-4 was designed in detail with reference to FIG.
4.

In step 6-1, the first character of the on-yomi element in register 5WB-hi and register YLBN.

Check out JIND in KIJN's Sen (YIJN, K
LEN) Set a pointer to the dictionary in PTR.

From JTOD at Stella 76-2 (YLEN.

KLEN) Finds the start code of the ascending movement and stores it in the register CD.

In step 6-3, (YI, E
N, KLBN) Find the starting address of the dictionary and set it in register AD.

Set to . The value of register 0BL8 is the value of register YI
It is determined by adding 1 to the value of register KLHN to the value of JN, and multiplying the result by 2.

By adding the numbers of the registers AI), the start address of number -CH is obtained.

In step 6-6, ``two lines i at l/s, jl(
i: Comparison of 1 single candy Sutsushima and Mi from T Food? j')
While doing so, (YIJN, KIJ3N) lW≠) → go to the notch.

In Saitama - A's acid 1 bamboo is 1 year old.

Next, one embodiment of the present invention will be explained.

That is, in the above-mentioned embodiment, the reading information of ll+%1'' (-:〒1A is classified as a partial dictionary).

and kanji information, but in addition to that, each word's Fi[%・Ki'9. If there are any items that give rise to iCf, they can also be included in the criteria and can be found in the M1-minute dictionary.
Although 6 types are adopted, the number can be increased as the required number l1Fx increases. ＃Recording a book 1 Vl) Reading V 舊 is divided into partial dictionaries classified according to the number of characters, so there is no wasted storage space for any word, In addition, no matter how long a new word is to be registered, it is only necessary to add a new partial dictionary to JII without changing the overall m-structure, which significantly improves memory usage efficiency and improves scalability. We were able to manufacture a character processing device with an excellent dictionary7.

[Brief explanation of drawings]

1] is a diagram showing the 11i candy structure of the dictionary. Fig. 2 is a diagram showing the extremely short length of a partial dictionary, Fig. 3 is a block diagram showing an embodiment according to the present invention, Fig. 4 is a 70-chart diagram showing a control Ikoma, and Fig. 5 is about 4 Fig. Figure 6 is a detailed flowchart showing step 5-4 shown in Figure 5. Figure 7 is a diagram showing the structure of a JIND table including an index of partial #books. , Figure 8 shows the structure of J'l'OD and J'P^■) which store the start word code and start address of the partial dictionary.
The figure shows the structure of the search word buffer that stores the input character data string. Figure 10 shows the 8W'l' old version in which the reading data, the word code of the corresponding word, and the grammar information are registered.・・・
...Random access memory 14-1...
・Dictionary 14-2...JTNI) Table 14-3...JTOD table 14-4...JTAD table 14-5...8WB register 14-6...8WTIIL table Applicant: Gyanon Co., Ltd. t1 4-1. +l! Ij 11 pieces 1 pay toy LEN
KLE pickpocket
[Karihi] prisoner 1m (U71) Mata]
==■A “B (2,1) m5 pieces=
E-“mouth (1,3,)hHW ding-door=
m 41.2)
(LL) JP 0U58-681
40(7)

Claims (1)

[Claims] A character processing device having a dictionary with on-yomi information as an index, and having a memory for storing each of the on-yomi elements divided into a plurality of on-yomi elements.