JPS5839335B2 - How to identify and print desired characters in a Chinese kanji search system - Google Patents

Description

[Detailed description of the invention] TECHNICAL FIELD The present invention generally relates to a search system for Chinese characters.

In particular, the present invention relates to a method for identifying and printing Chinese characters applicable to a computer search system equipped with a conventional input keyboard terminal device.

In Chinese, the basic spoken language is written in over 10,000 different characters, so conventional systems for printing Chinese characters were complicated.

Because of this large number of characters, it is practically impossible to use something like a conventional English typewriter to print Chinese characters.

This is because one key is required for each word to be printed, and in total more than 10,000 keys are required.

Conventional well-known Chinese search systems do not allow operators to select and print any Chinese kanji from a list of all available Chinese kanji without special training and based on their ordinary knowledge. System functionality is severely limited.

Telegraphic transmissions illustrate the difficulties of using such large numbers of characters.

One telegraph system utilized a codebook listing approximately 9,000 characters and indexing each character with a 4-digit number, so that the 4-digit number was transmitted by telegram and the telegram was translated into words at the receiving station. It is translated into Kanji using a four-digit number to find the corresponding character in the list.

The creator of the message is forced to use only the characters listed in the codebook.

This is because it is clear that characters not included in the codebook will not be transmitted.

One typewriter currently available includes a tray of movable type from which characters are selected one at a time, pressed onto typewriter paper, and returned to their original storage position. .

The selected characters are controlled by a single selection bar that moves over the tray of type.

The main difficulty in using this typewriter is that it must memorize dozens of type locations in order to operate efficiently.

Other Chinese typesetting machines currently available utilize a keyboard with 27 columns by 44 rows of keys, each key controlling two characters, resulting in a total of 2.376 characters. It will be done.

This system has a one-to-one correspondence between the keys and the frozen dessert list, so the operator must memorize the location of each character.

Other machines used for Chinese translation divide all the characters in the word list into two parts, and use the radicals that appear at the top and bottom of the valley character to create characters that all have these two parts in common. Define the group of .

These groups are optically displayed and the final selection is made by selection from the visual display.

Thus, identifying any character in a frozen dessert requires only three keystrokes, two of which are used to display the group containing the desired character, and a third. The one is used to select and print characters.

This system is expensive and, since the indexing system is not based on common knowledge, the operator requires extensive training to become competent.

Other Chinese typewriters have been developed that are based on indexing systems that use standard stroke sequences commonly used in writing characters.

The key sequences used to stroke the letters are entered into a digital computer, which matches the sequences of strokes to the letters in the word list.

These machines typically require an optical output to resolve ambiguities and for final confirmation of the selected character.

This is because the typist may not know the requested stroke sequence or may have made an error in typing the stroke sequence.

Northeast Chinese (Mandarin) is used by millions of Chinese people who have no difficulty understanding the spoken language.
Their words can be faithfully recorded using the widely known Chinese phonetic alphabet.

The Chinese Pronunciation Unification Committee has been promoting the use of Chinese phonetic scripts since 1932, when it published the first edition of the standard pronunciation list of Chinese characters.

Unfortunately, indexing kanji phonetically does not create unique characters.

because. Chinese characters are easy to pronounce, no pronunciation is longer than a triphone, and homonyms are much more common in Mandarin than in English.

The standard pronunciation list created by the Chinese Pronunciation Unification Committee is approximately −
Contains a word list of three letters arranged into approximately 1300 groups of homophones.

Therefore, the phonetic index only provides a unique identification of homophone groups.

These homonyms could be viewed optically to make the final selection, but this method would be expensive and would greatly reduce printing speed.

Mandarin Chinese is used by millions of Chinese people who experience no difficulty in understanding the spoken language.

The ambiguity that exists when hearing words does not exist in normal speech.

This is because phrases that identify unique sequences of words from the homophone group are heard and understood.

The Chinese phonetic characters are used in many dictionaries and are commonly studied by school children.

This is so widely used in classes that it is even printed in daily newspapers, with the main text written in kanji and accompanied by phonetic symbols.

These prior art typewriting systems are disadvantageous in that they are slow (and difficult to train operators to perform the indexing tasks required to use them).

Furthermore, for searching for information on Japanese sentences mixed with kanji, there is a known character search device that can use a kana typewriter to search for a desired kanji from information such as on-yomi, kun-yomi, and decomposed reading of kanji.

However, in this character search device, the on-yomi (on-yomi)
If there are two or more kanji with the same reading, a lamp or bell will notify you, and the Kun-yomi (or On-yomi) will be input until there is only one kanji that contains all the multiple readings. Due to the functional configuration of inputting kanji and decomposition readings, there is a drawback that frequent interaction between the operator and the device makes it impossible to perform quick typing key operations and, by extension, to perform quick kanji searches.

In addition, this character search device must always check the entire kanji search dictionary for each search in order to determine whether there is one kanji with the same reading or two or more. After determining that there is a kanji, the dictionary must be accessed again to read the output code of that kanji, which greatly limits the search speed.

It goes without saying that such a character search device cannot be applied to a search system for Chinese kanji, which have completely different pronunciations from Japanese kanji even if they are the same character.

SUMMARY OF THE INVENTION Accordingly, an object of the present invention is to provide a Chinese character search system that overcomes the drawbacks of conventional systems and devices as described above.

Another object of the present invention is to provide a search system that can print desired Chinese characters by minimizing the number of keystrokes on the keyboard, making it possible to use a Chinese character typewriter with a conventional keyboard.

Another object of the present invention is to provide a Chinese character retrieval system that can quickly search and print desired Chinese characters by continuous keystrokes.

Another object of the present invention is to provide a Chinese character search system that can simultaneously satisfy various requirements such as ease of learning key operations, simplicity and uniqueness of keystroke sequences, and psychological comfort for the operator. .

Another object of the present invention is to provide a search system that can tolerate differences in speech coding between various dialects to some extent.

Another object of the present invention is to solve the ambiguity of homophones (whether there are two or more kanji with the same reading) without requiring display means such as displays, lamps, bells, etc. The object of the present invention is to provide a system that allows desired Chinese characters to be retrieved in a rapid continuous type sequence without requiring frequent interaction between devices.

The present invention provides a first phonetic symbol that at least partially represents the pronunciation of Chinese characters and descriptive features of Chinese characters, such as simplified naming of strokes, radicals of characters, naming of parts, and implied meanings of characters. An indexing system is used which encodes Chinese characters with additional second phonetic symbols representing, etc.

Other objects and advantages of the present invention will become more apparent to those skilled in the art regarding the present invention from the following description taken in conjunction with the accompanying drawings.

A preferred embodiment of the invention is shown in FIG.
It includes a computer 30, such as an M370-158 or a much smaller general purpose computer.

This computer has a keyboard 34 and an output printer 36.
It is used in combination with the input terminal 32 having the following.

A special purpose internal computer may be provided if desired.

In this case, the entire typewriter would look as shown in FIG.

To implement the present invention, all Chinese characters are encoded using only Mandarin phonetic symbols.

All words in Northeastern Chinese can be written phonetically using 37 different symbols, along with one of five tone marks to indicate the accent used when pronouncing them.

The Mandarin Chinese phonetic symbols are shown in Table 2, along with the corresponding International Phonetic Alphabet and its English equivalents.

Phonetic symbols are used in one or two sequences of one to three symbols each.

The first sequence is the coding of character pronunciation in standard Mandarin without specifying tone marks.

Because of the many homophones in which different letters have the same sound, the use of phonetic coding alone does not uniquely identify the letters.

A second sequence of phonetic symbols is used to describe the shape of the letter to the extent necessary for unique identification.

Two sequences of symbols are typed into the turnout terminal 32 without interruption via the keyboard shown in FIG. 2, completion of the entry of a character being indicated by hitting the space bar.

The coding should be sufficiently unique to identify individual kanji, easy to learn, psychologically comfortable, and efficient to use, yet allow for some variation in phonetic coding due to pronunciation in various dialects. The coding rules for creating sequences of keystrokes that identify individual characters are shown below to enable typing of Chinese characters in a simple manner.

Coding Rules Each character is encoded by a keystroke sequence using only phonetic symbols.

All characters are divided into five categories for coding.

That is, there are five categories: (1) general, (2) special, (3) exceptional, (4) arbitrary, and (5) neighboring pronunciation.

(1) General Category Most characters are in this category and are encoded by typing two connected sequences of keystrokes.

The first sequence is the standard pronunciation of letters without tone marks, and the second is the coding of letters using phonetic symbols only.

(a)' All characters other than those belonging to special or exceptional categories are encoded according to the following rules in the following order of steps:

Step 10 The pronunciation of the characters is coded using phonetic symbols without tone marks and using standard Mandarin Chinese.

Step 2゜ Immediately following the phonetic encoding of Step 1, the stroke shape of the letter is (
i) Upper left corner, (11) Upper right corner, (n+) Lower left corner, GV)
Lower right corner, coded using only phonetic symbols in the order.

Step 3. Encoding of a given character is completed by using the space bar.

This indicates that the encoding of the given character is complete and that the encoding of the next character to be typed will begin with the next keystroke.

(b) After the audio sequence is encoded, the stroke or radical shape is encoded using the following rules.

(1) If a stroke or combination of strokes is similar to the phonetic characters used on the keyboard, these phonetic characters are used as coding in the shape sequence (an example is shown in Table 3).

(1)) In other cases, the first phonetic symbol (consonant) or the last phonetic symbol (consonant) of the pronunciation of the name of the stroke or stroke combination.
vowels) are used (examples are shown in Table 4).

In Tables 3-1 to 3-2 and Tables 4-1 to 4-4, the standard stroke is followed by a sequence of derived or modified or related stroke shapes, thereby forming a subset of the system; A typist can efficiently type codes without having to analyze in detail the structure of the characters for which the identifying sequence is typed.

(e) To achieve maximum efficiency, priorities in encoding are established as follows.

(1) The encoding of complex combinations of strokes has priority over the encoding of less complex combinations, and the encoding of single strokes has the lowest priority in encoding.

(it) Stroke combinations that occur in characters other than characters containing radicals are coded only once in shape coding.

If a stroke or combination of strokes is encoded and occurs again in a currently geometrically encoded corner, it is not used again.

That is, the currently coded corner is omitted if it is a duplicate.

If it occurs again in another corner, it is also omitted.

For example: (d) Except for symmetrical strokes with the indicated coding, all or most symmetrical characters or portions of characters shall be treated as follows:

(1) For symmetrical letters of the type where the left and right sides of the letter are the same (but perhaps slightly different depending on the style used for handwriting) and the central portion is different, only the central core (i.e., the asymmetric core) coded.

The general form of these Q characters is merikaku.

For example:

(!1) If a symmetrical character has a central nucleus that is inseparable from the bottom corner, then only the central nucleus is encoded.

Namely 1. If the shape is Q, only the convex portion is used for encoding.

This rule can also be applied to text parts.

For example: (it) If the character has a symmetrical top, only the upper left corner of the central portion will be encoded instead of encoding both upper corners.

Only one stroke for the top is used for encoding.

In other words, if the letter looks like a U, then () Only one keystroke in the upper left corner of is used.

This rule is also used to encode parts of characters.

For example:

(iV) All symmetrical characters without a central nucleus are considered to be ordinary characters that are not encoded by symmetry rules.

For example:

M A symmetrical character with an identical character that is not placed in a corner to be encoded is considered to be an ordinary character that is not encoded according to the rules of symmetry.

For example: (e) If the entire character has a box-shaped border, the box is used for encoding, and the lower corner of the boxed part is used as the lower corner for encoding.

For example: Note that if part of the text is in a box, the box can be thought of as a combination of strokes, not as a whole.

For example:

The maximum number of strokes to encode a general two-character character is six.

Thus, if three phonetic symbols are used to code an utterance, only the phonetic symbols for the first three corners are used to code the formation.

However, if the typist encodes all four corners, the sixth stroke will be ignored by the system.

(2) Special Category (a) Characters encoded by a single keystroke The five most frequently used characters are encoded by a single keystroke to maximize typing speed.

The encoding is partly phonetic and partly geometric, as shown in Table 5.

Note that when phonetic encoding is used here, only the phonetic symbol of the first phoneme is used.

The five characters listed in Table 5 must be encoded as shown.

(b) Frequently used letters that are coded phonetically only Letters representing personal pronouns are coded phonetically without the use of tone marks as shown in Table 6.

(3) Exceptional Categories In order to avoid uniqueness problems caused by inconsistencies in encoding when the general rule is used, a small number of rarely used exceptional characters are:

(1) Speech encoding is performed normally.

(i+)) - Shapes are encoded by using the pronunciation of the radical name spelled phonetically, without the -n mark, and in a few exceptional cases as shown in Table 7.

(4) Arbitrary Range In addition to using normal encoding, this machine can encode the most frequently used characters as shown in Table 8 using voice encoding without specifying tone marks. .

The characters in this table are arranged in order of frequency of use.

These characters are selected from the 300 most frequently used characters.

Notes on Arbitrary Categories All characters are properly indexed using the rules for encoding listed in General, Special, and Exceptional Categories.

Arbitrary and neighboring pronunciation ranges are additional codings provided to perform index shifting at the typist's convenience and optionally at high speed.

As the typist becomes familiar with the machine, the typing speed of the typist increases progressively as it increases the use of shorter arbitrary codings.

The shortest coding list is shown in Tables 91 to 9-
This is listed in number 2.

The letters in this table are listed according to the standard order of initial phonetic symbols with the required standard pronunciation.

(5) The standard pronunciation of Chinese characters used for phonetic encoding of neighboring pronunciation range characters is that of Mandarin Chinese.

However, many Chinese people are accustomed to speaking their dialects and do not usually use the full pronunciation, even if they know the standard pronunciation.

In order to make coding more convenient and psychologically comfortable for more people, a certain degree of inaccuracy is allowed in the phonetic encoding of letters.

Although standard pronunciation is required for the use of coding in any category, some variations in the pronunciation of letters in the general and exception categories are possible (as shown in Table 10).

Master coding list mask coding list is Table 11-1 or 11-2
5.

This is the coding that occurs when the coding rules described above are used in conjunction with standard pronunciation to encode basic Kanji that can be typed by the system.

It should be referred to by typists when coding questions arise, but should not be construed as complete.

This is because the list does not include any short coding or neighborhood speech coding, both of which are provided in the translator logic in computer 30 for the convenience of the typist.

It is clear that the lists in Tables 11-1 through 11-25 can be made considerably longer without changing the rules for encoding keystroke sequences in typing, the logic used for processing in translators, or the basic design of the system. It is.

Prepared random coding is not necessary but allows the typist to increase typing speed as he or she becomes familiar and comfortable with the task.

The typewriter or typesetting machine keyboard 34 should be designed for efficient and comfortable use.

A good design for an efficient and comfortable keyboard is shown in FIG. 2, and is based on the consideration of the adjusted frequency of the written Northeast Chinese language coded according to the above rules.

The phonemes are shown arranged on a standard English keyboard.

Space bars, rarely used punctuation, and special function keys are not shown.

Translator The preferred embodiment of the present invention is shown to include a general purpose computer.

This is because many large companies already have computers for other purposes, and their typing needs can be met relatively inexpensively through the use of the computer programs described below. It is.

A completely self-contained system can be created if desired.

Next, a Chinese kanji search system according to a preferred embodiment of the present invention will be described.

The input from the keyboard 34 on the input terminal 32 to the translator of the computer 30 is a coded sequence of keystrokes typed to identify the characters to be printed, each keystroke being coded as shown in Table 13. Ru.

Standard coding, such as the standard ABS11 code for currently available keyboards, can be used in place of those shown in Table 13, provided that the changes required to match the key coding do not accept keystroke data as input. shall be applied to the program that takes the program.

The sequence of keystrokes typed forms a series of numbers that are processed arithmetically to the extent possible one keystroke at a time.

The typed keystroke sequence is sent to the translator's buffer memory.

Within the translator, the sequence is used to identify the location or memory location where the typing instructions are stored for the identified character.

To avoid searching through the entire word list to find a sequence that matches what was typed, the list is divided into 6 according to the number of keystrokes required to type the entire code for a single character. It is divided into major word groups.

The valley major group is divided into 37 subsets according to the initial phoneme (ie, first keystroke).

Some of the subsets are empty because not all phonemes necessarily occur in the first position of a major group.

The number of keystrokes used to encode a given character is determined by counting the keystrokes until the space bar is pressed to indicate completion of encoding.

This number is the group number to which the character belongs.

The search begins within the group at the phoneme subject of the first keystroke.

The typed keystroke sequence is compared to each character in the subset.

To improve search efficiency, the sequences stored in each subset are arranged in descending order of character usage.

The number of comparisons made in the search until a match is found is used to identify the sequence number of characters in the major group that contains the searched subset.

This sequence number indicates the location of the memory cell in which the X-Y index of the character's location in the printer 36 (corresponding print address bits described below) is stored, and these indexes are then sent to the printer.

If no correspondence is found between the typed and memorized keystroke sequences, the possibility of neighboring pronunciations is tested.

If other possibilities exist, further searches are made for comparison.

If nothing is found, either there was a typing error or the identified character is not included in the printer, which may alarm the typist.

Next, the search system according to this embodiment will be explained with reference to the flowcharts shown in FIGS. 3 to 8.

This program uses punch card input.

However, in actual use of a general purpose computer, direct data entry from the typewriter console 32 can be used instead of card input.

FIG. 3: Loading and Storing (a) This reads and stores a list of used symbols, the maximum number of which is 61 when using the IBM Key Bunch.

Includes the symbols listed in Table 13.

(b) There are 6 groups in the master list.

N1 is the number of letters and punctuation marks in a group that can be encoded using only one keystroke.

(e) This is type font
), the single keystroke sequence used to identify the member, and the sequence number of the member in the main list.

The numbers IL1(1,I) and ILI(2,I) are the X and Y indices, respectively, of the location in the type font of the first character of the first major group.

AI (I ) is the coding of the first member of the first major group.

ILI(3,I) is the sequence number of the character in the mask list.

(d) J is an index number that defines the major group pick-up.

J is from 2 to 6. (e) This is group 1
(b) for group 2.

(f) To speed up the search, the search is performed in a specific subset identified by the initial phoneme of the coding and in a specific major group identified by the number of keystrokes (number of phonemes) used in the coding. It is done within.

The location of this subset is determined by specifying the first and last sequence numbers of the major group characters included in the subset.

The sequence number of the first character of the second major arc is identified by the number 15(K, 2).

Here, K is a phoneme sequence number and 2 is a major group pick-up.

For example, if K=5, the search is performed using sequence number l5.
Starts at (5,2) and ends at (IS(6,2)-1).

Thirty-eight numbers are required for each major group since the difference is taken to determine the range to be searched.

If the difference l5(K+1.2)-IS(K,2)=O, then the Kth subset is empty.

(g) This reads into memory the coding keystrokes, printer location index, and master list sequence numbers.

The master sequence number is not necessary for the functionality of the translator program.

They are included here only for experimentation and checking the accuracy of the mechanical construction.

In the definition of A2(I,K), A represents the alphanumeric, 2 represents the second major group, ■ is the keystroke number (in group 2,
I=1 represents the first stroke, I=2 the second stroke), and K represents the sequence number of the character in the second major group.

In the definition of IL2 (I, K), ■ represents an index,
L represents location, 2 represents major group 2, I-1 refers to the X index, I=2 refers to the y index, I-
3 indicates the sequence number of the master list.

K is defined relative to A in the preamble.

(h) This repeatedly reads data into memory for the valley major group.

Terms (e3) for various major groups.

The explanations for (e4), (e5) and (e6) are similar to those for (e) made for major group 2.

Terms (fi) and (gi) are treated similarly.

(i) TKYI (I), TKY2 (I), and TKY
3(I) are the possible changes that can be made to the typed keystroke sequence according to the allowed variants in the encoding.

Please refer to the neighboring pronunciation categories in the encoding rules mentioned above.

In the specification TKYI(I), ■ is the phoneme sequence number, and number 1 means the first keystroke typed.

This 1 can be 2 or 3, in which case it refers to the second or third keystroke typed.

Phonemes that cannot be replaced according to Table 10 are defined as themselves in this step.

(ie, the transformation TKYI(I)), etc. is the identity transformation in this case.

FIG. 4: Manual Entry of Keystroke Coding Message (a) KEY(1) is the entry for the first keystroke typed.

Control number KKH=0 means that no neighboring pronunciation alternatives were searched.

(b) The machine identifies the set to which the keystroke just typed belongs.

The symbols are defined as follows.

That is, E means belonging to..., PH means a set of phonemes, CR means a correction key, CL means a clear key, SP means a set of punctuation marks, S
F means special function key.

See Table 13. Because a single keystroke encoding includes not only the phonemes used to encode letters, but also the coding for punctuation and special function keys.
Treated separately from multi-stroke coding sequences.

(c) Special functions include modify, clear, backspace,
This includes jumping spaces, line breaks, repeating characters, etc.

See Table 13. (d) An alert is given if the keystroke sequence typed for this point is theoretically impossible.

(e) If the keystroke typed is a spacebar, the encoding of the character is complete and the completed keystroke sequence goes to the searched part of the program.

For one keystroke barger group,
This is [phase].

The main features described above for the first keystroke are applied to the second keystroke through the eighth keystroke as shown in the flowchart.

The maximum number of keystrokes is 8, but a maximum of 7 keystrokes including the space bar is all that is required.

If the 8th keystroke is a spacebar, the 7th keystroke is replaced by a spacebar stroke before the search is performed.

O indicates how to direct the search to one of major groups 2-6 according to the "calculated go-to" statement.

Recall that the search for group 1 is directed to O of (Top Wide).

)(f) ID(KEY(1)) identifies the subset sequence number by sequence number of the phoneme of the first keystroke.

(g) This defines the range to be searched in the selected barger group.

(h) This defines the initial value of I.

This number is the order number of the first member of the subset searched in the burger group of interest.

(i) The rcomputed go-to statement directs the search to the verger group identified by II.

Figure 6: Search and print The various burger groups are similar to each other.

The search proceeds by comparing the coding sequences stored in memory and the keystroke sequences typed for the valley characters in the master list, one keystroke at a time.

To improve the efficiency of the search for verger groups 2 through 6, the search is performed by searching for verger groups corresponding to the number of keystrokes used to encode a character only within the subset identified by the first keystroke of the sequence. It is done in a group.

Comparisons within this subset are made from the second keystroke.

Explanation is given for Berger Group 3 only.

Entries are identified by O.

Here the third digit, i.e. 3, is Berger Group 3
, which is a 3 keystroke group.

(a) The subset being searched has been selected using the first keystroke.

The comparison begins on the second keystroke.

KEY(2)=A3(2,1) compares the second keystroke of the coding stored in memory and the second stroke typed for the ■th member of the third mager group.

Note that the ■th member is in the correct subset.

If T (true), ie, if the second keystroke matches the code, then the third keystroke is compared.

If F (false), ie if the second keystroke does not match, then the coding of the second keystroke of the next character in the stored sequence is compared.

(b) The third stroke is compared.

(e) If all keystrokes typed match those stored, the typed character is identified as the ■th member in major group 3;

Location index I for identifying the location of characters in the printer 36
L3(1,I) and IL3(2,I) are sent to the printer and the characters are printed.

The input of the keystroke sequence for the next character begins.

See Figure 5.

(d) If the current comparison is not possible, increment the index number ■ by 1 and allow the coding of the next member of the major group to be selected for comparison.

(e) If the increased index belongs to the range being searched, a new comparison is made.

If the incremented index number exceeds the range being searched, then the entire list has been examined and no characters were found for the typed keystroke sequence.

In this case, the typed keystroke sequence is for a character that is not included in the type font, or there is an error in the typing.

If group number II is greater than or equal to 3, neighbor pronunciation is possible and the search is directed to O.

FIG. 8: Search using neighboring pronunciations (a) KKH=KKH+1° This control number indicates the number of trial searches using various neighboring voices.

(b) Branching is done according to the KKH control number.

(c) KKH=1oIsKEY(1) TKYl (KEY)(1))? This test is to see if the neighboring voice substitute for keystroke 1 is itself.

If so, this keystroke change cannot be made and returns to O.

If an alternative is possible for keystroke 1, go to (d).

(d) Keystroke 1 is replaced by its possible alternatives.

O is the start of the search in FIG. (e) KKH=2° This asks if the number of keystrokes typed is less than 4.

If so, the keystrokes cannot be changed.

■ Gives a warning. If the keystroke is greater than three, alternative possibilities for the second keystroke are examined.

(f) This asks if there are any possible alternatives to keystroke 2.

If there is a nearby voice replacement, no keystroke changes can be made and the search returns to O.

If a substitution is possible, a substitution is made.

(g) Keystroke 2 is replaced with its possible alternatives and the search is directed to O again.

(h) KKH=3o This switches keystroke 1 to its original stroke and leaves the second keystroke in its modified form.

Another search is performed.

(i) KKH=4° First, check the number of keystrokes.

If it is less than 5, an alarm will be raised.

If it is a dog than 4, the possibility of changing the third keystroke while leaving the first two unchanged is investigated.

(j) This asks whether keystrokes can be changed, similar to ((C) and (f)).

If possible, do so.

(k) This restores keystroke 2 to its original stroke, but changes keystroke 3 to its replacement,
Return to Exploration O.

(m) KKH=5o This asks if possible alternatives exist for the first stroke.

If it does not exist, an alarm will be raised.

If it exists, continue.

In this case, we investigate the possibility of changing the first and third keystrokes while leaving the second keystroke in the state it was originally typed in.

(n) Possible alternatives to keystroke 1 are used and the search is updated.

A Fortran program based on the flowcharts of Figures 3-8 using an IBM punch card as an input medium is attached at the end.

Input could be directed directly from a typewriter console, or from magnetic tape previously encoded by typing keystroke sequences, or from perforated paper tape made with a flexwriter or similar machine.

Those skilled in the art will be able to design various search devices that can be used in the system of the present invention from the flowcharts in FIGS. 3 to 8 and the programs at the end.

FIG. 14 is a block diagram showing an example of a search device that can be used in the search system of the embodiment described above.

Keyboard input device 34 converts valley keystrokes into corresponding 8-bit digital signals and sends a series of digital signals for incoming keystrokes to keystroke buffer memory 40.

Buffer memory 40 may consist of conventional registers and may have a capacity of 56 bits for seven keystrokes.

The output of buffer memory 40 is supplied via buses 48, 50, and 52 to keystroke number discriminator 42, last keystroke detector 44, and comparator 46, respectively.

The last keystroke detector 44 detects the completion of the input operation of the Chinese kanji by detecting the space bar stroke and provides a search enable signal on line 54.

This enable signal is provided to the set input of flip-flop 56 to place flip-flop 56 in the set state.

If the first keystroke is a caspase bar stroke, this is interpreted as a blank input and the detector 4
4 immediately generates a blank command signal on line 58 to a printing device, such as printer 36, which will be described below.

This command signal is also connected to line 58, OR gate 60.
, through line 62 to keystroke buffer memory 4
0 reset input to reset the memory 40.

The keystroke count discriminator 42 determines the total number of keystrokes (excluding the last space bar stroke) based on the number of bits of the input digital signal from the buffer memory 40, and determines the total number of keystrokes (excluding the last space bar stroke), and determines the total number of keystrokes (excluding the last space bar stroke) on the six output lines G0. G2...
...Gives an on signal to one Gj corresponding to G6.

This ON signal is sent to the memory 64 as a group instruction signal to make the corresponding group area in the memory 64 accessible.

For example, if the number of keystrokes is 6, a group 6 instruction signal is generated on line G6.

The upper eight bits of the buffer memory 40 (that is, the digital signal corresponding to the first keystroke) are supplied to the first keystroke decoder 57.

The decoder 57 demodulates this 8-bit signal and outputs 37 output lines S0. An on signal (subset instruction signal) is given to a corresponding one S. of S2...S3□.

This subset indication signal serves to make the corresponding subset area within memory 64 accessible.

The output signal of flip-flop 56 being set enables AND gate 66 to provide a clock pulse from clock circuit 68 to memory address index register 0.

Index register 0 outputs an address signal incremented by one in response to each clock pulse and is provided to memory 64 via bus 71.

This address signal specifies a specific memory area within the memory 64,
In other words, the memory area Mij specified by the group instruction signal Gj from the keystroke number discriminator 42 and the subset instruction signal Si from the first keystroke decoder 57 is searched, and as a result, the Chinese kanji characters allocated to the memory area Mij are searched. The coded bits are read out one after the other.

In this case, the corresponding print address bits (which correspond to the aforementioned ILi (1, I) and ILi (2, I)) are read out together with the respective coded bits.

The read coded bits are sent to comparator 4 via bus 74.
6 and are sequentially compared with the input digital signals sent from the keystroke buffer memory 40.

At the same time, the read corresponding print address bits are sent to print buffer memory 72 via bus 76 and held in order.

The coded bits may be set to 40 bits long and the corresponding print address bits may be set to 14 pits long, in which case only a portion of the 40 bits (e.g. 24 pits for group 4) are valid for groups other than noger group 6. Becomes a bit.

In the comparator 46, the coded bits are compared with the input bit (digital) signal one after another in the order of the sequence numbers mentioned above, and when any coded bit matches the input bit signal, the output line 78 of the comparator 46 matches. A signal is generated and this match signal is sent via line 80 to print buffer memory 72 to send the corresponding print address bits held therein to printer 36 via bus 82.

The printer 36 prints desired Chinese characters using operations as described below.

'The match signal from comparator 46 is also connected to line 80, OR
Index register 70 via gate 84 and line 86
is applied to the reset input of flip-flop 56 via line 88 to reset register 70 and is also applied via line 88 to the reset input of flip-flop 56.
Reset.

This closes the AND gate 66 and cuts off the supply of clock pulses.

Additionally, this match signal is passed through OR gate 90, line 92, O
Provided via R gate 60 and line 62 to the reset input of keystroke buffer memory 40 to reset buffer memory 40.

In FIG. 14, bus 94 is for supplying digital signals corresponding to various keystrokes other than phoneme keystrokes and space bar strokes to printer 36.

The file end error alarm 96 also indicates the search range (Mi
j), an alarm is generated in response to the end-of-count pulse / from index register 70, informing the operator that the typed character was not found.

The end-of-count pulse resets index register 70 and flip-flop 56 via OR gate 84, line 86, and line 88, respectively, and resets index register 70 and flip-flop 56 via OR gate 84, line 86, and line 88, respectively;
, through line 62 to keystroke buffer memory 4
Reset to 0.

As mentioned above, in order to improve search efficiency, the coded bits stored in the valley memory area may be arranged in descending order of the frequency of use of the corresponding character.

FIG. 15 shows another search device.

In this device, 37 ROMs corresponding to subsets are arranged, each ROM containing, for example, 1024 54-bit words (40 bits for the coded bits and 14 bits for the corresponding print address bit).

Although not shown, the above-mentioned keystroke number detector 42 is provided to detect R corresponding to a specific major group.
It may be configured to address OM groups.

Note that in FIG. 15, the 37 ROM groups are actually connected in parallel.

The input from printer computer 30 to printer 36 is the binary encoded X-y location of the character to be printed.

This coding is used to position a rectangular, flat, tubular or belted type string so that it is located below the location on the paper where the selected character is to be typed.

A printing mechanism then strikes the type, which prints an image by transferring ink from the ribbon to the paper in the usual manner.

Additionally, U.S. Patent No. 382,064 converts digitized characters into 120 hexadecimal digit code units.
A printer such as that shown in No. 4 can be used.

Examples of printers that can be used with the present system are shown in FIGS. 10-13.

The output printer 36 shown in FIGS. 9-13 is integrated with an input terminal 32 and a keyboard 34.

A conventional moving electric carriage 38 includes paper rollers 40,
It includes a spacer lever 42 and a conventional advance knob 44.

Control of carriage position is accomplished by conventional control keys on keyboard 34.

The paper roller 40 is attached to the support shaft 4 as shown in FIGS. 11 and 12.
6 of conventional rubber construction.

A typewriter ribbon 48 passes under paper rollers 40 to enable characters to be imprinted on a sheet of paper 50 being transported by carriage 38 in the manner of a conventional typewriter.

Typing head assembly 52 provides the desired ideograms;
An angularly spaced row of ideograph image blocks 56 locate the ideograph below the typing location at the intersection of the embedded tubular typing head 54 and the paper roller 40.

A hollow, stationary support shaft 58 allows the tubular typing head 54 to translate and rotate under the paper roller 40 (as shown in FIGS. 10, 11, and 12).

The cylindrical typing head 54 is formed from a relatively flexible elastomeric material, such as polyvinyl or silicone rubber, and is mounted as shown in FIG. 11 to provide a support surface for sliding and rotating back and forth on a stationary shaft 54. It is attached at both ends to a support flange 60 having holes 62.

A driven gear 64 is secured to the support flange 60 and positioned by a drive gear 66 mounted on a rotatable shaft 68 .

A longitudinal slot 70 in the outer surface of the rotatable shaft allows the splines 72 of the drive gear 66, as shown in FIG. 11, to translate axially along the shaft while allowing the drive gear to rotate with the shaft.

Translation of the drive gear along the rotatable shaft 68 is effected by a carrier 74 fitted with clips 76 at each end, the carrier 74 extending along opposite sides of the drive gear to translate the drive gear as the carrier translates. It has an arm 78.

A conventional x-y position indicator 80 rotates the rotatable shaft 68 to a desired position in accordance with a binary input (corresponding printed address bits) coming from the computer via a retractable wheel 82 and as shown in FIGS. Carrier 74 is translated with continuous flexible ribbon 84 secured to carrier 74 by set screw 86 shown.

Flexible ribbon 84 is translated to the desired position by binary input from computer 30.

As shown in FIG. 13, the ideogram block is generally square and has on its upper surface the type 88 of the desired ideogram that can be positioned by rotation and translation of the tubular typing spatula 1540 beneath the paper 50 on which the image is desired. .

The ideographic block is held in the tubular typing head by friction or adhesive and is pushed onto the typing ribbon 48 by plunger 90 when solenoid 92 is actuated after the desired type has been positioned, as shown in FIG.

The flexibility of the tubular typing head 54 allows individual type to be pressed against the paper without striking adjacent types into the paper.

Ideographic blocks can be formed from metallized plastic or light metal type to reduce system inertia and increase machine typing speed.

The x--y position indicator 80 is connected to the
2 binary pits and drive the rotatable shaft 68 to the desired angular position and the continuous flexible ribbon 84 to the desired axial position through appropriate gearing, thereby causing the plunger 90 to Two binary digit locators may be included to position the desired ideogram at the typing position by striking the back of the ideogram block and pressing it against the paper 50.

Operation of the printer involves placing the paper sheet 50 in the carriage 38 in its normal position.

However, printing occurs at the bottom of the roller rather than at the front of the roller as in conventional typewriters.

A sequence of keystrokes is typed onto the keyboard 34 to uniquely identify the desired ideogram as described above.

Once this is done, computer 30 provides a binary signal to XY position indicator 80 which rotates rotatable shaft 68 and translates flexible ribbon 84 to the column of type containing the desired character. The carrier 74, and thus the tubular type head, is moved to position the type containing the desired characters into a typing position above the plunger 90.

Solenoid 92 is then actuated to force the desired ideographic block 56 onto ribbon 48, thereby printing the desired character on paper 50.

The letters are typed from left to right on their sides, so when the paper is removed, the letters can be read from top to bottom and from right to left.

As is clear from the above detailed description, according to the present invention, it is possible to immediately search and print a desired Chinese character by continuous keystroke operation, and it is possible to immediately search for and print a desired Chinese character by continuous keystroke operation, and to avoid the need for intermittent keystrokes or frequent keystrokes in conventional systems. There will be no more man-machine dialogue, and search speed will be significantly improved.

Claims (1)

[Scope of Claims] 1. The combination of a first phonetic symbol at least partially representing nine predetermined pronunciations of one Chinese Chinese character and an additional second phonetic symbol representing a descriptive feature of the Chinese character In a search system that pre-codes Chinese characters and identifies and prints the desired Chinese character from a coding list of all available Chinese characters, the number of phonemes used to code the corresponding coded bits for available Chinese characters. assigning to a particular one of a predetermined number of coding groups according to the initial phoneme of the coding and to a particular one of a predetermined number of coding subsets according to the initial phoneme of the coding, and storing it together with the corresponding print address bits; sequentially inputting first and second phoneme sequences forming said first and second phonetic symbols, respectively, representing Chinese Chinese characters, converting valley input phonemes into predetermined corresponding bit signals, thereby generating a series of bit signals corresponding to the first and second phoneme sequences; identifying the number of phonemes of the first and second phoneme sequences based on the number of bits of the series of bit signals; the first indicating the corresponding specific coding group;
identifying a leading phoneme of the first and second phoneme sequences based on a leading bit signal of the series of bit signals, and instructing a particular coding subset corresponding to the leading phoneme; generating a second indication signal; in response to the first and second indication signals, the coded bits assigned to the particular coding group and assigned to the particular coding subset are transferred to the corresponding printed address; reading the bits one after the other; comparing the read coded bits one after the other with the series of bit signals, while at the same time retaining the corresponding print address bits read together one after the other; , generating a comparison match signal when the comparison Q result matches the read coded bit and the series of bit signals; and generating the correspondence corresponding to the matched coded bit in response to the comparison match signal. The present invention is characterized by comprising the steps of transferring print address bits to a printing device, and addressing and printing out the desired Chinese character in the printing device using the transferred corresponding print address bits. A method for identifying and printing Chinese characters. 2. Claims characterized in that the coded bits assigned to each of the coding groups and coding subsets are arranged and stored in order of frequency of use, and the coded bits are sequentially read out according to the arrangement order. The method described in paragraph 1.