FR2482747A1 - DEVICE FOR ENCODING IDEOGRAPHIC CHARACTERS - Google Patents

Abstract

ENCODING DEVICE INCLUDING A KEYBOARD WITH CODING KEYS 10, DISPLAY MEANS 12 AND A MASS MEMORY 11. IN RESPONSE TO THE ACTION OF THE KEYS AFFECTED TO DETERMINED PHONETIC ELEMENTS, THE CONTENT OF A MEMORY PAGE APPEARS . THIS CONTENT IS CONSTITUTED BY THE REPRESENTATION OF SEVERAL IDEOGRAPHIC CHARACTERS HAVING THE SAID PHONETIC ELEMENTS. ADDITIONAL KEYS ON THE KEYBOARD ALLOW YOU TO SELECT ONE OF THE CHARACTERS DISPLAYED AND TO PROVOKE THE EMISSION OF A REPRESENTATIVE CODE OF THIS CHARACTER.

Description

Device for encoding ideographic characters.

  The subject of the present invention is a device for

  encoding ideographic characters and she finds an ap-

  particularly important in the fields of text-typing and telex transmission.

  To allow a better understanding of the

  and the problems it solves, it can be useful

  to recall some indications on the languages of ideo-

  grams, taking Chinese as an example, which must be

  remember that it is spoken by a quarter of humanity.

  Chinese is a language with mono-syllable ideograms

  which there are several phonetic representations

  approximations (pin-ying transcripts, Wade, École Française

  of the Far East), all of which have the disadvantage

  deny not to account for homonyms, while the ideograms distinguish them perfectly. There are 412 different sounds in Chinese, which can be pronounced in

  4 tones, which makes about 1500 different sounds.

  However, there are about 49,000 Chinese characters, of which 5 to 8,000 are of fairly common use, which translates into an average of 20 homonyms per tone; but some of the characters are little used and it has been found that the 2,000 most frequent characters cover about 95%

current texts.

  Various devices are used or at least have been made at the moment to encode Chinese characters. A first device is purely mechanical. Tables are placed with 2,600 characters each

  a typewriter, one at a time, and they are moved manually.

  until an ergot can select the

  desired character and print it. A second device, de-

  Written in 'tComputers and People', Vol 27, May 1978, is electromechanical and includes a scroll with 4,356 written characters.

  until the chosen character is in front of a cur-

  sor. The operator then activates a pedal which causes the sending of a code, determined by the position of the roll,

to the printing device.

  Another approach is used for telex transmission. The operator has a collection in which the 000 most frequent characters are assigned a numerical code consisting of the digits from 0000 to 9999. The operator searches for the code corresponding to the characters

in the book and telex the figure.

  This last approach, like the first one, is very

  slow and requires a long apprenticeship.

  The electromechanical device, which is not

  not yet proven, only allows a gain in limited speed.

  We finally proposed (IEEE transactions on systems ...

  SMC3 No 3, May 1973) an encoding system that involves

  decompose each character into components and subcomponents

  juxtaposed, the first approach being all-

  time a phonetic decomposition of the character between a

  initial consonant and a final vowel. The keyboard used

  is a classic keyboard typewriter.

  The Japanese language, also with ideographs, has a

  structure and a slightly different writing. She

  carries 140 syllabic characters, grouped into 2 genera, "cur-

  siVe "and" square ", representing the sounds of 70 syllables of

  two different ways: it also includes features such as

  only 1650 since 1945,

but no tones.

  The problems encountered in encoding the characters

  However, Japanese styles are essentially the same as in the case of Chinese. We still find them in

the case of Korean.

  The present invention aims to provide an encoding device that responds better than those previously known

  the requirements of the practice, in particular

  puts a fast rate encoding, in that it is efficient, that is to say requires few bits to represent a character, in that it is very flexible in that it is suitable for encoding any number of characters up to a maximum of 49000 and in that it is readily achievable from commercially available components, at an acceptable price although naturally higher

  than that of purely mechanical machines.

  The invention also aims to provide a device for particularly easy encoding,

  so that a sinophone even beginner can use the

  positive after a few minutes of explanations, and to learn

  automatic weaving (the repeated use of the machine rapidly increasing the encoding speed)

  For this, the invention uses a phonetic approach

  that, no longer semantic or based on the form of

  This approach makes use of the fact that each of the monosyllabic sounds forming the Chinese character set can be regarded as composed of an initial, a final or a termination, and a tone. For example, in the dialect

  Mandarin, which is the official language in the People's Republic of

  In China and Taiwan, there are 21 initials, 35 finals and 4 tones that can be combined to form the syllables represented by characters, respecting certain rules. There are indeed characters whose sound is only a termination and two or three sounds that

  can not be pronounced with an initial.

  For this purpose, the invention proposes in particular a

  positive encoding type including a keypad

  each assigned an address in the

  moire, a device characterized in particular by

  gate means associated with the mass memory and the keyboard so as to show, in response to

  the operation of keys assigned to specific components

  the contents of a memory page constituted by the

  representation of several characters having the said ele-

  specified phonetic elements and keys, which can be operated manually, to select one of the displayed characters

  to cause the issuance of a code representative of the

selected.

  The code is advantageously constituted by the address in memory, broken down into the number of the page of memory and number of the character in the page, although other codes

  possible, for example, the telegraphic code

  but the translation into a language other than Chinese of the meaning of the character or any combination. The code will usually be provided in electrical form, but it can also be

  well be optical, acoustic or mechanical.

  An encoding device according to another aspect of the invention, of the above-defined type, is characterized in that the keys for coding the keyboard are provided to allow each of the initials to be selected and

  each of the finals of character and in that the

  tif further comprises display mdfs associated with the mass memory and the keyboard so as to display, in response to the operation of coding keys, the

  content of a page of memory in which is located the

  character or the most frequent characters defined by the initial, the final and possibly the tone defined using the keys, the keyboard having manually operated selection keys to select one of the displayed characters and validate it so as to provoke

  issuing a code representative of the selection character

  born. The approach thus used is very natural because it uses an old and widespread method for

  explain the pronunciation of a character, named "fan-qie".

  This method consists of taking two characters of sounds

  known, and to point out that the character of pronounced

  unknown at the beginning of one and the end of the other.

  If, for example, the character to be explained has the pronunciation "mu, (fourth tone)", two very strong characters are used.

  known and frequent, for example "my first tone" and "zhu.

  fourth tone ", - and we say" ma-zhu-mu ".

  Keyboard coding keys will bear characters, preferably very common, which will exhaust all possible initials and finals or their transcription.

  The code may in any case be transmitted to

  tance or used locally, for example to operate a printer that will recognize the code and print the corresponding character. The decoding will in any case be very simple since it suffices to assign to each code 2 48 2 7 4 y digital a character, which can be done for example

  using a read-only memory constituting lexicon.

  The invention will be better understood on reading the

  following description of devices that constitute

  particular modes of implementation, given as examples

  non-limiting ples. The description refers to the drawings

  accompanying it, in which: - Figure 1 is a block diagram, showing the

  main components of the device.

  FIGS. 2 and 3 show a possible arrangement

  the keys of a device according to Figure 1,

  encoding Chinese characters, -the indications on the keys being in Latin characters on the

  Figure 2 and in ideograms in Figure 3.

  FIGS. 4 and 5 show a visualization mode

  possible ideograms on the display means of the

device of Figure 1.

  FIG. 6, similar to FIG. 2, shows a

  variant embodiment of the keyboard.

  FIG. 7 is a flowchart of principle

  the sequence of successive encoding operations.

  The block diagram in Figure 1 shows only

  the main constituents of the device and their

  links, without separating bus lines of data and address.

  The encoding device illustrated can be viewed as comprising an input keyboard 10, a data processing system whose essential element is a read-only memory 11 and display means 12 which will be

successively described.

  The keyboard input, including the origina-

  to use a phonetic approach, is provided with all

  codes or keys whose distribution may be that shown in FIG. 2 or 3. These keys comprise a set of initial or final keys 13, of which

  the number is sufficient to exhaust all the

  possible. On each key is written an ideo-

  gram, preferably among frequent, representative

  of the corresponding initial or final in the present

  "fan qie" (Figure 3) or its transcription in Latin characters (Figure 2) or other, for example Cyrillic. All the phonemes chosen must exist in the four tones and at least one of the four-tone meanings advantageously corresponds to frequent characters. It may especially be the phonemes ma, bo, ke, shi, cai, fei, dao, cabbage, tan, gen, li, jia, piao, niu, xian, qiang,

  rank, hong, zhu, sui, zuo, following the transcription pin-

  ying, worn by twenty-one keys. We see that the choice

  is made so that keys with different initials

  annuity bear characters which are also terminating

all different sounds.

  a set of keys 14 of phonemes used without

  tial, that is "er" and "weng", keys carrying the in-

  phoneme labeling in Latin characters (Figure 2) or ideograms (Figure 3); these phonemes are not used in all the tones: those which are missing are indicated in parentheses in figure 2; a set of keys 15, corresponding only to phonemes containing finals not appearing on the keys 13; for greater clarity, these keys 15 have been separated from the keys 13 by a dashed line in FIG. 2; these finals are still symbolized on the keys by frequent ideograms (Figure 3) or their

  pin-ying transcription. The keys may bear the symbol

  bolisation in the four tones, as shown in the figures, although it is not necessary. More generally, the keys 15 bear characters in the

  four shades and at least one of these meanings

to a frequent character

  four tone keys 161 ... 164 allowing

choose the appropriate tone.

  The keyboard furthermore has selection and printing keys 17, a key 18 whose actuation indicates that the character to be encoded does not include an initial, a

  error key 19 and a subsequent page call key

  te whose role will appear later.

  The mass memory is an organic read-only memory

  f -j2747 nized by pages, each page containing a determined number of ideograms encoded according to a matrix. Other methods of storing characters are possible mechanical, optical or storage of a set of programs for calculating matrices of points. Most characters

  can be satisfactorily represented on a machine.

  16 x 16. Figures 4 and 5 show respectively

  the representation, on such a matrix, of two ideograms

  my having the pronunciation "mu" one and the other.

  The memory It can notably be organized in pages of 256 bytes each comprising 8 characters each represented on a matrix 16 x 16 and each assigned a code consisting of the address of the page and the serial number of the

character in the page.

  The ideograms are distributed between the pages of

  in order to locate an ideogram determined by a

  approach that is statistically as short as possible.

  For this, the homonyms are grouped: all the characters corresponding to a given phoneme, for example "m", are

  grouped in successive pages of memories n, n + 1, ..

  whose number depends on the number of homonyms, the characters

  being arranged in order of decreasing frequency.

  The display means 12 may be either on a separate device 10 - (Fig. 1) or carried by it (Figs 2 and 3). They are intended to display

  simultaneously all the contents of a page of memory, that is,

  that is eight characters. In this case, this result -

  can be obtained with 4096 points in the case of matrices 16 x 16 taking into account the necessary intervals. The selection keys 17 make it possible to choose, from among the eight characters visible on the means 12, the one that is to be encoded; if this character does not appear, the support on the

  key 20 to display the next page-.

  It can be noticed that the actions on the key 20 do not cause an increase in the number of bits of the

  code, defining the page, but only an incremental

  the page number register.

  The encoding device is completed by a central processing unit 21 provided with its working memories 22 and 23, a printing or transmission unit 24

  of the code selected in mass memory 11 and two

  registers, an entry register 25 and a page register 26.

  It will be assumed in the following that the mass memory 11 comprises 512 pages of 256 bytes each. In this case, the page register 26 may comprise 9 page representation bits (part 27) and 3 character location bits in the page (part 28). The input register associated with the

  keyboard, may have 5 bits of initial representation of

  the, 6 bits of final representation and 2 bits of tone.

  The device is implemented according to the logic flow diagram of FIG. 7. Before any action

  on one key, the eight most frequent characters

  appear on the screen; if the character to be coded is one

  of these eight characters; the action on the key 17 corres-

  The other cases, it will be explained by assuming that the character to be encoded has the pronunciation "m" and the representation In this case, the operator successively presses the key 13 marked "my". "(Fig. 2) or (Fig. 3), then the

  key 13 marked "zhu" and finally on the key 164 (four

third tone).

  In other cases, the final will have to be searched for

  15. The 13-bit word thus generated is trans-

  put in the register 25 which presents it to the calculation unit 21.

  This determines the number of the first page where the pronunciation characters "m" are found and stores it in section 27 of the register on page 26. This page

  contains the most frequent representations in terms of

  Sound defects "m", as shown in Figure 2.

  In the example shown, the character searched for

  on the page, in fourth position: it is then sufficient to

  press key 17 marked "4" to enter the number

  fre 4 in binary in part 26 of the register on page 26 and send the corresponding code (address in memory 11)

to the device 24.

  The arrangement of the keys 17 corresponds to the arrangement

  characters on the screen: the operator does not have

  need to count the characters or keys to choose

  take the appropriate touch. After the action on one of the keys 17, the device is returned to the initial state, in which it is sufficient to actuate again a key 17 to encode

  one of the very frequent characters in the particular pae

  which is displayed at the beginning of the maneuver.

  Other situations may arise and correspond to

  different results of tests shown in 30 and 31 (Fig. 7) - the operator finds an error after display on 12: there may be no display because there are no characters in memory pronunciation

  defined by the initial, the final and the tone, the characteristics

  shown are obviously wrong, which reveals

  a keyboard manipulation error. In this case the operation

  controller presses the error key 19, which returns

  zero the registers and returns to the beginning of the sequence.

  - the searched character is not on the first page of memory. The operator then presses the next page key 20, which increments the register 27 of a unit and causes a new addressing. The calculation unit may be provided to display a particular message after calling the last corresponding memory page

to a definite pronunciation.

  We see that the approach adopted ensures speed

  maximum access to the searched character. If this character pos-

  few homonyms, it will appear on the screen as soon as

  first page, and the action on the "selection-impres-

  "print the character." If the character

  searched is an infrequent homonym and does not appear, the ac-

  pressing the "next page" key 20 causes the

  the following eight characters, until the

  The desired character appears and can be printed. About

  1960 characters out of the 4096 possible in this

  are visible on the first page. As will be the

  the most frequent characters, the use of the following pages

  will be infrequent since the 2000 characters

2482-747

  more common Chinese language include about

  % of those appearing in the texts.

  The embodiment shown in Figure 6 implements the same coding principle, but the keyboard does not include tone keys. In return, it has four separate sets of coding keys,

each game corresponding to a tone.

  The keyboard of Figure 6 (where the organs already

  in figure 2 bear the same reference number} com

  carries 256 character keys grouped into four sets of 64 keys with the same tone, designated 321V 32 2

  323 and 324. Only a few of the keys are

  felt for clarity. In each game, for example the game 323, a fraction of which is shown on a large scale in FIG. 6a, there are:

  - 21 keys 33 to represent initials

  tials and finals (similar to keys 13 in Figure 2),

  - 14 keys 34 corresponding to the finals

  not among the keys 33 (similar to the keys 15 in Figure 2): - The 4 x 29 = 116 additional keys are the most frequent characters that are not among

  the characters visible in the four coding zones

  (keys 36), Chinese numerical characters and

  Western, punctuation marks.

  In the embodiment shown, each

  It bears two characters. Access to the characters

  the top row is done with a touch of 37

  blable to the "capital letters" key of a typewriter

  or the "digital" key on telescrip-

  tors. The upper characters are chosen according to their

  frequency and proximity of their sound to the corresponding character

  laying on the bottom line. The action on each of these keys brings up a different page of memory

  on the screen, and another page by pressing the key 37.

  Finally, the keyboard has a standard spacing bar

  than 38, and possibly additional keys, by it

  example when the device has to perform other functions

* such as recording and billing, or where it is intended for word processing involving the

  on page, editing, getting started, correcting

reurs, etc ...

  This constitution of the keyboard will lead, -vantageuse-

  to place the same character in memory in several pages, each of the codes thus determined being valid, and to provide the calculation unit so as to show, from the action on a first key 33 to select

  an initial, the first page containing the character de-

  signed on the sidelines. It is possible that the

  sought character then appears immediately among the

  eight characters displayed on the display means.

  An advantageous distribution can be explained by referring to a very frequent character with many homonyms, such as "î11". The most common "1l" will then appear in the "normal" position. The action on this key will cause the visualization of the most frequent character, and-of seven other characters which will be the third, the fourth ... until the ninth in frequency: the second in frequency will appear when the keyboard is

  in the "capital" position. The action on this key in majus-

  this character with its procession of

  10th, 11th ... 16th characters in order of frequency

  this. With this arrangement, 512 characters or signs are visible to the operator on the keyboard and a little more

  two thousand on the first page. 400 characters suffice

  to print 73% of a Chinese text and 2000 for

  print 96.5%. After a certain apprenticeship he

  the vast majority of the characters in a tex-

  the current only requires the action on two keys (one coding, one printing), which leads to a speed

  very high typing rate compared to current systems.

  The keyboard can be supplemented with LEDs

  keys 36, which indicate to the operator who has

  searched, then selected a character by successive action-

  sive on two keys 33 and / or 34, and possibly one or more actions on the key 20, while this character

  among the keys 36, that he had an appro-

  It is faster by lighting indicators that are suitably arranged to indicate the approach that would have been the fastest. Thus the use of the encoding device is ergonomic, because it helps to memorize the positions of the characters. The use of the "next page" key 20

  decreases with use. In this provision, where

  same character can appear several times on several pages, each of the codes thus determined must be valid,

  which implies an appropriate structure of 24 (Fig. 1).

  Simplified and economical achievements not re-

  presented are also possible: for example, the

  It may be a group of initial keys and a separate group of final keys each bearing a phonetic code (transcription in Latin characters or other). The codification approach appears

perfectly on such a keyboard.

  The device is directly transposable to Japan.

  but in which the finals can be mono or disylla

  each Chinese character has two pronunciations

  one of which is monosyllabic or at most disyllabic.

  A simple and practical encoding device uses a keyboard with 70 coding keys, each of which has a squared syllabic character and additional keys for the numbers and typographic signs. Three keys indicate whether

  wish to write a square syllabic character, a character

  syllabic syllabic or a Chinese character.

  In the case of syllabic characters, the character is encoded by one-touch keystrokes. Each key on the keyboard also has a Chinese character, the subsequent action on the "Chinese character" key, and a keyboard key on the keyboard.

  the screen the page of memory in which appears the character

  wanted. Again we choose the character on the page

  by pressing the appropriate key of selection and

  The characters not appearing in clear on the

  are, as in the previous cases, coded by ac-

  following two coding keys which are

  the page of the memory containing the character whose initial and final correspond to the two keys. The

  Homonyms are treated as in the case of Chinese.

  But it only takes a much smaller memory,

  only 2048 characters, that is, for an encoding matrix of 16 x 16 (32 bytes), a

64 Kbytes is enough.

  It has so far been assumed that each character is encoded on a 16 x 16 matrix (Figures 5 and 6). But this choice is not essential. We can represent the characters on a different matrix according to their complexity: we can notably represent a part of the characters on matrices of 8 x 8, another part on matrices of 12 x 12, and another part on matrices of 24 x 24 by grouping the characters whose graphics are the simplest in the same page, and thus resulting in having

  pages of 144 bytes (instead of 256), which may

  ter 18, 8 or 2 characters respectively.

  We can also group characters with the most complex graphics, in special pages-where they are

  encoded on a 32 x 32 matrix; every page of 256 oc-

tets then contains 2 characters.

  In these cases, the "select-print" keyboard must be designed to accommodate different numbers of

  characters per page of memory. The size of a page

  fixed, which greatly simplifies the software, the number of printing keys would correspond in all

  Cases with the maximum number of characters per page. This re

  sultat can be reached by the software of the calculation unit 21 which affects a selection function only to the keys 1 and 2 in the case of display of two characters,

  keys 1 to 4 in the case of displaying four characters

  tères and so on. In the case where only matrices

  these 16 x 16 and 32 x 32 are provided, the keys 1 and 2

  may have a different color for individual

  ser. In addition, methods can be

  to visualize the characters by other means: electro-

  optical, electromechanical, or by not storing the characters themselves but simple algorithms to compute

  characters, saving on memory size.

* .. 7-; 7

Claims (14)

  1. Encoding device comprising a keyboard   coding keys each assigned to an item   character, and a memory storage mass of characters each assigned an address in the memory, characterized in that it comprises display means (12) associated with the mass memory (11) and the keyboard (10) so as to display, in response to the operation of keys assigned to specific phonetic elements,   the contents of a page of memory-constituted by the repre-   of several characters having those elements   phonetics, and keys, operable manually.   nally, selecting one of the displayed characters to cause the issuance of a code representative of the character selected.   2. Encoding device comprising a keypad with   each of which is assigned to a phonogram   tick, and a mass storage memory of characters each assigned an address in the memory, characterized in that the keys for coding the keyboard (10) are   planned to enable each of the initiatives to be selected   the and each of the finals of character and in that the   device further comprises display means associating   to mass memory and the keyboard so as to make   appear in response to pressing   the content of a memory page in which the most frequent character or characters defined by the initial, the final and possibly the tone are defined using the keys, the keyboard comprising   selection keys that can be manually operated to select   display one of the displayed characters and validate it   to cause the issuance of a code representative of the selected.   3. Device according to claim 1 or 2,   characterized by the fact that the keys of. codification   take a set of keys (13) each assigned to a   different initial setting, a set of keys (15) assigned each   to a different end, and ways to differentiate 16- tones.   4. Device according to claim 3, characterized   risqué by the fact that the keyboard carries four keys of tone   (16 -1 for determining the tone.   64). 5. Device according to claim 3, characterized in that the keyboard has four sets of coding keys juxtaposed, each set (321-324) corresponding to a tone to allow determination of the tone.   6. Device according to any one of the claims   3 to 5, characterized in that the keyboard comprises a particular key (18) whose actuation indicates that   the phoneme to be coded has no initial.   7. Encoding device for Chinese characters   any of claims 3 to 6, characterized   the keyboard incorporates special keys   (14) for the phonemes "er" and "weng" which did not   never initial in the four tones.   B. Device according to any one of the   cations 3 to 7, characterized in that on each initial coding key is written an ideogram having a pronunciation element corresponding to the element codified by the key, advantageously chosen from among the most frequent characters having the same sound, or his transcription phonetic.   6. Device according to claim 8, characterized   in that each of the keys assigned to an initia-   it bears a character with a termination different from that of the characters of the other coding keys detinitiale .--   10. Device according to claim 8 or 9,   characterized by the fact that the sounds corresponding to said characters carried by the keys have meanings   in each of the four tones, at least one of these meanings   cations corresponding to frequent characters.   11. Device according to any one of the   cations 3 to 10, characterized in that it comprises a calculation unit provided for fatle appear on the means of   visualization, in response to the action on a first touch   coding system with indication of the tone, the content   a page of memory containing the most frequent characters   quents having the pronunciation corresponding to this coding key.   12. Device according to claim 11, characterized   risqué by the fact that the keyboard has a key "page   next "to go from one page of memory to the next page   with homonyms from the first, and possibly   other characters, which can be   appear a particular message if the page is the last Niere.   13. Device according to claim 11 or 12,   characterized in that the coded representation of a character is the page number in memory followed by the serial number   in the page, its representation according to the telegram code   Chinese language, its translation into another language, or   any combination of the previous representations.   14. Device according to any one of the   cations, characterized in that the keypad   additional keys corresponding to the characters most common Chinese.   Device according to any one of the   cations, characterized in that the memory is organized by pages each containing the same number of   characters coded according to a matrix of   mined, for example a 16 x 16 matrix, each page of   256 bytes containing 8 characters.   16. Apparatus according to any of the claims   1 to 14, characterized in that the characters are   coded according to different matrices according to their   plexity, for example a part of the characters on   trices of & x 8, another part by 12 x 12 matrices,   and another part by 24 x 24 matrices, in re-   grouping characters whose graphics are the simplest in the same page.   17. Apparatus according to any of the claims   cations 3 to 14, characterized in that each ignition key   is also assigned to a final among those not associated with a final touch, the calculation unit of the dis- =   positive being planned to trigger the selection of an initial   the first key pressed and the selection of a final by the second. g