FR2599670A1 - Control method for a writing system and system for writing in Nagari (Sanskrit) - Google Patents

Abstract

The writing system receives, in the form of a coded message MC, a series of characters comprising mainly consonants, matras and signs. A circuit 2 recognises groups with a defined configuration among the series of characters. Each recognised group is transcoded by a transcoding circuit 3, 4, 5 corresponding to the configuration of this group. A transcoding circuit converts each group of signs into a group of graphics. The graphics are stored in a memory 6, addressed by the transcoding circuits. The memory 6 controls a printer 7 and a display screen 8. The invention applies, for example, to data transmission systems terminals.

Description

The present invention firstly relates to a method for controlling a devanagari write writing system, receiving, in the form of a coded message, a sequence of characters of a set of characters divided into several categories of characters. characters.

Such a method applies for example to the terminals of data transmission systems, provided with a printer or a display screen.

By devanagari writing, we mean writing allowing the use of several official languages among the fourteen used in India, including the Hindi language. In the rest of the text, and for the sake of simplification, we will speak more specifically of the Hindi language, but the possibilities of Devanagari writing are not limited to this language.

The Hindi language is actually a phonetic transcription of sounds. Thus, all or even any foreign word can be transcribed in Hindi.

The writing principle consists in choosing the consonant corresponding to the sound and then modifying it by adding vowels. Most of the time, the vowels are not actually written but indicated by accents, called "matras", placed above, below, before or after the consonant.

For some sounds, it is necessary to choose a group of several consonants then to modify this sound by the addition of matras. In such a case, the consonants are said to be interrelated and the writing of these consonants is simplified.

In addition to the matras, accents transcribing signs, each corresponding to a secondary modulation of the sound, can also be placed around the consonant. The same consonant can have several accents corresponding to the matra and several signs.

Unlike Latin languages, Hindi characters are represented graphically in different forms depending on the context in which they are located (ie, according to the characters that precede or follow them).

 It is thus necessary to distinguish characters and graphics, a given character being able to be represented by a graphic or another according to the context in which it is.

Moreover, in the transposition of characters into graphics, which is done in a natural way for an Indian, a graphic corresponding to a matra or a sign, written after a graphic corresponding to a consonant, for example, can be placed before the latter
Since the known Devanagari writing systems are of the same type as those used in the Latin languages, it is necessary that the sequence of characters received, in the form of an encoded message for example, is in fact a series of graphics, in the order in which they must be placed on the printed sheet or on the screen Indeed, in a system of the type used for Latin languages, each received character corresponds to a graphic and only one, and the graphics are written one after the other in the order in which the characters were received.

The use of a system of this type for devanagari writing, however, has the following drawbacks: need for each consonant, each matra, each sign, to have as many codes as possible graphics for this consonant, this matra, or this sign, which makes the coding more complex, - the need for the operator who codes the message, to represent a mental image of each written word before starting to code, given the fact that the place of a graphic in the written word is not necessarily the same as that of the corresponding character in the sequence of written characters.

The present invention aims to overcome these disadvantages.

For this purpose, its object is a method of the type defined above, characterized in that the character string is transformed into a series of groups of characters, such as the sequence of the categories of ca @ acters of a group has a configuration belonging to a set of several predetermined configurations, - one eliminates the characters not belonging to any of said groups of character, - one transcode each of said groups of characters in a group vde graphics, according to a law of transcoding predetermined depend on the configuration of the sequence of categories of the characters of said group of characters, and the sequence of the groups of graphics.

In the method of the invention, the graphics are written only after grouping the corresponding characters into groups of predetermined configurations. These configurations are obviously chosen so that the context influencing the graphics of a character of a group does not extend beyond the limits of this group.

Then, within a group, the transcoding rules of graphic characters, taking into account the influence of the context, are translated by a special transcoding law that defines the configuration of the group. Because of this, to code.

the message, we can use a single code by character, independent of the graphics that will take this character in the written text since it provides a transcoding operation characters in graphics after receiving the message.

The number of codes used is therefore reduced to what is strictly necessary and the coding can be done naturally by an operator, as if he was writing by hand.

In the preferred implementation of the method of the invention, the categories of characters comprise at least the following categories
. matras:
. signs: S consonants: C
. transforms: x - the predetermined configurations comprise at least the following configurations
consonant-based configuration
C, C + S, C + S1 + S2, C + M, C + M + S, C + M + S1 + S2,
. vowel-based configuration
x + M, x + N + S, x + M + Sl + S2,
. configuration based on related consonants
X, X + S, X + S1 + S2, X + M, X + M + S, X + M + S1 + S2 with X = Cl + x + C2
or Cl + x + C2 + x + C3
or C1 + x + C2 + x + C3 + x + C4 - the predetermined transcoding laws comprise at least the following transcoding laws
consonant-based one-group transcoding law
. transcoding law of a vowel-based group
transcoding law of a group based on consonants
linked.

The subject of the invention is also a devanagari write writing system for implementing the preceding method, comprising: means for generating, in the form of a coded message, a sequence of characters of a set of distributed characters in several categories of characters, means for generating, in response to said coded message, a write control signal, means, controlled by said write control signal, for writing a series of graphics representative of said sequence. character, characterized in that it comprises - moyens, provided with an input receiving the coded message and several outputs, for recognizing, in said sequence of characters, groups of characters such as the sequence of characters d a group has a configuration belonging to a set of several predefined configurations and to send each of said groups to one of said outputs, depending on t of the configuration of the sequence of the categories of characters of said group of characters, eliminating the characters not belonging to any of said groups of characters, and, between each of said outputs of said recognition means and said generating means of said control signal of writing, means for transcribing each of said groups of characters and outputting control signals from said generating means of said write signal.

In the preferred embodiment of the system of the invention, said recognition means are arranged to deliver, on an additional output, a signal E representing the sequence of said recognized groups and said generating means of said coded message comprise a keyboard.

With such a system, the operator strikes the consonants, matras and signs keys without worrying about the different graphical forms that these characters can take.
The system chooses the right graphics depending on the context.

The invention will be better understood with the aid of the following description of the preferred implementation of the control method of a devanagari writing system of the invention, and of the preferred embodiment of the writing system. in devanagari of the invention, with reference to the accompanying drawings, in which the single figure shows a block diagram of a teletypewriter according to the invention.

Referring to the single figure, a teletypewriter for devanagari writing comprises, in a known manner - a generator 1 of a coded message representative of a sequence of devanagari characters to be written; in a conventional manner, using a switch 12, this coded message comes from another teletypewriter via a line of transmissions transmitting a signal R, or is generated locally by an operator manipulating a keyboard 11.

- And writing means, here a display screen 8 and a printer 7, provided with a print head of the known type here comprising a vertical bar of needles and moving horizontally to write a horizontal line of graphics.

Between the generator 1 and the previous writing means are arranged: a character group recognition circuit 2, receiving, at a single input, the coded message MC and delivering on four outputs four signals E, C, V and CL a transcoding circuit 3 of the consonant-based groups, receiving the signal C and delivering a transcoded signal C ', a transcoding circuit 4 of the vowel-based groups, receiving the signal V and delivering a transcoded signal Vt, a transcoding circuit 5 of the groups based on connected consonants, receiving the CL-signal and delivering a transcoded signal CL ', and a write control signal generating circuit, here a REPROM memory 6, having the addresses are controlled by the signals C, V 'and CL', to control the screen 8 and the printer 7.

Before describing in more detail the recognition circuit 2, the transcoding circuits 3, 4 and 5, and the memory 6, let us discuss the operation of the teletypewriter shown in FIG.

The characters that may be generated by the keyboard 11, or transported by the signal R are the characters defined for example by Table I on page 17 which, as a person skilled in the art will be able to notice, is an extension of the ASCII standard. .

As it appears on Table III page 19 where each character is named by its hexadecimal code
ISSCII, Hindi characters can be divided into five categories
- the C consonants
- the matras M
- S signs
- the character "transforms" x
- the "other characters".

In Table III, it appears that consonants, matras and signs are further divided into classes and types. The meaning and role of these subdivisions will appear later.

The character "transforms" x is not strictly speaking a character. Its role is to transform the character that follows it.

Thus the hit of the key "transforms" x followed by that of the key of the matra M, which strike will be noted
x + M corresponds to the striking of the vowel corresponding to the matra M. The character "transforms" x thus transforms the matra M into the corresponding vowel. Similarly, for the transcription of sounds where it is necessary to bind together several consonants, we intercalate the character "transforms" x between the consonants to be linked. To link three C1 consonants,
C2 and C3 for example, so we will hit
Ci + x + C2 + x + C3
The "other" characters are characters that correspond to a single graphic, and whose transcription is made as that of Latin characters. This will not be the case.

The character group recognition circuit 2 is arranged so that the groups it recognizes are such that the sequence of the character categories of a group has a configuration belonging to a set of predetermined configurations. These predetermined configurations are as follows
consonant-based configuration
C, C + S, C + S1 + S2, C + M, C + M + S, C + M + S1 + S2
. vowel-based configuration
x + M, x + M + S, x + M + Si + S2
. configuration based on related consonants
X, X + S, X + S1 + S2, X + M, X + M + S, X + M + S1 + S2 with X- = Ci + x + C2
or Cl + x + C2 + x + C3
or Cl + x + C2 + x + C3 + x + C4
These configurations were determined because they correspond to the groups of characters actually encountered in the Hindi language and because the influence of the context on the graphics that a character of such a group must assume is limited to the characters belonging to the same group. Thus, when a group has been recognized, we know all the factors likely to have an influence on the graphics of the characters of this group, and we can thus transcode the sequence of characters of the group into a series of graphics through a law. transcoding. There are three transcoding laws, corresponding to the three previous configuration types, ie - a transcoding law for consonant-based groups - a transcoding law for vowel-based groups - a law transcoding for groups based on bound consonants.

In practice, consonant-based groups and vowel-based groups are simple groups for which transcoding is done in a single step, whereas transcoding for complex consonant-based groups is done in two steps. Upon receipt of the matra of the group of connected consonants, we transcode a first part of the complex group, that is to say the first bound consonants, whose graphics is influenced by the matra, without transcribing the last linked consonant and the matra1 which are then considered as the beginning of a simple consonant-based group, possibly terminated by signs, and encoded as such.

For example, the following character group Cl + x + C2 + x + C3 + x + C4 + M + S1 + S2 will be transcoded in two steps.

In a first step, a first part of the complex group Cl + x + C2 + x + C3 + x + C4 + M will be transcoded, it being understood that only the graphics corresponding to the consonants C1, C2 and C3, influenced by the matra M will be writings.

In a second step, a second part of the complex group: C4 + M + S1 + S2 will be transcoded as a simple consonant-based group.

Since each of the three preceding transcoding laws is implemented in each of the transcoding circuits 3, 4 and 5, it can be said that the group recognition circuit 2, after having recognized these groups, sends the groups of consonants to the output outputting the signal C, the vowel groups on the output delivering the signal V and the first parts of the connected groups of consonants on the output delivering the signal CL.

The recognition circuit eliminates the characters that could not be assigned to one of the preceding groups and additionally delivers the signal E which is the continuation of the recognized groups, that is to say the MC signal in which the possible errors of syntax have been corrected.-If the MC signal is the output signal of the keyboard 11, the signal
E can be used as a signal sent to another teletypewriter, for example.

To explain the operation of the decoding circuits, it is necessary to specify certain points relating to the graphics used to represent the consonants, matras and signs.

As we have had the opportunity to point out about Table III, the matras are divided into classes. These classes, here five in number, correspond to the position of the graphic that represents them in relation to the consonant whose sound it modifies, according to a rule explained in Table III page 19 and Table IV on page 20. The table
IV shows moreover that each matra can be represented by as many graphics as there are tables in the class to which it belongs. The graphic corresponding to a determined matra and to a given table is designated by an alphanumeric code which refers to Table II on page 18, representing all the graphics used. For example, the matra = whose code, in the table
I is E8, is according to Table IV, of class 5. This matra, when it corresponds to a vowel (Table 1) is represented by a graphic whose code, in Table II, is 28, it is to say the graphic <. On the other hand, this matra, when modifying the sound of a consonant of type 1 (table 2) is represented by a graphic whose code, in the table
II, is 98 that is to say the graphic
The signs are divided into three classes according to the position of the graphic that represents them in relation to the consonant whose sound they modify, according to a rule explained in Table III page 19 and Table V on page 21. Table V shows from us ~ that each sign, as previously each matra, can be represented by as many graphics as there are tables in the class to which it belongs.

 L @ s consonants can be of two types according to the form of the graphic that represses them. The consonants always have a horizontal bar and a vertically vertical one. If the intersection between these two bars is lateral, the consonant is of type 1, if the intersection is central, the consonant is of type 2 as shown in Table VI page 22. Like the matras and bookmarks, the consonants are moreover divided into classes, and can be represented by as many graphics as there are tables in the class to which they belong.

Table VII page 22 represents the other characters and a special graphic, called the twelfth sign S12, which does not exist as a character, but which is, in some cases, useful as a graphic.

The transcoding circuits 3, 4 and 5 are arranged to determine the graphics corresponding to the groups of characters by respecting the following rules
Flasks and accents can be placed before, after, above or below the consonant
Class 1 matras are placed after
2 It "" before and on
"3 3 after and above
"4" below
5 "5" above
Class 1 signs are placed on them
2 "2" below 't "3" after
In case of conflicts of place we superimpose the graphics except in the following cases - 8 there is a matra of class 2 and a sign of class 1 we use the "short" graphics of the matra and the graphics "above right" of the sign.

class 1, we use the graphic "whole place" of this twoiaBSi-ge.

- if there is a matra of class 1 or 3 and a sign of class 1, we use the whole place graphic ", of the sign.

Similarly, if there is a second sign of class 1.

- if there is a matra of class 5 and a sign of class 1, a special graphic combining both is used. If there is a second class 1 sign, use the "whole place" graphic for this second sign.

- if there are two signs of class i, use a special graphic containing both signs.

The sign: is always placed at the end of the group.

For vowels corresponding to class 1 matras, the accents are positioned in the same way as for consonants followed by a class i matra.

For vowels corresponding to class 2, 3, 4 and 5 matras, the accents are positioned in the same way as for type 1 consonants.

The transcoding circuits 3, 4 and 5 deliver sequences of alphanumeric codes, which constitute the addresses of the graphics stored in the memory 6, which controls the printer 7 and the display screen 8.

Naturally, the graphics are printed by the printer with some delay due to the fact that the choice of the correct graphics for a character requires knowledge of the characters following the latter. This delay is on average two characters received. The rat is null after typing a Latin character or a typographic sign, for example.

Practically, and in order to avoid to the maximum the returns of the trolley of the printer, this one is arranged so that it is formed a electronic image ("bit map") of the group of graphics to write, before the head n ' printed. Naturally, the head being provided with a vertical bar, the width of the graphics hit is variable.

To facilitate the task of the operator who strikes a text on the keyboard 11 and controls his work using the printer 7, it is.prévu a light, indicating at any time whether characters are pending, and a key allowing the printing, in square brackets, of the pending characters.

On the display screen, however, there is no delay between the display of graphics and the typing of characters, but, as typing, graphics already displayed change to reflect new characters. The display is performed as if the last Hindi character struck was followed by a blank.

For example, there are shown above six successive keystrokes and the corresponding displays
Show Strike
T '
I
Ut
C fiv
Of course, if the operator uses the display screen, the light and the print key for pending characters are useless.

Thus, thanks to the invention, a Hindi language operator only needs to know the following three principles to perform a correct typing - Typing order: consonant + (matra (s)) + (accent (s))
vowel + (accent (s)) (characters enclosed in parentheses are optional) - Vowel selection: key "transforms" x followed by
the key of the corresponding matra - Liaison of the consonants: key "transorme" x between the
consonant keys
In the above, it is remarkable that the typing order defined corresponds to the natural order of the characters written to X main.

In the described embodiment of the system of the invention, the group recognition circuit 2 and the transcoding circuits 3, 4 and 5 are made using a single microprocessor. The different operating algorithms of this microprocessor are described in appendix 1 page 23. They describe, in detail, the sequence of operations performed for the recognition of groups, transcoding of a group based on a consonant, the transcoding of a vowel-based group and transcoding of a group based on bound consonants or more exactly the transcoding of the first part of such a group, the second part of such a group being transcoded as a consonant-based group.

The main conventions used for the presentation of these algorithms are specified below - "?" = any character - italic characters denote sub call
corresponding program - pre (C) = character preceding consonant C
finsi = end of the "if" test started earlier with the same horizontal offset.

- pointer: + n = the pointer marks the beginning of the group whose recognition is in progress; as soon as the group which has been recognized is transcoded, the pointer of the number of characters indicated is advanced to start the recognition of the next group - graphic (M, 2) = graphsime of the table 2 of the matra M.

The other conventions used are conventional and known to those skilled in the art who can therefore, from these algorithms, write without difficulty the corresponding calculation programs.

Examples of transcoding particularly representative of the operation of the algorithms are described in Annex 2 page 29.

Naturally, the scope of the invention which has just been described is not limited to the description that has just been made of the method of the invention and the associated teletypewriter.

On the contrary, it understands all the variants. In particular, from the presentation that has been made, it would be possible for a person skilled in the art to modify the algorithms to recognize and transcode a group of five linked consonants for example, instead of the four to which one has voluntarily limit. Such a modification is not outside the scope of the present invention.

0 <SEP> 1 <SEP> 2 <SEP> 3 <SEP> 4 <SEP> 5 <SEP> 6 <SEP> 7 <SEP> 8 <SEP> 9 <SEP> A <SEP> B <SEP> C <SEP> D <SEP> E <SEP> F
<tb> 0 <SEP> NULL <SEP> DLE <SEP> SP <SEP> 0 <SEP>&commat;<SEP> P <SEP>#<SEP>#<SEP> 3 # <SEP> 3 # <SEP>#<SEP>#<SEP>#<SEP>#
<tb> MUNT
<tb> 1 <SEP> SOH <SEP> OCI <SEP>! <SEP> 1 <SEP> A <SEP> Q <SEP> a <SEP> @ <SEP>#<SEP> 3 # <SEP>#<SEP>#<SEP>#<SEP>#
<tb> 2 <SEP> STX <SEP> OCZ <SEP>"<SEP> 2 <SEP> B <SEP> R <SEP> b <SEP> r <SEP>#<SEP>#<SEP>#<SEP>#<SEP>#<SEP>#
<tb> LIHK
<tb> 3 <SEP> E @ X <SEP> DC3 <SEP>&num;<SEP> 3 <SEP> C <SEP> S <SEP> c <SEP> ss <SEP>#<SEP>#<SEP>#<SEP>#<SEP>#<SEP>#
<tb> 4 <SEP> ECT <SEP> DC4 <SEP> R @. <SEP> 4 <SEP> D <SEP> T <SEP> d <SEP> t <SEP> D-SP <SEP> 3 # <SEP>#<SEP>#<SEP>#<SEP>#
<tb>###
<tb> 5 <SEP> ENO <SEP> NAK <SEP>% <SEP> 5 <SEP> E <SEP> U <SEP> e <SEP> u <SEP> @ <SEP> 3 # <SEP>#<SEP>#<SEP>#<SEP>#
<tb> AVAG @ AH
<tb> 6 <SEP> ACK <SEP> SYN <SEP>& 6 <SEP> F <SEP> V <SEP> f <SEP> v <SEP> 3 # <SEP> 3 # <SEP>#<SEP>#<SEP>#<SEP>#
<tb> 7 <SEP> BEL <SEP> ETB <SEP>#<SEP> 7 <SEP> G <SEP> W <SEP> g <SEP> w <SEP> 3 # <SEP> 3 # <SEP>#<SEP>#<SEP>#<SEP>#
<tb> 8 <SEP> BS <SEP> CAN <SEP>(<SEP> 8 <SEP> H <SEP> X <SEP> h <SEP> x <SEP>#<SEP>#<SEP>#<SEP>#<SEP>#<SEP>#
#### SEP>#<SEP>#<SEP>#
<tb> A <SEP> LF <SEP> SUB <SEP> * <SEP>: <SEP> J <SEP> Z <SEP> j <SEP> z <SEP> 3 # <SEP>#<SEP>#<SEP>#
<tb> B <SEP> VT <SEP> ESC <SEP> + <SEP>;<SEP> K <SEP>[<SEP> k <SEP>{<SEP>#<SEP>#<SEP>#<SEP>#<SEP>#
<tb> C <SEP> FF <SEP> FS <SEP>, <SEP><SEP> L <SEP> \ <SEP> l <SEP> @ <SEP>#<SEP>#<SEP>#<SEP>#<SEP>#
<tb> D <SEP> CR <SEP> GS <SEP> - <SEP> = <SEP> M <SEP>] <SEP> m <SEP>} <SEP>#<SEP>#<SEP>#<SEP>#<SEP>#
<tb> E <SEP> N <SEP> N <SEP> N <SEP> N <SEP> N <SEP>#SEP># SEP >#<SEP>#
<tb> F <SEP> S @ <SEP> US <SEP> / <SEP>? <SEP> O <SEP> - <SEP> o <SEP> DEL <SEP>#<SEP>#<SEP>#<SEP>#<SEP>#<SEP> LED
<tb> TABLE I (Characters)

<SEP> i0 <SEP> li <SEP> I, & I, i,; 0i (JiI; iAI; I <SEP>i;;,<SEP> lE <SEP> IF
<tb><SEP> 3
<tb><SEP> 4 <SEP> t <SEP> I <SEP>'w
<tb><SEP> .iiw
<tb> 5 <SEP>;<SEP> --- <SEP> I <SEP> I <SEP> r "
<tb> 6 <SEP> 1 <SEP>, <SEP>, <SEP> ~ <SEP>'I<SEP>
<tb> iD <SEP> D <SEP>;
<tb><SEP><SEP> v
<tb> m <SEP> ib <SEP> s <SEP> e- <SEP>M;<SEP> r <SEP> c
<tb><SEP> I
<tb><SEP>!<SEP>.<September>
<Tb>

A <SEP>% t '<SEP> v <SEP> c- <SEP>&<SEP> D
<tb> ~ <SEP>. <SEP>. <SEP> ~
<tb> 13 <SEP> h <SEP> .- <SEP> - <SEP> r <SEP> O <SEP> o <SEP> 2
<tb> c <SEP> b <SEP> --- t- '<SEP> - <SEP> Fw <SEP> o <SEP>6;<SEP> 6
<tb><SEP> D <SEP> 0 <SEP> I <SEP> +
<tb><SEP> E <SEP> - <SEP> tr <SEP> z <SEP> jazzzen &'<SEP> Fk $ <SEP> fF
<tb><SEP> F <SEP> I <SEP>, <SEP>, <SEP> I
<tb> cti <SEP>. <SEP> h <SEP> Fb <SEP> t <SEP>><SEP> Dt <SEP> s <SEP> t <SEP> Fs
<tb> TABLE II (Graphics)
Matra class 2 = (E2) (before-above) 1 matra
"" 3 = (E3) (after above) 1 matra
"" 1 = (E1, EC, ED, EE, EF) (after) 5 flasks
"" 4 = (E4, E5, E6) (below) 3 flasks
"" 5 = (E8, E9, EA, EB) (top) 4 matras Total: 1 4 flasks
Sign of class 1 = (60, A1, A2, F5, FA, F13, FC) (top) 7 signs
"""2 = (E0, F0, F1) (below) 3 signs
"""3 = (A3) (after) 1 signs
Total: 11 monkeys
Class I terms = type 1: (A6, BE, CO, C4, D8, F6-F9)
+ type 2: (BC, C2, C3, C6, CB, DE) 15 consonants
Consonant class 2 = type 1: (B9-BB, BD, BF, C1, C6-CA, CC,
CD, CF, D1-D4, D7, DA-DD)
+ type 2: (B8, D0) 23 consonants
Consonant class 3 = type 2: (D5) 1 consonant
39 consonants
Other = (A0, A5, F3, F4, Latin, typographic signs) code transforms = F2.

TABLE III
Matras of class 1 (after):
table l table 2
vowel shape
par
E1T 20 + 82 (3 #) 82
EE # 20 + 83 (3 #) 83
EF # 20 + 84 (3 #) 84
EC, # 20 + 85 (3 #) 85
EDT 20 + 86 (3 #) 86
Matra class? (front on)
table l table 2 table 3 table4
vowel consonant consonant sign class 1
type 1 type 2 (short) E2 # 21 89 88 87
Matra class 3 (after above):
table l table 2 table 3
vowel consonant consonant
type 1 type 2
E3 # 22 8A 8B
Matras class 4 (below) ::
table l table 2 table 3 table 4
vowel consonant consonant consonant
type 1 type 2 D5 () E4 # 23 8C 8D 92 E5 # 24 8E 8F 93 E6 # 25 90 91 91
Matras class 5 (top):
table l table 2 table 3 table4
vowel consonant consonant on the left
type 1 type 2 E8 #, 28 98 99 9C E9 # 29 9A 9B 9D EA # 26 94 95 9E
EB # 27 96 97 AO
TABLE IV Class 1 Signs (Top) Table 1 Table 2 Table 3 Table 4 Table 5 Table 6 Table 7 Table 8 Table 9 Table 10 Top Top Place With Sign With Sign With Sign With Sign With Sign With Sign Consonant Right Antiere A2 A2 F5 FA 60 FB FC
A1 # A8 AF B6 BF C0 C1 C2 C3 C4
A2 # A9 B0 B7 BF C5 C6 C7 C8 C9
F5 # Reph AD B1 B8 C0 C5 CA CB CC CD
FA # AB B2 B9 C1 C6 CA CE CF D0 60 # AC B3 BA C2 C7 CB CE D1 D2
FB # AD B4 BB C3 C8 CC CF D1 D3
FC # AE B5 BC C4 C9 CD DO D2 D3
Class 2 signs (below) table 1 table 2 consonant consonant type 1 type 2
E0 # A4 A5
F0 # A2 A3
F1 # A6 A7
Sign of class 3 (after) table 1
# A3: A1
TABLE V
CLASS 1 SOUNDS:

<tb><SEP> type <SEP> table <SEP> 1
<tb><SEP> form
<tb><SEP> normal
<tb><SEP> BC # <SEP><SEP> 2 <SEP> 2E
<tb><SEP> BE # <SEP><SEP> 1 <SEP> 30
<tb><SEP> CO # <SEP><SEP> 1 <SEP> 32
<tb><SEP> C2 # <SEP><SEP> 2 <SEP> 34
<tb><SEP> C3 # <SEP><SEP> 2 <SEP> 35
<tb><SEP> C4 # <SEP><SEP> 1 <SEP> 36
<tb><SEP> C6 # <SEP><SEP> 2 <SEP> 37
<tb><SEP> CB # <SEP><SEP> 2 <SEP> 3A
<tb><SEP> DB # <SEP> 1 <SEP> 4F
<tb><SEP> FROM # <SEP><SEP> 2 <SEP> 4A
<tb> F6 # <SEP><SEP> 1 <SEP> 4B
<tb><SEP> F7 # <SEP><SEP> 1 <SEP> 4C
<tb><SEP> F8 # <SEP><SEP> 1 <SEP> 4D
<tb><SEP> F9 # <SEP><SEP> 1 <SEP> 4E
<tb><SEP> A6 <SEP> 3 # <SEP><SEP> 1 <SEP> 20
<Tb>
CLASS 2 SITUATIONS:

<tb><SEP> type <SEP> table <SEP> l <SEP> table <SEP> 2 <SEP> table <SEP> 3
<tb><SEP> Form <SEP> 1/2 <SEP> Form <SEP> Modified <SEP> By
<tb><SEP> normal <SEP> consonant <SEP> D5
<tb><SEP> B8 # <SEP> 2 <SEP><SEP> 2A <SEP> 50 <SEP> 67
<tb><SEP> B9 # <SEP><SEP> 1 <SEP> 2B <SEP> 51 <SEP> 68
<tb><SEP> BA # <SEP><SEP> 1 <SEP> 2C <SEP> 52 <SEP> 69
<tb><SEP> BB # <SEP><SEP> 1 <SEP> 2D <SEP> 53 <SEP> 6A
<tb><SEP> BD # <SEP><SEP> 1 <SEP> 2F <SEP> 54 <SEP> 6B
<tb><SEP> BF # <SEP><SEP> 1 <SEP> 31 <SEP> 55 <SEP> 6C
<tb><SEP> C1 # <SEP> 1 <SEP> 33 <SEP> 56 <SEP> 6D
<tb><SEP> C8 # <SEP><SEP> 1 <SEP> 38 <SEP> 57 <SEP> 6E
<tb><SEP> C9 # <SEP><SEP> 1 <SEP> 39 <SEP> 58 <SEP> 6F
<tb><SEP> CA # <SEP><SEP> 1 <SEP> 3A <SEP> 59 <SEP> 70
<tb><SEP> CC # <SEP> 1 <SEP> 3C <SEP><SEP> 5A <SEP> 71
<tb> CD # <SEP><SEP> 1 <SEP> 30 <SEP> 5B <SEP> 72
<tb> CF # <SEP><SEP> 1 <SEP> 3E <SEP> 5C <SEP> 73
<tb><SEP> D0 # <SEP><SEP> 2 <SEP> 3F <SEP> 5D <SEP> 74
<tb><SEP> D1 # <SEP><SEP> 1 <SEP> 40 <SEP> 5E <SEP> 75
<tb><SEP> D2 # <SEP><SEP> 1 <SEP> 41 <SEP> 5F <SEP> 76
<tb><SEP> D3 # <SEP><SEP> 1 <SEP> 42 <SEP> 60 <SEP> 77
<tb><SEP> D4 # <SEP> 1 <SEP> 43 <SEP> 61 <SEP> 78
<tb><SEP> D7 # <SEP><SEP> 1 <SEP> 45 <SEP> 62 <SEP> 79
<tb> DA # <SEP><SEP> 1 <SEP> 46 <SEP> 63 <SEP> 7A
<tb><SEP> DBZT <SEP> 1 <SEP> 47 <SEP> 64
<tb> DC # <SEP><SEP> 1 <SEP> 48 <SEP> 65
<tb><SEP> DD # <SEP> 1 <SEP> 49 <SEP> 66
<Tb>
CLASS 3 CONSON

<tb> D5 # <SEP> 2 <SEP> 44
<Tb>
Others:
table 1 A0 3 # 7F A5 # 7E
F3 # 80
F4 # 81
Special graphics:
table 1 table 2 S12 # BD BE
TABLE VII
APPENDIX I: Algorithms used - algorithm of the recognition of the groups of analysis: if in the chain of the received characters, 2 codes x follow each other
destroy one of the two codes x finsi do box start string of received characters: = C S1 S2: transcodingcon form of (pre (C) C S1 S2) -pointer: +3 = C S1? "" (pre (C) C S1) "+2 = CS nothing to do = CM S1 S2""(pre (C) CM S1 S2)" +4 = CM S1? "" (pre (C) CM S1) "+ 3 = CM?" @ (pre (C) CM) "+2 = CM: nothing to do = C1 x C2 x C3 x C4?: transcoding bound consonants of (C1 , C2, C3, C4,?)
if? = matra E2 () then destruction of?
pointer: +6 (= on C4) = C1 x C2 x C3 x C4: nothing to do = Cl x C2 x C3 x?: transcoding bound consonants of (C1, C2, C3) point .: + 4 = C1 x C2 x C3 x: nothing to do = C1 x C2 x C3? : transcoding bound consonants of (C1, C2, C3,?).

if? = matra E2 (@) then destruction of?
pointer: +4 (= on C3) = Cl xC2xC3: nothing to do = Ci x C2 x? : transcoding bound consonants of (C l, C2) -pointer: +2 = C1 x C2 x: nothing to do = C1 x C2? : Transcoding Related consonants of (C1, C2,?)
if? = matra E2 () then destruction of?
pointer: +2 (= on C2) = C1 x C2: nothing to do = C x: nothing to do = C? : transcoding consonant of (pre (C), C) -pointer: +1 = XM S1 S2: vowel transcoding of (x M S1 S2) -pointer: +4 = x M S1? : vowel transcoding of (x M S1) -pointer: +3 = x M? : vowel transcoding of (x M) -pointer: +2 = x M: nothing to do = x? destroy x = M: destroy M = S: destroy S =? : immediate transcoding of? -pointer +1
End of box transcoding algorithm of a simple group to consonant definition of (transcoding consonant) -in entry: group of structure (preC) C (M) (Sl (52)) beginning treatment of the matra sy consonant if M = matra class 2 (before above)
if S1 or S2 = class 1 sign (top)
graphic design (M, 4)
if not
if C = consonant type I
graphic design (M, 2)
if not
graphic design (M, 3)
end if
thus processing the consonant if preC <> transforms or C <> consonant D5
graphism (C, 1) finsi tralement of the matra after the consonant if M = matra of class 1 (after)
graphic (M, 2) otherwise
if M = class 4 matra (below) and C = consonant D5
graphic design (M, 4)
if not
if M = class 5 mastra (top) and S1 or S2 = class 1 sign (top)
graphic design (M, 4)
if not
if C = consonant type 1
graphic design (M, 2)
if not
graphic design (M, 3)
end if
end if
finsi finsi treatment of sign processing signs (CM S1 S2) end definition of (sign processing) -in entry: structure group (H) (M) (S1 (S2)) where H = consonant or x start ranking of both signs in 1 order class 1. class 2. class 3 if 52 and S l exist
if S1 = sign of class 3, inversion of S1 and 52 finsi
if S2 = sign of class l, inversion of S1 and S2 finsi finsi treatment of the first sign siSl sign class l (top)
if M = matra class l (after) graphics (S1,3) finsi (whole place)
if M = matra class 2 (before above) graphism (S1,2) finsi (right)
if M = matra class 3 (after above) graphics (S1,3) finsi (whole place)
if M = matra class 5 (top) graphics (S1,2) finsi (right)
if M = matra class 4 (below) or does not exist
if S2 = class sign l (top)
if S2 = sign Al, graphic (S1,4) finsi
if 52 = sign A2, graphic (S1.5) finsi
if 52 = sign F5, graphic design (51,6) finsi
if 52 = sign FA, graphic design (51,7) finsi
if 52 = sign 60, graphic (S 1.8) finsi
if S2 = sign FB, graphics (51,9) finsi
if 52 = sign FC, graphics (51,1 O) finsi
if not
graphic design (S1,1) (consonant top)
end if
so if S1 = sign class 2 (below)
if H = consonant type 1 or H = x
graphic design (S 1,1)
if not
graphic design (S 1,2)
so if S1 = sign class 3 (after)
graphic design (S1, 1) finsi
Treatment of the 2nd sign if 52 = class sign 1
if M exists and <> matra class 4
graphics (52, 3) (whole place)
so if 52 = sign class 2 (below)
if H = consonant type 1 or H = x
graphic design (52, 1)
if not
graphic design (52,2)
so if 52 = sign class 3 (after)
graphism (S2, 1) finsi end transcoding algorithm of a simple group vowel definition of (vowel transcoding) -in entry: group of structure xM (S1 (S2)) early treatment of vowel graphism (M, 1) processing accents treatment of signs (x M S1 52) end transcoding algorithm of a first part of a complex group definition of (transcoded bound consonants) - input: structure group C1 C2 <C3 (C4)) (M) start processing of the matra to be placed before the consonants if M = matra E2 (f)
if C1 = consonant type 1
graphic design (M, 2)
if not
graphic design (M, 3)
finsi finsi treatment of consonant formation half-consonant (input: C1, C2 output: info) if C3 exists and info = a consonant treated
half-consonant formation (entry: C2, C3 exit: info) otherwise
info = a consonant processed finsi if C4 exists and info = a consonant processed
half-consonant formation (input: C3, C4 output: info) finsi end definition of (half-consonant formation) -in entry: structure group Ca Cb (where Ca and Cb are consonants) -on the output: info info = either '1 1 consonant processed'
either '2 consonants treated' beginning case where the consonant ler is the consonant (D5) if Ca = consonant D5
graphic design (Cb, 1)
if M = matra class 1 or 3
graphic design (sign F5,3) (whole place)
if not
graphic design (sign F5,1) (on top)
end if
info = 2 consonants processed otherwise case or the second consonant is the consonant (D5)
if Cb = consonant D5
if Ca = class 2 consonant
graphics (Ca, 3)
if not
graphics (Ca, 1)
if Ca of type 1
graphic design (sign S12,1)
if not
graphic design (sign 512,2)
end if
end if
info = 2 consonants processed
if not other cases
if Ca = class 1 consonant
graphics (Ca, 1)
if Ca of type 1
graphic (sign FO, 1)
if not
graphic design (sign F0,2)
end if
otherwise (class 2 Ca)
graphics (Ca, 2)
end if
info = 1 consonant processed
thus, end
APPENDIX 2: EXAMPLES
Transcoding consonants # + # = # (C + M) # + # = # (C + M) # + # = # (C + S)
# + # + # = # (C + M + S) # + # = # (C + S) # + # + # = # (C + S1 + S2)
# + # + # = # (C + S1 + S2)
# + # + # = # (C + S1 + S2)
Vowel Transcoding x + # = # (x + M)
x + # = 3 # (x + M)
x + = # (x + M)
Transcoding of bound consonants + + x + # = if (C + x + C)
# + x + # = # (C = x + C)
# + x + # = # (C + x + C)
# + # + # = ## (C + x + C + M)

Claims (10)

1.- A method for controlling a devanagari writing writing system, receiving, in the form of a coded message, a sequence of characters of a set of characters divided into several categories of characters, characterized in that the character string is transformed into a series of groups of characters, such that the sequence of the categories of the characters of a group has a configuration belonging to a set of several predetermined configurations, the characters which do not belong to any one of these are eliminated; groups of characters - one transcode each of said groups of characters into a group of graphics, according to a predetermined transcoding law, depending on the configuration of the sequence of categories of characters of said group of characters, and - one writes the following groups of graphics . 2. The method of claim 1, wherein the character categories comprise at least the following categories.  matras: M  signs:  . consonants: C  transforms: x - the predetermined configurations comprise at least the following configurations  consonant-based configuration  C, C + S, C + S1 + S2, C + M, C + M + S, C + M + S1 + S2,  . vowel-based configuration  x + M, x + M + S, x + M + Sl + S2,  . configuration based on related consonants  X, X + S, X + S1 + S2, X + M, X + M + S, X + M + S1 + S2 with X = Cl + x + C2  or Cl + x + C2 + x + C3  or C1 + x + C2 + x + C3 + x + C4 - the predetermined transcoding laws comprise at least the following transcoding laws  . transcoding law of a consonant-based group  . transcoding law of a vowel-based group  transcoding law of a group based on connected consonants. 3. A method according to claim 2, wherein the characters of the matras category are divided into at least five classes according to the position of the graphic that represents them in relation to the consonant whose sound they modify, according to the rule of next distribution  Position Class  after 1  before and above 2  after and above 3  below 4  above 5 - a matra of class 2 is represented by at least one so-called "short" graphic, - the characters of the "signs" category are divided into at least three classes according to the position of the graphic that represents them in relation to the consonant whose sound they change according to the following distribution rule  Position Class  above  below 2  after 3 - a sign of class 1 is represented by at least one so-called "whole place" graphic and a so-called "top right" graphic - the characters of the "consonants" category are divided into at least two types according to the form of their graphics, according to the following distribution rule  Graphic form Type of lateral intersection of the horizontal bar and the substantially vertical bar Central intersection of the horizontal bar and the substantially vertical bar - the predetermined transcoding laws are such that, in the event of space conflicts, the graphics except in the following cases. if there is a class 2 matra and a class sign  1 we use the "short" graphics of matra and graphics  "On the right side of the sign, if there is a second sign of  class 1, we use the graphic "whole place" of  this second sign.  sign. Similarly, if there is a second sign of class 1.  class 1, we use the whole place graphics "of the  . if there is a matra of class 1 or 3 and a sign of the  whole place "for this second sign.  there is a second sign of class 1 we use the graphic  we use a special graphic combining both. if  . if there is a matra of class -5 and a sign of class 1  special phisme containing both signs.  if there are two signs of class 1, we use a gra The sign: is always placed at the end of the group. For vowels corresponding to class 1 matras, the accents are positioned in the same way as for consonants followed by a class 1 matra. For vowels corresponding to class 2, 3, 4 and 5 matras, the accents are positioned in the same way as for type 1 consonants. 4.- devanagari write writing system for implementing the method according to one of claims 1 to 3, comprising - means (1) for generating, in the form of a coded message (MC), a sequence of characters of a set of characters divided into several categories of characters, - means (6) for generating, in response to said coded message, a write control signal (CE), - means (7, 8), controlled by said write control signal (CE), for writing a series of graphics representative of said sequence of characters, characterized in that it comprises - means (2), provided with an input receiving the coded message (MC) and several outputs, for recognizing, in said sequence of characters, groups of characters (C, V, CL) such that the sequence of characters of a group has a configuration belonging to a set of several predetermined configurations and to send each of these groups (C, V, CL) on one of said outputs, depending on the configuration of the sequence of the categories of characters of said group of characters, eliminating the characters -n't belonging to any of said groups of characters  (C, V, CL), and, between each of said outputs of said recognition means (2) and said generator means (6) of said write control signal (CE), means (3, 4, 5) for transcode each of the said groups of characters  (C, V, CL) and outputting control signals (C ', V', CL ') from said generating means (6) of said write signal (CE). 5. Writing system according to claim 4, wherein said recognition means (2) are arranged to output to an additional output, a signal E representing the sequence of said groups (C, V, CL) recognized. 6. Writing system according to one of claims 4 or 5, wherein said writing means comprises a printer (7) provided with a print head, and control means of said head, arranged to forming an electronic image of the graphics group to be written before controlling the print of the head.  The writing system according to one of claims 4 to 6, wherein said writing means comprise a display screen (8), and control means of said screen for writing immediately the graphics controlled by said generating means ( 6) of said write control signal (CE), and then modify this graphic, if necessary, to take account of the new characters arrived. 8. Writing system according to one of claims 4 to 7, wherein said generating means (1) of said coded message (MC) comprise a keyboard (11>. The writing system according to claim 8, wherein means are provided for signaling that characters generated by said keyboard (11) have not yet been written by said writing means (7, 8) and a key to order, at any time, the writing of such characters, in square brackets. 10. Writing system according to one of claims 4 to 9, wherein said generating means (1) of said coded message (MC) comprise a line of teletransmissions receiving a signal (R) from another system d 'writing.