WO2009103208A1 - Sentence component device and reading foreign languages and producing universal language and text conversion method - Google Patents

Abstract

A sentence component device, including a sentence component storing part (101), comprising a sentence component base of a sentence component constituted by an electronic data form; a sentence pattern base (300), for storing a sentence pattern component, including a sentence pattern code, a multi-language category sentence pattern field, comprising at least one record, the cabin semanteme of the different language category in the same record are same; a sense cluster base (500,502), for storing a sense cluster component, including a sense cluster code, a multi-language category cluster field, comprising at least one field, the sense cluster semanteme of the different language category in the same record are same; a phrase base (600), for storing the small phrase component, including a phrase code, a multi-language category phrase field, comprising at least one record, the small phrase semanteme of the different language category in the same record are same; a sense understanding code programming part (103), connecting with the sentence component storing part (101), programming a sense understanding code for each record of the sentence component base.

Description

 Statement component device and native language read foreign language and generate world text and text conversion method

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an apparatus and method for processing or converting natural language; and more particularly to an apparatus and method for text conversion; namely, a statement component device, a statement component making method, and a text conversion method based on a sentence component for a native language reading foreign language . BACKGROUND OF THE INVENTION Information processing of computer language words represents machine translation, which is the highest point of technical difficulty. Its operational object is a text file, produced by the computer's word processing technology. The existing word processing technique is to encode characters of various languages and then generate a text file by using a character code (internal code). The shortcomings are the same as the files on paper, which can only be read, written and communicated by people of their own language. Therefore, people of different languages must rely on translation. Machine translation such as "Computer Processing of Natural Languages" Feng Zhiwei, Shanghai Foreign Language Education Press, published in October 1996, the eighth chapter of the first section of machine translation. The full text elaborates on the beginning of the idea of the use of machines for language translation in the early 1930s by the French scientist Alcuuni; by the time the world's first computer was introduced in 1946, the study of machine translation began in the same year, and it flourished for a while; 1966 In November, the Academic Automatic Processing Advisory Committee of the American Academy of Sciences published a report entitled "Language and Machines (ALPAC)", negating machine translation and pointing out that machine translation encountered insurmountable "semantic barriers"; Unprecedented depression; recovery in 1970-1976; and the prosperity after 1976, etc., 20,000 words, and finally pointed out that "in 1964, the "semantic barrier" encountered by machine translation in the US ALPAC report still exists today, machine translation technology It still seems that there is still no breakthrough progress.""The practical and commercialization of machine translation systems is facing a severe test." Volkswagen Software, the second issue of 2004, author Wang Wei, interviewed many experts in the industry and reported that "Machinery Translation, Where is the Road" wrote: "The current machine translation mainly has two forms, MT and TM. MT (machine translation) is our common rule-based machine translation software, its main purpose is to help users with poor English to provide translation reference, but the accuracy is not high. The key technology of MT has four aspects: word analysis, syntax analysis , meaning analysis and literary analysis. Its working process is to divide the sentence into several words, find out the meaning of the word through the electronic dictionary stored in the machine database, analyze the meaning of the statement according to the grammar rules, and transform it into a conceptual construct. Then use the language model to generate the target language. Although in principle, it is not difficult to implement this series of steps, but due to the particularity and variety of language, and the limitations of the development level of artificial intelligence technology, it is currently required to be in different languages. Correct translation is impossible, which is why the current machine translation software cannot .

 Memory, translation memory, mainly designed for professional translators and organizations, with translation memory and human-computer interaction as the core, requiring users to have independent translation capabilities. The principle of TM is based on a database, and all translated materials are stored in the database in sentences. The system automatically analyzes the electronic documents during translation. The 100% matching sentences can be automatically replaced. The partially matched sentences can be translated according to the matching degree, and the new sentences are manually translated through the translation suggestions provided by the system. Scientific research shows that the amount of repetitive work in translation is about 30%, and the TM translation software makes "the same sentence never needs to be translated a second time", thus improving work efficiency. "But 70% still have to rely on human workers... The last author pointed out: "The problem with machine translation technology itself is a flaw that hinders its development. Objective ^ Not only in China, there is no major breakthrough in machine translation technology throughout the world. Attempts to improve translation accuracy through machines with limited rules and corpus are not possible in the short term. Under the condition that the theory of linguistic intelligence research is not mature, MT software research encounters bottlenecks in technology, can not solve the problem of word meaning selection in different language environments, and can not correctly select grammar in complex and variable context. Rules, therefore, the level of translation cannot be significantly improved. "In view of the prior art, the hope lies in another way! SUMMARY OF THE INVENTION The technical problem to be solved by the present invention is:

1. The expression of the language of the existing computer, the meaning of the words, words and sentences in different languages is not uniform;

2. Existing machine translation and text conversion; the results are poorly readable and cannot express the original text. The object of the present invention is to overcome the deficiencies of the prior art, and to provide a sentence component device and a method for fabricating a two-sentence component; and using the technical means of human-computer interaction to solve the language expression of the existing computer, which is also a word, a word, a sentence expression Technical issues that are not equal or uniform. Thereby, the technical effects of the sentence components stored in the sentence component library and the different language words, words and sentences are equivalent and unified are generated. A further object of the present invention is to provide a text conversion method based on a sentence component for a native language reading foreign language; to solve the existing machine translation and text conversion, the result is poorly readable, and cannot express the technical problem of the original text meaning; thereby improving the translation or conversion The readability of the text, the technical effect of the original meaning consistent with the original text. The conversion also generates Esperanto, which produces technical effects that can be read in multiple languages. It has greatly changed the status quo of the existing word processing text for the reading and writing of people in this language. The technical solution adopted by the present invention to solve the technical problem thereof is:

</ RTI><RTIgt;</RTI><RTIgt;</RTI><RTIgt;</RTI><RTIgt;</RTI><RTIgt;</RTI><RTIgt;</RTI><RTIgt;</RTI><RTIgt;</RTI><RTIgt; Stored multilingual semantics , : sentence pattern library 300, used to store sentence structure components, there are sentence patterns, English sentence patterns, Chinese sentence patterns, Russian sentence pattern fields, which contain at least one record, the same semantic sentence pattern is the same record, corresponding The sentence patterns of the genre are stored in the corresponding sentence pattern field, and the sentence pattern code represents the language of each sentence sentence type in each sentence pattern field in the same record; the cabin model library 400 is used for storing the cabin model member , there is a cabin model, an English cabin model, a Chinese cabin model, a Russian cabin model field, which contains at least one record, the same semantics of the cabin model with the same record, the corresponding language of the cabin model is stored in the corresponding language cabin model field Inside, the cabin model code represents the language of each language module in each language model field in the same record; the meaning group library 500, 502, used to store the meaning group components, intentional group code, English string, Chinese A string, a Russian string field, which contains at least one record, a syntactic group of the same semantics is in the same record, and the meaning group of the corresponding language is stored in the corresponding text string field, and the meaning group code represents the text in the same record. a string of words in a string field The idiom library 600 is configured to store the small idiom component, and has an idiom code, an English idiom, a Chinese idiom, a Russian idiom field, which includes at least one record, and a small idiom of the same semantics is recorded at the same time. The small idioms of the corresponding language are stored in the corresponding idiom field, and the idiom code represents the semantics of each idiom in each idiom field of the same record; the ideology code preparation unit 103, and the statement The component storage unit 101 is connected to receive the notification of the component adding unit 106, and when the new record is generated by any one of the above four libraries, the current library representative number is added to the high-order word plus the current library record number to generate the meaning code, and Fill in the code field of the current library as the unified double-byte fixed-length multi-language semantic interworking code of the statement component. The Italian code is unique to the same semantic representation of each language component in the current record of the current library; The component reading unit 104 is connected to the sentence component storage unit 101 for receiving the read command, determining a certain record of the library by the number of segments included in the code, and reading the required words into the corresponding record of the corresponding library. a member matching unit 105 is connected to the sentence component storage unit 101 for receiving a matching command, and querying the corresponding language index field of the corresponding component library according to the sentence or the sentence content of the given language and the guidance of the current operating point. Matching, giving the required language component of the match or returning the no match signal; the component adding unit 106 is respectively connected to the sentence component storage unit 101 and the Italian code compiling unit 103, and is configured to receive the add new component command, and confirm the corresponding in the query. After the component library has no identical components, the new component is added to the corresponding language component field of the corresponding component library. When a new component is added to a new record, the information is simultaneously notified to the code generation unit 103; component library operation control, interface unit 107. The component reading unit 104, the component matching providing unit 105, and the component adding unit 106 are connected to the sentence component storage unit 101, and receive calls of various applications based on the present sentence member or , , ,

Other application devices are connected. The statement component in the statement component device: the statement component is obtained by means of expert operation and human-computer interaction, and is obtained from the parsing comparison bilingual training sample corpus; another source of the statement component is that the user feedback information is reviewed by the expert. The statement component is a standard component for assembling a language sentence or a standard component for encoding a sentence, including the following four types:

A sentence structure component 201, 301, which is used to form the basic structural framework of a sentence, represents the basic semantic category of the sentence, and also determines the order and number of sentence boxes contained in the sentence, and includes the sentence. More complicated grammatical phenomena;

2 cabin module 202, 401, which is used to form the basic structural framework of the complex sentence cabin, represents the basic semantics of the sentence cabin, and also determines the number and number of cabins contained in the sentence cabin. The more complicated grammatical phenomena of this type of sentence cabin;

The syllabus members 501, 503 are components that are used by the ensemble string to fill the components of the simple sentence capsules 203 to 204 or the bonnets 205 to 207. The simple sentence cabin and the cabin eye are the same as the upper and lower concepts. , are not more than three meaning clusters except for the meaning of the word;

4 small > J-language component 601, which is too short to be used to separate sentences and sentence sentences as small components, used to directly form short sentences. The statement component library in the statement component device: the language included in the library, except for English, Chinese, and Russian, each additional sentence, the sentence library, the cabin model library, the Italian group library, The idioms are each added with a certain sentence pattern, a literary module, a string, and an idiom field. The newly added genre can only be added to the same record if it has the same semantics as the existing genre. Extracting the sentence pattern library, the cabin model library, the meaning group library, the syllabus in the idiom library, a document module, a string or an idiom and a certain code to form a certain language library , a first language library or a second language library for language translation or text conversion.

<2> Providing a method for making a sentence component, using a bilingual or multilingual text version of the same content as a training sample, and using human-computer interaction to analyze the sentence and sentence levels at the two levels. Words, words, and sentences are equivalent and uniform statement components, including the following steps:

S1. Using the corpus of the bilingual or multilingual text version of the same content as the training sample, each round of A and B bilingual as a sample pair, where the A language is assigned to the pinyin text or the already-matched language, the B language can be assigned. The first round of bilingual analysis of the training samples, in which the B language of the sample is English and the B language is Chinese. From the second round, the new language pair must have been analyzed and compared. When joining Russian, only Chinese-Russian or English-Russian materials can be used as bilingual training samples. In the second round of analysis, the B-language in the sample should be compared with Chinese or English, and B should be new. Russian; each round of training corpus samples should be large enough to add new sentence/sentence ratio < 1% before considering adding new languages, conducting second round of analysis, and, in other respect, industry based on training sample corpus The source or application scope source to mark, divide the sentence pattern library, the cabin model library, the meaning group library, and the idiom library to form the corresponding library for the industry or special version;

52. Analyze the sentence at the sentence level, read the bilingual sample sentence pairs, divide the sentence pattern, sentence cabin, and store the sentence pattern as a sentence structure in the sentence pattern library, and put a small sentence that is not enough to separate sentence patterns and sentence cabins. As a small idiom component, it is stored in the idiom library;

53. Analyze the sentence at the sentence level, compare the sample sentences that have been divided into sentence patterns and sentence cabins, and then take out the contents of the sentence cabin, and then divide the cabin model and the cabin eye into a cabin, and store the cabin model as a cabin module. The model library, the contents of the cabin or the simple sentence cabin aligned by the intention group are stored in the group of meaning clusters as the group of the meaning group; after processing all the sentence cabins, the next bilingual sample sentence pair processing Then, step S2 is performed. The step S2 of the sentence level analysis in the method for fabricating the sentence component further includes the following steps:

521. Read a bilingual sample sentence pair;

522. Calling the sentence-type subroutine to find the sentence type library returns A, B language matching sentence pattern, if no, there is no matching sentence pattern, step S23 is performed, if yes, there is a matching sentence pattern performing step S26;

 523. Taking the current bilingual sample sentence pair as an example, a new sentence pattern is created, and a window is opened, the A statement is displayed on the horizontal line, the B statement is displayed on the horizontal line, and the two command buttons are displayed under the horizontal line, and the sentence type is saved, and the prompt is displayed. The expert clicks, the B-term example of the head and tail of the sentence cabin to be excavated, the digging cabin counter N=0;

 524. When the command to receive the digging cabin is clicked, N=N+1, check whether both A and B are clicked on the two points and whether the two points are valid. If not, the prompt is repeated, if the click is correct And effective, the content between the two points of the B statement is dug and filled in "[N]", the round of the sentence is finished, and the next round of repeating step S24 is to dig a sentence cabin;

525. When the received sentence type command button is clicked and N ^ l, indicating that the new sentence pattern operation is completed, the display of steps S23 and S24 is cleared, and the two new sentence patterns are respectively written as sentence structure members. Type A text and B sentence type fields, if the received sentence type command button is clicked, but N=0, it means that the current bilingual sample sentence pair is not enough to separate the sentence pattern and the sentence cabin and is judged as a small idiom. Clearing the display of steps S23 and S24, and writing two small idioms as small idiom components into the idiom A idioms and B idiom fields; And "the sample sentence sentence that has been divided into the sentence pattern and the sentence cabin is read by the storage unit S3 step, and then the step S21 is performed. The step S3 of the sentence level analysis in the sentence member preparation method further includes the following step:

531. Read a sample sentence sentence that has been divided into sentence patterns and sentence compartments saved by the S26 step;

532. Take the sentence cabin, and then take out the sample sentence sentences that have been divided into sentence patterns and sentence cabins, and use one of the sentence boxes as the current sentence cabin. The upper part of the window displays the sample sentences of A and B, and the lower part shows the 4, B language. The current sentence cabin content; At the same time, the current sentence compartment of the A language is segmented by the word string and sequentially filled in the reference field A language field, and then a word string is sequentially taken to find the A text string field of the meaning group library, and then found and taken out The content of the B-string field of the same record, if the content of the B-string is contained in the current sentence of the B-language, the content of the B-string is filled in the B-language field of the reference table, and does not contain it to be empty; The A text string field has the same record, and the corresponding reference table also has one A language field with a heavy record candidate, which is a complete reference table, the open window two displays the reference table, the group compound word command button, and the groupable compound word operation prompt Accept the expert to click on the reference table and mark it in the marked field of the mark; when the group compound command button is clicked and the reference table has continuous records to be clicked, the content of the A language field with the mark record in the reference table is connected with " _ " group In the compound word, and merge the marked records into one record, the A language field is filled in the compound word, and the B language field is filled in the same semantic string;

533. Determine whether the current sentence cabin is a simple sentence cabin, if it is to perform step S37, if not, further query the cabin model library to determine whether the current sentence cabin contains a cabin model, if not, without the cabin model, performing step S34, if yes, the cabin The module is used as the current cabin model and is incorporated into the current sentence cabin content, and step S36 is performed;

 534. Open window 3 as an editable window, display the current double-sentence cabin content again, accept the expert to write the cabin model, and also display the save mode command button;

535. When the deposit module command button is clicked and the editable window has been edited, the new cabin model also meets the format requirements, and the newly edited A and B language modules are stored as cabin modules in the cabin model A. The mold and B cabin model fields, at the same time, fill the current cabin content into the current cabin model, or fill in the new cabin model as a complex sentence cabin display that has been divided into cabin molds and cabin eyes;

536. Take out the contents of one of the cabins in sequence, and then proceed to step S37;

537. Intention group alignment, the alignment determination command button is displayed under the second window reference table, and the reference table accepts the expert to extend or add the meaning of the word according to the example, and does not change the original word or the word, the length of the string is added or subtracted, and the tape is accompanied by the word. , the modification of the meaning of the inflection of the inflected word, or the modification of the meaning group alignment, or the preferred record; , ,

The group string has been formed, and then the contents of the A and B language fields are stored as the group string component in the A string or the B text field of the meaning group library.

S39. If the current operation is the cabin eye, and the current sentence cabin has no operation of the cabin eye, step S36 is performed until all the cabin eyes of the current sentence cabin are completed, and then the sample sentence sentence sentence sentence and sentence sentence are determined. If there is still an unprocessed sentence compartment in the middle, yes, step S32 is executed to continue processing the sentence cabin. Otherwise, all the sentence cabins are processed, and step S31 is performed to perform the next round of sentence pair operation.

<3>, providing a method for reading a foreign language in a native language based on a statement component, comprising the following steps:

54. Interface, the user specifies the native language and the source language (foreign language, the same below), which is the language of the component library, and divides the screen into upper, middle, lower or front, middle and back windows. For displaying the current operation sentence and related information in the operation, the lower or rear window is used to display the source language to be read, the upper or the front window is used to display the read native language text, and in addition, the repentance operation is displayed on the prompt line. Exit and other command buttons and one or one shift word order buttons, or make them as floating bars immediately below the middle or user can move;

55. The source statement reads in, reads a sentence from the source language as the current sentence is displayed in the middle window, the native text displays the last sentence of the processed sentence, and the source text displays the content of the current sentence.

56. The conversion operation, using the four sentence component library to obtain the conversion operation of the foreign language, the corresponding component of the native language and the Italian code by looking up the current sentence, and giving the content of the native language field of the same record;

57. The sentence is determined, if the entire sentence of the current sentence is processed, the query buffer and the command button are: When the feedback buffer is not empty, the feedback buffer information is added to the source language, the native language, the current source sentence, and the like. E-mail feedback to the support website, clear the feedback buffer, and store the "feedback sentence" flag in the world text cache; when the repent command button is clicked, according to the user's click on the remorse string, the corresponding in the regret selection cache is taken out. The content allows the user to re-select the string and make related changes; when receiving the save exit command, save the contents of the world cache to World, file name = source file name. SJW, if the source text is not finished, the file header is recorded Lower source text offset; when the repentance operation, the save exit command button are not clicked, determine the current sentence is given, perform steps

S5. The method for reading a foreign language in a native language based on a sentence component, wherein the four sentence component library is used to obtain a conversion operation of the foreign language, the corresponding component of the native language and the Italian code by looking up the table, and giving the content of the native language field of the same record. Step S6 further includes the following steps:

S601. Judging a small idiom, querying the source idiom field of the idiom library with the current sentence, if not, not found, executing , , reading the recorded idiom code into the world volume buffer, and then performing step S5;

 5602. Calling the sentence-type subroutine, querying the source sentence type field of the sentence pattern library with the current sentence, if a matching sentence pattern is found, performing step S603, if multiple matching sentence patterns are found, displaying corresponding in the lower part of the middle window The parent sentence type, after accepting the user selection, executing step S603, if there is no matching sentence pattern, the stored sentence library represents the number in the feedback buffer area;

5603. Give the sentence code of the same record, the parent sentence type and the source sentence type. The mother sentence type is mainly displayed in the upper part of the middle window, and the source sentence pair is nested in the source sentence type, and the note type is displayed in the native language of the window. Below, and read the sentence code into the world buffer;

5604. Take the sentence cabin, from left to right, in the middle window, mark the current sentence cabin in the parent sentence type, store the current sentence cabin label in the world text buffer area, and mark and extract the corresponding sentence cabin content of the source language as the current sentence cabin content. , determining whether the current sentence cabin content belongs to a simple sentence cabin, if not, executing step S605, if yes, performing step S608;

 5605. Inspect the matching cabin module, and query the source language module field of the cabin model library with the current sentence cabin content. If a matching cabin model is found, step S606 is performed, and if multiple matching cabin modules are found, the middle window is extended downward. Displaying the corresponding native module in the extension, accepting the user selection and then performing step S606, if there is no matching cabin model, the storage module represents the number in the feedback buffer;

5606. Giving the same recorded cabin model code, mother tongue module and source language module, the mother tongue module is displayed in the extension of the middle window, and the source sentence cabin content is nested in the source language module, and the note is displayed. Below the native module of the window, and reading the cabin model into the world buffer;

5607. The eyes are taken from the left to the right in the native mode, and the current cabin eye is marked one by one on the native language module. The current cabin eye is marked in the world space buffer, and the corresponding cabin contents are marked and extracted. Go to step S608;

 5608. Determine the meaning of the word, read the simple sentence box of the source language or a word string in the cabin eye from left to right, and query the source language string field of the meaning group library. If only one identical word string is found, step S609 is performed, if Find a plurality of identical word strings, respectively take out the contents of their same-recorded native language string fields, back them up in the regret selection cache, and display them in the lower part of the expanded middle window. After receiving the user selection, perform step S609, if the same words are There is no string, and the current source word string is stored in the feedback buffer area;

5609. Take out the content of the currently recorded native language string field and fill in the current parent sentence box or the current native language cabin, and take out the intention group code and store it in the world text buffer; continue to step S608 until the current simple sentence cabin or the current cabin operation is completed. According to the information of the personality loss table, the current sentence or cabin eye loss compensation operation; then correct the current sentence or the cabin's native language word order according to the information of the native language word list; last query →, one move word order button, when → button Clicked to click on the word string of the current sentence or cabin user ,

Before the word string, the word sequence after the move is added to the original word list for use, and step S610 is performed subsequently;

 S610. Judging, if the current sentence compartment is still unprocessed, step S607 is performed. If no, the current sentence and the sentence compartment are not processed, step S604 is performed, and if all the sentence compartments of the current sentence are all processed, the subsequent step S7. The method for reading a foreign language based on a sentence component of a sentence component: wherein the email is fed back to the support website, and when the support website receives the feedback email from the user, the new component is added to the corresponding component library after being processed by the expert in real time, and The new component and related information are fed back to the user in real time, and the original "feedback sentence" flag is replaced by the user's participation; wherein the content of the world text buffer is saved as the world text, in the process of the user reading the foreign language in the mother tongue, the world The text buffer is stored in real time with the code code of the sentence pattern, idiom code, cabin code, and group code, as well as the sentence number, the cabin number, etc., which are saved in the world and read directly in the mother tongue. In foreign languages, after reading, he also produced the world text. As long as one foreign language has been read by one person, the thousands of people in the back can read the world text. Reading the world text is faster than reading the foreign language in the mother tongue, without intervention, and the meaning is accurate. The user-selected user language, the multilingual reading process of Esperanto is only the decoding process, the specific steps are:

1 sequentially remove the code one by one;

2 use the switch statement to classify the code separately;

3 where if it is a sentence number and a cabin eye label, it is used to indicate the current sentence compartment or the current cabin eye;

4 Decompose the sentence pattern, >" code, cabin model code, and meaning group code into a certain record number of a certain library, and give the contents of a certain read field of a certain library or a record. If it is an intention group code, follow the instructions. Given to the current sentence compartment or the current cabin eye;

5 Continue to execute 1 until the end of the text. The method for reading foreign language based on the sentence component of the sentence component is actually one of the applications based on the sentence component, and referring to the coding step of the current sentence using the four sentence component library, and the decoding step of the world text reading, Generate a variety of sentence-based component-based application systems: a method based on sentence component-based world text generation, used to convert traditional text into world text, followed by multi-lingual readout; text processing method based on statement component, for Converting a source text into a textual text, or converting it into a multi-language; a statement-based machine translation method for translating a source language into a target, or translating into a multilingual. Compared with the prior art, the beneficial effects of the present invention are:

1. The statement component storage unit of the statement component device is provided with four libraries for storing sentence patterns, cabin models, and meaning clusters respectively. , ,

A certain code field, used to compile the code. The Italian code not only uniquely represents the common semantics of the same component, but can be decomposed into a certain record of a certain library. Such a design has the beneficial effect that the component and the component can be directly converted or changed by the meaning of the code.

2. The sentence pattern and the cabin model provide a framework for the sentence, which involves a complicated grammar and determines the position of the sentence cabin and the cabin eye. This avoids the prior art using artificial intelligence for syntactic analysis and grammar analysis. The bitterness. Benefits The results are consistent with the original text.

3. According to the industry source or application range source of the training sample corpus, the corresponding sub-library is separated, which is suitable for the industry or special version, and has a support website; it is beneficial to user segmentation and has more beneficial effects for users.

4. The step features of the method for making the sentence component necessarily produce the beneficial effects of the meaning of the sentence component by the equalization and unification of the meanings of words, words and sentences of different language characters.

5. The characteristics of the statement component library are more reliable and more reliable than the electronic dictionary and rule library of the prior art machine translation, and the translation quality and the text conversion quality are more and more reliable; The benefits of improved text quality.

6. The foreign language reading method is one of the applications based on statement components. The foreign language can be read directly in the mother tongue and can be read by everyone. And after reading, it also generates Esperanto; in this way, as long as one foreign language has been read by one person, thousands of people in the back can read the Esperanto - read the multilingual mother tongue without intervention; this has always been a dream of.

7. With website support, it can guarantee the application of product based on statement component, and establish user contact, which has beneficial effects on service quality and version upgrade. BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a schematic structural view of a sentence member device; FIG. 2 is a schematic diagram of a sentence member; FIG. 3 is a schematic diagram of a sentence library; FIG. 4 is a schematic diagram of a cabin model library; FIG. 5a is a schematic diagram of a cluster of clusters (English single string); 5b is a schematic diagram of the Yiqun string library (English complex string); Fig. 6 is a schematic diagram of the idiom library; Fig. 7 is a sentence pattern level comparison flowchart; Fig. 8 is a sentence block level comparison flowchart; The present inventor believes that for computer language text information processing, grammar, voice, and vocabulary are not difficult problems, and the key is that words, words, and sentences in different natural languages are not equal or uniform; Similarly, the computer has to be understood and analyzed like a human being, resulting in "the semantic barrier" still exists today, and the results of the translation are often ridiculous." Sentence structure theory (the discovery of the inventor, the end of the disclosure): We understand that computers can not understand the semantics like the human brain; the length of the computer lies in storage and search. The human brain is just the opposite of the computer, it can understand the semantics, but the storage and search capabilities are far less than the computer. The human brain and the computer have a good complement, but can achieve a good complement, lazy language expression in the computer. Fortunately, the essence of natural language is the meaning of meaning, and the meaning of human beings. All kinds of language and characters are composed of characters, vocabulary is composed of sentences, and sentences are composed of articles; its essential attributes are: sentences are the basic unit for expressing complete semantics; sentences of different languages can express the same semantics. Natural language is the product of the development of the entire human society. When human beings produce language and characters, people are dispersed in many independent societies because of the isolation of time and space. Languages and characters in these many independent societies slowly evolve and develop in their own independent systems. It is precisely because various languages and languages have evolved slowly in their respective independent systems. Between different languages, the voice is different, the grammar is different, and the vocabulary characters are more diverse. This also results in the inconsistency or inconsistency of words, words and sentences between different languages. If the expression of the language is expressed in the computer (representation), the meaning of words, words and sentences can be equal and unified; translation and conversion between different languages is not difficult. How can we make words, words and sentences between different languages and words consistent and unified? The inventor started the study from the sentence pattern, not the sentence pattern on the grammar book, and is a sentence pattern that is convenient for computer operation. After many years of discussion, the theory of sentence components was obtained. Here, the main thesis involved in the statement component theory, and the concept definitions and explanations related to the present invention are as follows: Sentences - In natural language, the basic unit of expressing complete semantics is called a sentence; sentences of different language words Can express the same semantics. Sentences can be divided into two parts: sentence pattern and sentence cabin. A sentence pattern contains at least one sentence cabin. Sentence pattern - abstraction from a type of sentence, relatively stable in the sentence, embodying the basic semantics and generics of the sentence; constitutes the sentence structure of the basic structural framework of the sentence. The sentence pattern shows that the basic semantics and generics of the sentence are oriented to all human beings and cross-lingual; and the basic structural framework is oriented to the specific natural language, and it involves the complex and individualized grammatical phenomena of natural language. The sentence cabin - those flexible alternatives embedded in the basic structural framework of the sentence pattern is called the sentence cabin. The sentence cabin accepts the choice and restriction of the sentence pattern; the sentence cabin can be filled or replaced with a group of meanings to form a colorful and specific sentence. The number of sentences and their semantic content are all-human and cross-lingual; but their position, order and meaning clusters in the basic structure of the sentence structure are specific to the natural language; Grammatical phenomena are also extremely Example sentence sentence sentence explanation (# line number):

1# As long as the sentence pattern (01074) 2 cabins

2# As long as you {1}, you will {2}. 1{Perseverance} 2{success}

3# {1} and you will {2} . 1 {Persevere} 2 {succeed}

4# ECJIH {1}, TO {2} . 1 {BbmacTaHBaeTe} 2{6y eTe c ycnexoM}

5# slightly (other languages)

6# 象一句 (00892) 4 cabins

7# {1} Like {4} {3} {2}? 1{John}4{Henry}3{Effort}2{Work}

8# Does{l} {2}as{3}as{4}? 1 {John} 2 {work} 3 {hard} 4 {Henry}

9# {1} {2} τακ{3} κακ {4} 1 {PaGoTaeT} 2{3⁄4>KOH} 3{ycepHo} 4{ΓεΗΐΐΗ}

10# 略 (Other languages)

11# high sentence type (00922) 3 cabins

{#} of 12# {2} is higher than {3}. 2 let } 1 wake up } 3{ ruby }

13# The {1} of {2} is above {3} • 1 {price} 2 {wisdom} 3 {rubies}

14# {1} {2} BMine {3} . 1{CTO ocTb} 2{MyApocTH} 3{py6nna}

15# 略 (Other languages)

Example 1 1# ₆ #11# The three lines each represent the basic semantics and generics of the three sentence patterns, facing the world and cross-lingual parts; for example, "as long as the sentence pattern" means generic and basic semantics, (01074) Indicates the model number, which is the low-digit decimal number of the Italian code.

2 2~5#, 7~ pendulum, 12~15# represent the structural framework of the three sentence patterns, which are oriented to specific natural language. The left front part of each line is the frame structure of the sentence pattern, the sentence brackets are inside the sentence box; the right rear part is the corresponding sentence box and content examples. 2#7#12# is for Chinese; 3#8#13# for English; 4#9#14# for Russian; 5#10#15# for other languages.

3 The numbers in or above the curly braces are the sentence numbers. The number of sentence cabins (such as 2 cabins for 1# and 4 cabins for 6#), and the semantic meaning of each sentence cabin is world-oriented and cross-lingual; its position, order, and structure in the sentence structure The meaning clusters used for filling are language-oriented (for example, {2} in 7~9# is different in the Chinese-English Russian sentence patterns; the filled meaning group strings are: work, work, JI > κ 0 H ). , ;

The above is a concise example of the definition and interpretation of sentence patterns and sentence cabins. The sentence compartment is filled or composed of a cluster of meanings that are dominated by the group (also preliminarily understood as being filled or composed of a string of words). However, the size of the sentence cabin is quite different. The smallest sentence compartment contains only one cluster of meanings; the largest sentence compartment can contain a clause or clause. We divide the sentence cabin into two kinds of simple sentence cabins and complex sentence cabins: Yiqun group - Yiqun is the equivalence and unity of the "meaning" of words, words, phrases or phrases in natural language; it is the basic unit of human thinking activities. . The meaning of the meaningless group is τ all human beings; it is also metabolized with the development of human dubbing. The meaning group string - Yi Qun in the language text corresponding to the meaning of the group text string, referred to as the meaning group string. The meaning group is divided into single string and double string; only one original string is a single string (as shown in the English string in Figure 5a); consists of two or more original strings, and uses "_" Connected to a complex string (as shown in the English string in Figure 5b). Simple sentence cabin - except for the words that do not express the meaning of the word, no more than three sentence groups are called simple sentence cabins (as shown in Figure 2, 203 ~ 204). English such as "a an the in on to and" is neglected; the Chinese quantifiers are also excluded. Other languages are analogous. Between different languages, these three strings only need to have corresponding, semantically identical strings, and they are not required to be consistent. Complex sentence cabins - larger than the simple sentence cabin, the sentence compartment containing the cabin model is called a complex sentence cabin. The sentence compartments of the example sentences listed above are all simple sentence cabins. The following example sentence contains a complex sentence:

1 {the fisherman} consents to 2 { return the_f eather_suit} , on condition that 3 {fairy dance and play heavenly music for him} . Under the condition of 3 {fairs dancing for him and playing the music of the heavens}, 1 {fisher} Promise 2 {return to the feathers}. This example cabin 1, sentence cabin 2 are simple sentence cabins, sentence cabin 2 English with multiple strings; sentence cabin 3 is larger than the simple sentence cabin, contains the cabin model, is a complex sentence cabin. Cabin and cabin eye - further analysis of the contents of the complex sentence compartment; the frame structure of the sentence pattern is called the cabin model; the replaceable part embedded in the frame structure of the cabin is called the cabin eye. The sentence cabin and the cabin eye are the upper and lower concept; however, the simple sentence cabin and the cabin eye are equal in size, and the same is not more than three meaning clusters except for the unspoken words. The above example boat 3 {fairy dance and play heavenly music for him} 3 {fairy dance for him and play the music of the heavens} Using the sentence pattern theory analysis, you can get the cabin model:

(00205) {l} + {2} +and+ {3} +for him

{ 1 }+ is his +{2}+ and +{3 } where (00205) is the cabin model number; this cabin model contains 3 cabin eyes, and the contents of the three cabin eyes are no more than three clusters: 3 {1 {fairy} 2 {dance} and 3 {play heavenly music} for him}

3 {1 {Fairy} for him 2 {dance} and 3 {playing the music of the heaven} } , . . i

Do? Hello! Get away! Get out! "Wait. Statement component - statement component is the equivalence and unity of words, words, sentences and sentences between different languages. According to the ideology of natural language, human beings interpret and compare pairs of sentences; Unified sentence patterns, cabin models, meaning clusters, and small idioms. The statement component encoded by the library can be another component of the assembled sentence or a standard component that encodes the sentence. The statement component includes a sentence component, Cabin module, esthetic string component and small idiom component. Yitong code——for multilingual, unified coding of interpretive and intercommunicative statement components called Yitong code. Esperanto - generated by Yitong code, reflecting many The language is interoperable, and can be used for multi-text reading or text conversion of special text files. This special file is expected to be used in the world and is called world. Sentence structure theory simplifies the complexity of natural language and adapts to it. Flexibility; and resolve the grammatical inconsistency between them. However, we assign grammar analysis and semantic understanding to the human brain. Principle, the sentence is sentenced, the meaning of the string of two levels of semantic analysis, comparison finishing. These difficult to understand is also a once-and-for-all thing done by the human brain. At the same time the regular, monotonous, cumbersome memory, search, Matching and other work to the computer. Let the computer provide a convenient operation platform, using the form of human-computer interaction, the brain and the computer can complement each other well. The analysis, comparison process generated sentence patterns, cabin models, Statement group and small idioms are used to build the library, and the Yitong code is compiled. The statement component library is generated. The statement component store is the statement component. These statement components are the meaning between multiple languages. Equivalent, can be assembled, spliced into sentences (Figure 2). We can use these components to assemble sentences; we can also use sentences to encode sentences, use the Italian code that matches them to generate World Essay, etc. In the process, the computer only needs to do simple table lookup and judgment; encoding or decoding operations are all possible. The translation or conversion of natural language; or text conversion; etc.; the obtained translation or the converted text is not only readable, but also ideographically consistent with the original text. Hereinafter, the contents of the present invention will be further developed by using embodiments with reference to the accompanying drawings. The description is as follows: 1. A sentence component device FIG. 1 is a schematic diagram of a structure of a sentence component device. As shown in FIG. 1 , the sentence component device includes: a sentence component storage unit 101, an original component 102, an Italian code preparation section 103, and a component read. The ejection unit 104, the component matching giving unit 105, the component adding unit 106, the component library operation control, and the interface unit 107 are seven components:

The ( ) sentence member storage unit 101 is a central component of the device. A two-dimensional database table containing statement components constructed in the form of electronic data and storing multilingual semantic equivalents. They are a sentence library, a cabin model library, an Italian group library and an idiom library (see Figures 3 to 6).

1. A sentence pattern library 300 for storing sentence structure components, having a sentence pattern, an English sentence pattern, a Chinese sentence pattern, a Russian sentence . ,

The sentence patterns are stored in the corresponding sentence pattern field 301. The sentence pattern described in this sentence actually refers to the framework part of the sentence pattern, which is oriented to each natural language. The curly braces indicate the sentence cabin, the middle number is the number of the sentence cabin, the sentence cabin is filled by the meaning group, the position, the order of the sentence cabin in the sentence pattern and the filled meaning group are all oriented to the natural language; It can be seen from the content 301 in the library of Fig. 3 that the same sentence cabin has the same label but its position and order in the sentence patterns of the various languages are not consistent. The sentence pattern field stores the sentence pattern code, and the sentence pattern code represents the semantic meaning of each sentence pattern in each sentence type field in the same record. The sentence pattern reflects the basic semantics and generics of the sentence is all-human and cross-lingual; it contains the number of sentence cabins and the meaning of the sentence cabin are all human-oriented, cross-lingual; human-oriented, cross-lingual representation Is the sentence code. That is to say, the sentence code represents the sentence form semantics, which insinuates the sentence patterns of each language; each sentence pattern can infer another sentence pattern through the sentence pattern. As for grammar, which belongs to various natural languages, the framework part of the sentence model is a complex and individualized grammatical phenomenon of natural language. However, even if there is a grammatical phenomenon in the sentence cabin, it is extremely simple.

2. The cabin model library 400, used to store the cabin module, has the cabin model code, the English cabin model, the Chinese cabin model, and the Russian cabin model field, as shown in Figure 4. It contains at least one record, the same semantics of the same model, and the corresponding model of the cabin is stored in the corresponding language model field 401. The cabin model is the frame structure part of the complex sentence cabin and is oriented to the natural language. The square brackets indicate the cabin eye, the middle number is the number of the cabin eye, and the cabin eye is also filled with the cluster. The position, order and filling of the cabin eyes in the cabin mold are all oriented to the natural language; as can be seen from the contents of the library in Figure 4, the same cabin eye is labeled the same but it is in each language class. The position and order in the mold are not consistent. The cabin model code field stores the cabin model code, and the cabin model code represents the semantics of each type of cabin model in each cabin model field in the same record. The basic meaning of the cabin model is for all human beings, cross-lingual; it contains the number of cabin eyes and cabin language for all human beings, cross-lingual; its expression is the cabin model code. That is to say, the cabin model represents the semantics of the cabin model, which insinuates the cabin models of each language; each type of cabin module can infer another type of cabin module through the cabin model code. As for the grammar in the sentence cabin, it belongs to the natural language. The cabin model covers the grammatical phenomenon of natural language. However, even if there is grammatical phenomenon in the cabin, it is extremely simple.

3. The meaning group library 500, 502 is used to store the meaning group component, the intention group code, the English string, the Chinese string, the Russian string field, as shown in Fig. 5a-b. It includes at least one record, the same semantic group of the same group of records, and the meaning group of the corresponding language is stored in the corresponding text string field 501, 503. The meaning cluster is the content of the sentence cabin or the cabin eye. The sentence cabin and the cabin eye are the concept of the upper and lower position. The sentence cabin is divided into a simple sentence cabin and a complex sentence cabin. The complex sentence cabin is extracted as a sentence structure and is the cabin eye. The sentence cabin and the cabin eye are the upper and lower concept; however, the simple sentence cabin and the cabin eye are equal in size, and the same is not more than three meaning clusters except for the unspoken words. The pinyin string has a single string and two strings. The single string is an original word string 501. The compound string is composed of more than one original word string and is connected by " _ " to form 503. The meaning group code field stores the meaning group code, and the meaning group code represents the semantic meaning of each group of meaning group strings in each group of the meaning group strings in the same record, which is oriented to humans and cross-language; For all natural languages. That is to say, the meaning group code represents the semantic meaning of the group of meaning groups, and the group of meaning groups of each type of text is inferred; each group of meaning groups can infer another group of meaning groups through the group code. , , , , ,

The idiom field, as shown in Figure 6. It contains at least one record, a small idiom of the same semantics is co-located, and a small idiom of the corresponding language is stored in the corresponding idiom field 601. The idiom code represents the semantics of the small idioms of each language in the idiom fields of the same record. That is to say, the idiom code represents the semantics of the small idioms, and the small idioms of each language are inferred; the idioms of each language can insinuate the idioms of another language through the idioms. The structure of the above four libraries emphasizes that only the same components of the same semantics are in the same record, and the same record is designed with a certain code field for compiling the code. The Italian code is inline with the same components of the same record. Such a structure ensures that the component and the component can be directly converted or changed by the Italian-style code conversion; that is, different languages can be converted into each other. The relationship between the above four libraries is parallel, and they do not interfere with each other and are shared among the statement component storage units. All operations or controls of other components are accepted.

(2) The original department 102 stores the index files related to the above four libraries; it also includes the original CPU and the like.

(3) The code generation unit 103 is connected to the sentence component storage unit 101 and the Italian code preparation unit 103. Only when any of the above four libraries has a new record, the current library representative number: for example, the idiom library = FF00H, sentence library =?00011, cabin library = EF00H, group library =0001H (also the library label and numerical segment of the four libraries; the starting point of the segment, the last library number -1 ) Make the high word plus the current library record number to synthesize the code; and fill in the code field of the current library, as the unified two-byte fixed-length multi-lingual semantic interoperability code of the statement component. The representation of the same semantics of the current library and the components of the language in the current record is unique;

(4) The component reading unit 104 is directly connected to the sentence component storage unit 101, and is configured to receive the read command, and determine a certain record of the library by using the number of segments included in the code, that is, the number of the code of the passbook and the minimum number of the library. = The record number of the library indicated by the minimum library number. Then read the required language components to the corresponding records in the corresponding library.

The component matching matching section 105 is directly connected to the sentence component storage section 101 for receiving a matching command. According to the sentence or sentence content of the given language and the guidance of the current operating point, the corresponding language records are searched in the corresponding language index field of the corresponding component library, and the required language components are matched. If there is no matching record, no match signal is returned.

(6) The component adding unit 106 is directly connected to the sentence member storage unit 101. Used to receive the Add New Component command. After the query confirms that the corresponding component library does not have the same component, the new component is added to the corresponding language component field of the corresponding component library. When a new component is added to a new record, the information is simultaneously notified to the code generation unit 103.

(7) The component library operation control and interface unit 107 is connected to the sentence component storage unit 101 by the component reading unit 104, the component matching operation unit 105, and the component adding unit 106. It is used to receive calls from various applications based on the statement component or to receive related commands, return the caller required statement components, or connect to other application devices based on the statement component through the interface. The statement component described above (see FIGS. 2 and 3 to 6) is another component for assembling a linguistic sentence, and is also a standard component for splitting and encoding another sentence. There are four types as follows: , , .

The genus of Italians also determines the order and number of sentences in the sentence, and it involves the more complicated grammatical phenomena of such sentences.

2. Modules 202, 401, used to form the basic structural framework of a complex sentence compartment. It represents the basic semantics of this type of sentence cabin, and also determines the number and number of cabins contained in such sentence cabins, and covers the more complicated grammatical phenomena of such sentence cabins. Both sentence patterns and cabin modules provide a framework for sentences, which involves a complex grammar and determines the order of the sentence cabins and cabin eyes. This avoids the use of artificial intelligence for syntactic analysis and grammatical analysis. bitter. There is a contribution to the results that can be consistent with the original text.

3. The group of characters 501 to 503 is a component that is played by the group of meanings. For filling the simple sentence cabin 203~204 or the cabin eye 205~207, the simple sentence cabin and the cabin eye are the same as the upper and lower concept, and they are not more than three meaning clusters except for the vocabulary;

4, the small idiom component 601, by the sentence is too short to separate the sentence pattern, sentence sentence as a small idiom component. Used to form short sentences directly. The sentence component device described above, in addition to the corresponding corresponding language (English, Chinese, Russian), for each additional language, the sentence library, the cabin model library, the Italian group library, the idiom should be first used. The library respectively adds a certain sentence pattern, a document module, a string, and an idiom field. And the newly added components can only be added to the same record if they have the same semantics as the existing language components. That is, it is emphasized again that only statement components of the same semantics can share a record. For the above four libraries, only two fields can be extracted to form a corresponding sub-library; that is, a sentence-type library, a cabin model library, an Italian group library, a sentence pattern in a corpus library, a document module, a string or An idiom and a certain code form two fields, which constitute a certain language library or a first language library and a second language library, and are applied to a language translation or text conversion. The source of the statement component described above, one is through the expert operation, human-computer interaction, analysis and comparison of the bilingual training sample corpus. Second, another source is that the user's feedback information is added after being reviewed by experts; it is implemented through the support website. Second, a method of making a statement component: The method of making a statement component: 1 Prepare a sample corpus, and take the corpus of the bilingual or multilingual text version of the same content as a training sample. The human-computer interaction method is used to first perform the analysis of the sentence-level analysis; then, the analysis is performed on the sentence-level level. Thus, the word, word, and sentence meanings are equivalent and unified statement components, including the following steps:

() Prepare sample corpus and use corpus of bilingual or multilingual text versions of the same content as training samples. Each round compares A and B as a sample pair. Where the A language is assigned to the pinyin text or the already compared text The first round of bilingual analysis of the training samples. Among them, the B language of the bilingual sample is English, and the B language is Chinese. From the second round, the new language must have been analyzed and analyzed. For example, when joining Russian, only Chinese-Russian or English-Russian materials can be taken as a bilingual training sample. The second round of comparative analysis of the bilingual pair of samples should be compared with Chinese or English, and B should be new. Canadian in addition. This feature is emphasized above. It is a powerful measure to ensure that the same semantic components are shared together. The training corpus sample for each round should be as large as the new sentence/sentence ratio < 1% before considering adding new languages; for example, during the operation, one working day is counted, and the new sentence number is divided by the new one. The ratio of sample sentence pairs is < 1%. After 1%, consider the next round of analysis. On the other hand, it is marked according to the industry source or application range source of the training sample corpus. By dividing the sentence pattern library, the cabin model library, the meaning group library, and the idiom library into a number of corresponding sub-libraries. These sub-libraries are used in the respective industry or proprietary versions. Plus a support site; useful for user segmentation; also beneficial for version upgrades. It is a natural matter to collect the above-mentioned corpus and, if necessary, to enter it.

(2) Sentence level analysis comparison Figure 7 is a sentence level comparison flowchart. Analyze the sentence at the level of the sentence. Read bilingual sentence pairs, divide sentence patterns, sentence cabins, store sentence patterns as sentence components in sentence patterns, and put small idioms that are not enough to separate sentence patterns and sentence cabins into idioms as small idioms. At the same time, the bilingual sample sentence pairs that have been divided into sentence patterns and sentence cabins are saved, in order to further compare the sentence cabin level. The specific steps are shown in Figure 7: Start, first read a bilingual sample sentence pair 701. Then, the sentence pattern word program 702 is called, and the sentence pattern library is returned to return 4, and the B language matches the sentence pattern. A matching sentence pattern 703 is judged. If yes, there is a matching sentence pattern, and the following pattern is matched with 707; if no, there is no matching sentence pattern, and the sentence box type operation 704 is continued. That is to make a new sentence pattern by taking the current bilingual sample sentence pair as an example, popping up a window, displaying the A statement on the horizontal line and the B statement on the horizontal line, and displaying the two command buttons of the sentence box and the sentence pattern under the horizontal line, and prompting the expert Click the head and tail points of the sentence box to be excavated in the A and B example sentences, and dig the cabin counter N=0. When the digging cabin command button is clicked, set N=N+l to check whether both A and B words are clicked and whether the two points are valid. Pinyin text between two points ^ a word string, ideogram ^ a word is valid. If no, prompt to redo; if the click is correct and valid, the content between the two points of the A and B statements is dug and filled in "{N}", the round of the digging cabin ends, the next round of repeated digging cabins The sentence operation 704 then digs a sentence cabin. When it is judged that the sentence type 705 command button is clicked and N^ l, it indicates that the new sentence pattern operation is completed. Clear the corresponding display above, and save the sentence pattern or save the small idiom 706. Add two new sentence patterns as sentence structure components to the sentence database A sentence sentence type, B sentence sentence type field. If N=0 at this time, it means that the current bilingual sample sentence pair is judged as a small idiom because it is not enough to separate the sentence pattern and the sentence cabin. Then, to clear the corresponding display, the two small idioms are added as idiom components to the idiom A idioms and B idiom fields. The continuation type 707, the current bilingual sample sentence number is filled in the current matching sentence pattern, or filled in the current new sentence pattern, as a sample sentence sentence that has been divided into sentence patterns, sentence compartments Compare the readings. This step ends. next , . A preferred example of a sentence-type word program is: Change the sentence pattern to a sentence string in advance (such as "Does {1} {2} as {3} as {4}?" becomes " does as as ") , then from the left to the right, English (Pinyin text) word by word, medium (non-phonetic text) one by one word, to find the sentence type first word or first word; set the matching sentence pattern in the temporary library. Then use the loop statement to test the sentence type by sentence. In the loop, a switch statement is set, and the number of the sentence type string is used as a switch statement (such as a sentence string " does as as " = 2 empty 3 segments), and each segment of the sentence string is sequentially compared with the example sentence. Both are discarded after the comparison (the same has no effect). When each segment can be found in the example sentence, it is the selected one, that is, the sentence pattern matching the example sentence, and the list is given. If there are several sentence patterns, select the sentence pattern with the longest sentence string. If there is a case where the string length is the same, manual intervention is accepted. The following is a detailed explanation of the above steps: Start, read a bilingual sample sentence pair 701, if the sentence pair is "Does John work as hard as Henry? John worked hard like Henry? ": Call the sentence-type word program 702, find the sentence pattern library returns 4, B-word matching sentence pattern. A matching sentence pattern 703 is determined. If no, there is no matching sentence pattern, and the sentence box operation is continued. That is to say, a new sentence pattern is created by taking the current bilingual sample sentence pair as an example. Pop up a window, the horizontal line shows the A statement, "Does John work as hard as Henry ?", the horizontal line shows the B statement "John worked hard like Henry?", and then displayed the digging sentence cabin, the sentence sentence type Command buttons, and prompt the expert to click on the first and last points of the sentence box to be drilled in the A and B example sentences, and dig the cabin counter N=0. When the digging cabin command button is clicked, set N=N+l to check whether both A and B words are clicked and whether the two points are valid. Such as "Does I John work as hard as Henry ?,, " | John I work hard like Henry? "(' Ι ' is the clicked point.) Their clicks are correct and valid, and the content between the two points of the Β statement is dug, then N=l is filled with "{1}", which is "Does {1} Work as hard as Henry ? "" {1} Work hard like Henry? ". The sentence-type command button is not clicked; the next round of repeating the sentence-sentence operation 704 and then dig a sentence cabin; such as "Does {1} work | as hard as Henry? "" {1}Working like I Henry I work I? "N=2, fill in "{2}", for " Does {1} {2} as hard as Henry ? "" {1}Working like Henry {2}? "If the save sentence command button is not clicked; continue, N=3, N=4. When is "Does {1} {2} as {3} as {4} ? ", " {1} Like {4} {3} {2}? When it is judged that the sentence type 705 command button is clicked and N ^ l, it indicates that the new sentence pattern operation is completed. The above corresponding display is cleared, and the sentence pattern 706 step is performed; the two new sentence patterns are used as the sentence pattern. The widget "Does {1} {2} as {3} as {4} ? "Add a sentence pattern A sentence sentence type field; " {1} like {4} like {3} {2}? "Add a sentence pattern B sentence sentence type field. Then follow the example with 707 steps, the current bilingual sample Fill in the current new sentence pattern, as a sample sentence pair that has been divided into sentence patterns and sentence cabins, such as "Does 1 {John} 2 {work} as 3 {hard} as 4 {Henry} ? 1 {John} Like 4 {Henry} 2 {Effort} 3 {Work}?", and save it, the profiling of the statement level is compared to the reading, this step ends. The next bilingual sample sentence begins, then Execute reading the bilingual sample sentence pair 701. If the sentence pair read above is "How do you do?", "Hello! ", when it is judged that the sentence type 705 command button is clicked, then N=0, indicating that the current bilingual sample sentence pair is insufficient to separate the sentence pattern and the sentence cabin and is judged as . , . you do?" Join the idiom A idiom, "Hello! "Add B language idiom field. Continue to type 707, this step ends. The next bilingual sample sentence pair begins, then read the bilingual sample sentence pair 701. The main point of sentence pattern comparison is to dig the sentence cabin Sentence pattern, how to dig, how to make a sentence pattern. The requirement is to pursue the representativeness under the premise of ensuring operability. The operability means that the computer does not need to understand and analyze the operations such as table lookup, judgment, storage, etc. as described above. The representative sentence pattern covers the number of sentences, which can cover the more representative the better the sentence. The principle measures required by the experts in the sentence-level comparison process are as follows:

1 Multilingual vs. Semantic Considerations Semantic considerations from multilingual pairs, at least bilingual pairs. If the conditions permit, it is natural to take as many words as possible while extracting the sentence; it is because it is impossible to ask at least a bilingual pair. Such as:

We used to go to the movies about once a week. Usually we go to see a movie about once a week. In this sentence, English can use "go to the movies" as a sentence cabin, but the Chinese corresponding "watching movie" has "one time" inserted in the middle. " Once a week" as a sentence cabin, Chinese "once a week" is separated by other words. Neither of these situations can be done, and it must be considered in terms of bilingual semantics. This sentence pair can:

1 {We} used to 2 {go to the movies about once} a week. Usually 1 {us} 2 per week {about to watch a movie}. 2 representative consideration principle The number and size of the sentence cabin directly affect the representativeness of the sentence pattern. Our principle is to pursue representation under the premise of ensuring operability. Regarding the size of the sentence cabin, it is explained below. Here is a brief explanation of the representative problem:

How many are there in your fami ly? How many people are there in your family? If you only use "your fami ly" as a sentence cabin; you can use "his fami ly; John's fami ly; your class" and so on. But because "How many" is related to "family" semantics. Chinese asks "a few people" for "home"; but "how many students" or "how many people" should be asked for "class". This only makes "your fami ly" a poor representation of the sentence cabin. If you make "How many" and "Fami ly" into two sentence cabins, you can not only take care of each other in terms of semantics; but also increase the representativeness.

3 Simple and accurate consideration principle In view of the positioning of the Yitong text is "simple and accurate transfer of semantics", when the digging of the sentence cabin is not able to resist the time, you can modify the gorgeous translation of the sentence into a simple literal translation, and then dig the sentence. Such as: The soldiers are not deceptive. The translation of the sentence "Bold is not deceptive" is both gorgeous and concise, but the sentence collection is difficult to operate. It is not too much to change Chinese to a simple translation. "There are no more tricks in the war."

There can never be too much 1 {deception} in 2 {war} . 2 {war} More than 1 {诡} is not too much.

For al l their great size , the elephants moved absolutely noiselessly . Although the body of the elephant is huge, it does not sound at all. This "no sound at all"; it is difficult to implement into a sentence pattern or a sentence cabin; change it to a simple literal translation "walking but quietly" The problem is solved:

For al l their 1 {great size} , 2 {the elephants} 3 {moved} absolutely 4 {noiselessly} . Although 2 {image}'s 1 {body is huge}, 2 {it} 3 {walking} up 4 but Quietly}.

4 Principles of grammar and simplification From the grammatical point of view, complex and individualized grammatical phenomena are taken into the sentence pattern; the grammar in the sentence cabin is extremely simple. The example above is not difficult to understand. In the specific operation, the number of sentence cabins can be appropriately increased to reduce the complexity of the sentence cabin, and the sentence compartment should be minimized (see below).

1 {She} never 2 {comes} but 1 {she} 3 {brings something for the chi ldren} .

1 {She} Not once 2 {耒} Not on 3 {Bring something to the children}. If you increase the sentence cabin, the sentence {3} is changed to {3} {4} complexity is reduced. Such as:

1 {She} never 2 {comes} but 1 {she} brings 3 {something} for the 4 {chi ldren} . 1 {She} Not once 2 {耒} Not for 4 {children} Bring 3 {- Something}.

The principle sentence of 5 part of speech and replaceability is the part that can be replaced by other words. The more alternative words, the stronger the substitutability; the grounding makes the representation of the sentence form stronger. The part of the vocabulary in the sentence cabin is limited to numbers, nouns, adjectives, and plurals. In a few cases, other words (such as verbs, adverbs, etc.) are considered. If you want to give priority to the class of words in the sentence cabin, then the first is the string, the special string, the second string, the adjective string..., and finally consider the verb string. The least considered are prepositions and conjunctions. That is to say, prepositions and conjunctions are almost all part of the sentence. It is very simple to require the sentence pattern collection to reach the sentence grammar, and the complex grammar phenomenon is all in the sentence pattern; the sentence patterns left after the sentence cabin is dug, the sentence patterns should not be too small, because there are too few inconvenient sentence types. Out. The ideal situation is that there is a sentence before and after each sentence, that is, there is no continuous sentence. The principle is that any sentence pattern, the frame structure in any language must have one or more text strings as sentence patterns. It is not allowed in a multilingual pair, a certain language does not even have a sentence type, because Chinese is the most succinct, this happens from time to time - but it has to be reworked and must be avoided. Sometimes there are several consecutive sentence compartments, that is, continuous sentence cabin problems. There is also the size of the sentence cabin, as little as possible for the big sentence cabin problem, these are not too relevant to the sentence patterns, these conditions are explained under the relevant headings.

The principle of avoiding more than 7 consecutive sentence cabins

1 {1} 2 {get to work} at 3 {nine o' clock} every morning. 3 {9 o'clock} 1 {{}} {start work} every morning. In the previous example, there were consecutive {1} {2} two sentence cabins; the corresponding Chinese became {3} {1} {2} three consecutive sentence cabins. Two or more sentence compartments are connected to each other as consecutive sentence compartments; three or more sentence compartments are connected to each other as multiple consecutive sentence compartments. Continuous sentence cabins not only have the disadvantage of too few sentence patterns, but also have to intervene manually. In particular, three or more consecutive situations should be avoided. The above example is reduced to two sentence cabins without this consideration:

I 1 {get to work} at 2 {nine o' clock} every morning. Every morning 2 {9 o'clock } I 1 {start work}.

8 The principle of making the sentence space as small as possible. The sentence cabin is large and small. The smallest sentence box contains only one group of meanings; the largest sentence box can contain a clause or a clause. When we are making sentences, we should try to make as few sentences as possible. So, how to master this as little as possible, what is the scale? Take "only this" as the standard. For example, the following sentence pairs:

Can, you guess 1 {what I was doing} 2 {this morning} ? Can you guess 2 {this morning} 1 {What am I doing }?

I have forgotten 1 {what time} he said he 2 {had dinner} 3 {last night} . I forgot that he said he 3 {last night} was 1 {when} 2 {eat dinner}.

What were you doing when 1 {I cal led you on the telephone} ?

1 What do you do when I call you }?

I have forgotten 1 {what he said his address was} . The above four sentence pairs, respectively, by a when, lead a clause. The first pair can separate the subject-predicate and the time-sentence into two sentence cabins. The second pair can be made into three sentence cabins. The third pair can't separate adverbials, can only be made into a sentence cabin; the fourth pair of questions on the topic, can not be separated, can only be made into a sentence cabin. The third, the four pairs are "only like this", which is the scale to be grasped in the lesser case.

9 The principle of verification after the excavation of the sentence cabin After the excavation of the sentence cabin, it must be carefully examined: the sentence pattern, and even each sentence cabin will be examined separately. First look at the sentence meaning, must come from the sentence form, and there is no adhesion to the sentence cabin content. Each sentence is then validated, they must be replaceable, separate from the sentence, and not confined to the sentence. If a sentence box is stuck with a sentence, it must be modified. E.g:

When do you think 1 {the meeting wi l l be held} ? What do you think 1 {when is the meeting held}? In this way, the sentence type and the sentence cabin are incorrect, and the sentence content "when" has a glue with the sentence pattern "When". Should be repaired to:

When do you think 1 {the meeting} wi ll be 2 {held} ? When do you think 1 {conference} will be 2 {convened}? (3 sentence compartment level analysis comparison Figure 8 is the sentence cabin level comparison flow chart. As shown in Figure 8, start running, read the bilingual sample pair 801 that has been divided into sentence patterns and sentence cabins. , group compound word 802 steps, sequentially take out the sentence sentence form, the sentence sentence of the sentence sentence to one of the sentence cabins, open the window one shows the A, B language sample sentence example, the lower part shows the A, B language current sentence cabin At the same time, the current sentence compartment of the A language is segmented by the word string and then filled in the A language field of the reference table in turn, and then a word string is sequentially taken to find the A text string field of the meaning group library, and the same record is taken out after being found. B string field content. If the content of the B string is contained in the current sentence of the B language, the content of the B string is filled in the B field of the reference table, and does not contain it to be empty. Make a complete reference table, open window 2 Display reference table, group compound command button and groupable compound operation prompt. Accept the expert to click the reference table and mark it in the marked record field. When the group compound command button is clicked and the reference table has continuous records, click There are signs in the reference table The contents of the recorded A language fields are connected by " _ " to form a compound word, and the marked records are merged into one record, the A language field is filled in the compound word, and the B language field is filled in with the same semantic word string. Cabin 803, if it is connected to the group alignment 808; If not, query the cabin model library to determine whether the current sentence cabin contains the cabin model 804. If the cabin module is not executed to write the cabin module 805, open window three as an editable window, the current double statement The contents of the cabin are displayed again, and the expert is used to write the cabin model. The display module command button is also displayed. If the cabin module is included, the cabin module is used as the current cabin module. Steps: Divide the cabin mold and the cabin eye 806. The save mode command button is clicked and the editable window has been edited , ,

 In the case of the syllabus, the current sentence space is filled into the current module, or the new model is filled as a complex sentence display with the cabin and cabin. Continue to take the cabin eye 807, and then take out the contents of one of the cabins in turn, and then perform the steps of 808 alignment. In the second window reference table, the alignment determination command button is displayed. The reference table accepts the expert to extend or add the meaning of the word according to the example, does not change the original word word, and adds or subtracts the string length, the adhesive tape accompanying word, and the word form change supplementary word meaning item. A modification of the equivalent group alignment, or a preferred record. When the alignment determination command button is clicked, it indicates that the reference string of the reference table Α Β, Β 已经 has been intentionally aligned, that is, the group string has been formed, and the escrow group 808 operation is performed, and the contents of the Α and Β 字段 field are recorded as records. The group string component is stored in the string or the string field of the meaning group library. If the current operation is the cabin eye, the judgment eye is finished 809, no, if the current sentence compartment has no operation of the cabin eye, then perform the cabin eye 807 step; until all the cabin eyes of the current sentence cabin are completed. After the judgment of the sentence cabin is completed 810, it is judged whether there is an unprocessed sentence cabin in the sample sentence sentence sentence sentence sentence, and there is still an unprocessed sentence cabin, and the sentence box 802 step is executed; The sentence cabin. If all the sentence cabins have been processed, perform the re-reading sentence pair 801 step, and read in the next bilingual sample sentence pair that has been divided into sentence patterns and sentence boxes. Go to the next round of operations. The following is a detailed example to illustrate the above-mentioned process: Start running, read in (sentence level comparison) has divided the sentence sample, sentence cabin bilingual sample pair 801, for example "1 {the fisherman } consents to 2 { return the_feather_suit}, on condition that 3 {fairy dance and play heavenly music for him} .,,, " Under the condition of 3 {fairs dancing for him and playing the music of the heavens}, 1 {fisherman} promised 2 {return the feathers}. "There are 1, 2, 3, and two sentence cabins for each example sentence. Execute the sentence box, display, group compound word 802 steps, and then take out the sentence sentence sentence sentence sentence sentence sentence sentence sentence to one sentence Cabin, the upper part of the open window shows the sample sentences of A and B, and the lower part shows the current sentence of the 4th and B. At the same time, the current sentence of the A language is divided into the word string and filled in the reference field A field. Then, take a word string to find the A string field of the meaning group library, and then find out the content of the B string field of the same record. If the content of the B string is included in the current sentence box of the B language, the content of the B string is Fill in the reference table B language field, does not contain it to be empty. Make a complete reference table, open window two display reference table, group compound word command button and group compound word operation prompt. If sentence box 2 is the current sentence cabin, The current sentence A is "return the feather suit"; B is "returning feathers"; the reference table at this time is:

A i my field: return the feather suit

B-language field: Return V reference table B has three records that are empty, but corresponding to "feather" has "feather"; corresponding to "suit" has "clothes"; and these three records have the "" mark of the expert click . And the compound word command button has been clicked, and the contents of the A language field with the mark record in the reference table are connected by "_" to form a compound word, and the marked record is merged into one record, and the A language field is filled in the compound word, B The language field is filled in with an equal semantic string. At this point the reference table becomes: — —

B language field: Returning the feathers now, the current sentence A is "return the_feather_suit"_; B is "returning the feathers"; the subsequent judgment is whether the simple sentence 803, is, the continuation of the group alignment saves the group 808 step; has a meaning of aligned, but are intended to save string group, the "return", "return", is intended to join the group string library A, B text string field, as shown in FIG. 5A; the "the- feather_ _S uit", "plumage" added The meaning group library A, B text field, as shown in Figure 5b. No, if the current sentence compartment is a sentence cabin 3, not a simple sentence cabin, it is judged that there is no cabin module 804, and the cabin model library is judged whether the current sentence cabin contains a cabin model. No, without cabin model, execute the step of writing cabin module 805, open window three as an editable window, and the current double statement cabin content "fair dance and play heavenly music for him,,""The fairy dances for him and plays the music in the sky. "Re-display, accept the expert to write the cabin model, and also display the storage module command button. If the cabin module is included, the cabin module is used as the current cabin module: the cabin mold and the cabin eye 806 are divided. The cabin mode command button is clicked and the window can be edited, such as "{1} {2} and {3} for him", "{1} is his {2} and {3}", has been edited The new cabin model also meets the format requirements, and the newly edited A and B language modules are stored as cabin modules in the cabin model A and the B cabin model fields, as shown in Figure 4. The sentence cabin content is filled in the current cabin model, or filled into the new cabin model, such as: "1 {fairy} 2 {dance} and 3 {play heavenly music} for him", "1 {fairy} for him 2 { Dancing} and 3 {playing the music of the sky} "As a complex sentence cabin display that has been divided into cabins and cabins. This complex sentence Contains 3 cabin eyes. Continue to take the cabin eye 807, take out the contents of one cabin eye in turn, and then perform the step of performing the group alignment 808. In the second window reference table, the alignment determination command button is displayed, and the reference table accepts the expert extension by example or Adding the meaning of the word, changing the length of the string, adding the tape, attaching the word, changing the word form, adding the meaning of the word, etc., or modifying the meaning of the group alignment. The three cabins A and B have The corresponding original word string does not need to be explained. Among them, the "play" Chinese original dictionary only has "game, competition, sports, gambling, script; play, play, play, play, play", etc. without "playing" acceptance. The expert "extends or adds words according to the example" plus "play" and "play". The current cabin eye is the cabin eye 3. The reference table is:

A language field: play heavenly music

B-language field: Playing the music of the sky When the alignment determines that the command button is clicked, indicating that the word string of the A and B words in the reference table has been formed into a group string, the save meaning group string 808 operation is performed, and the A and B language fields are recorded one by one. The content is stored as an A-string or a B-string field of the meaning group library as a group of meaning strings. If the current operation is the cabin eye, the judgment eye is finished 809, no, if the current sentence cabin and the cabin eye are not operated, the step of taking the cabin eye 807 is performed; until all the cabin eyes of the current sentence cabin are completed. After the judgment of the sentence cabin is completed 810, it is judged whether there is still an unprocessed sentence compartment in the sample sentence sentence sentence sentence sentence, and there is still an unprocessed sentence cabin, and the execution of the sentence cabin 802 step; The sentence cabin. If all the sentence cabins have been processed, perform the re-reading sentence pair 801 step, and read in the next bilingual sample sentence pair that has been divided into sentence patterns and sentence cabins. Go to the next round of operations. The sentence level comparison operation is as described above for "group compound words", and the original word string is connected by "-" into a compound word (multiple string); ;

In case of circumstances, experts are required to master:

1 When the direct word translation can't be expressed, the group of the meaning group is composed of you are gone. The literal translation should be "you (yes) go"; "the" comes from the past tense, so the composition of the string "you- were- gone, you are gone. ,, cal l-on- me visit me knew_nothing_about_it know nothing about compel led-to-go non-de-show-self-in_his- true- colours

2 succinctly aligned, complex representatives form a series of "per-mu grain yield, per mu yield of grain"; in the middle of "mu production." Put the British into a series of "per-grain-yield", and look at it in the middle. Another example:

Late-at-night late night down-to-the-countryside to the countryside fight- it-out_to_the-end

3 escaping means a thing, or a combination of regular fel l_asleep falling asleep know-nothing - no knowledge of come-back - from_the_front under the line the - sweat - was - pouring - down sweaty; pictures - it_have_just_taken Proximity put_my_f inger_on indicates

4 words have coincidence or repetition of the composition of the string very-wel l is good; doubts_of_questions doubt; bear-fruit results; the-far-distance far; after the above group of compound operations, often in a sentence cabin, part of the string Formed a complex string, making some seemingly complicated sentence cabins into simple sentence cabins. Group compounding is also one of the measures of ensemble alignment and should be applied flexibly. The principles and measures that the expert requires in the operation of the sentence-level alignment operation as described above are as follows:

1 in accordance with the example, extension or addition of meaning 2 {Applied} It is impossible to {{perfect}. Among them, the meaning of "useful" has "useful, helpful, and beneficial", and it is not suitable to choose them; add the word "applied".

1 {She} was 2 {strong} , for al l 1 {she} was so 3 {smal l} . 1 {3⁄4} Although 3 {thin small}, but very 2{strong}. Such as: where small small, add "skinny" word meaning.

I am very i ll. I am seriously ill. Add the verb "am" to the meaning of "affected".

Learn the truth To understand the truth, Learn adds the word "understand".

2 Adding or subtracting the length of the string without changing the original word or word, easy to splicing

I ask you to teach me every other day. I ask you to teach me every other day. Where 〃 teach〃

V has the meaning of "teaching, teaching"; the length of the word string is "teaching" for splicing, and the meaning of "teaching" is added.

Sttend school; sttend force "people,, word meanings, brief interpretation of "power mouth people"; school power mouth "learning," meaning, "school".

3 sticky tape accompanying the string of words such as "good good" sticky into "benefits, good things, kindness, easy to use", etc. (Chinese with the accompanying words, the latter is handled separately). For example, "word word" sticks into "word" and so on.

4 The change of the word form adds the meaning of the word (the semantics of the expression of the word segmentation, comparative level, etc.), so that the word form analysis and processing can be omitted. Has been added "still, how" word meaning; punished added "accepted"meaning; best added "best"meaning; had added "previous"meaning; done added "completed, completed"meaning; vi l lage^ t has supplemented the meaning of "multiple rural"; third, a method of reading foreign languages based on sentence components. Figure 9 is a flow chart of foreign language reading methods based on sentence components. As shown in Figure 9, preparations for some interfaces are started. The native language and the source language specified by the user are each of the documents contained in the component library. Divide the screen into upper, middle, lower or front, middle and back windows; the middle window is used to display the current operation sentence and related information in the operation, and the lower or rear window is used to display the source language to be read, on or before The widget window is used to display the native language text that you have read. In addition, at the prompt line, a command button such as a repentance operation, a save exit, and the like, and a shift word sequence button are displayed; or they are made as a float bar immediately below the middle or the user can move. Then, run, the source statement reads in 901, reads a sentence from the source language as the current , , ,

content. The small idiom 902 is judged, and the source idiom field of the idiom library is queried with the current sentence. If yes, give the small idiom 903, take out the native idioms in the native idiom field of the same record, display it in the middle window, and read the idiom code of the same record into the essay buffer, and then execute the step source. The statement reads in 901. If not, it is not found, and the sentence-type subroutine 904 is continuously called. The source sentence field of the current sentence query query library, if there is no matching sentence pattern, the stored sentence library represents the number in the feedback buffer. If multiple matching sentence patterns are found, the corresponding parent sentence type is displayed in the lower part of the middle window, and the user is selected to select a sentence pattern. Or only find a matching sentence pattern, give the sentence pattern of the same record, the parent sentence type and the source sentence type, highlight the parent sentence type in the upper part of the middle window, and nest the source sentence pair into the source sentence type. , the note is displayed below the native language of the window, and the sentence code is read into the world cache. Then, the step 905 step is executed, and the current sentence box in the parent sentence type is marked from left to right in the middle window, and the current sentence box number is stored, that is, (the original sentence box number + FFE0H in the sentence pattern) is in the world text buffer area. At the same time, the corresponding sentence cabin content of the source language is marked and taken out as the current sentence cabin content, and it is judged whether the current sentence cabin content belongs to the simple sentence cabin 906. If it is a simple sentence cabin, perform step word determination 909. If not, perform a step to match the cabin mold 907. Query the source language mode field of the cabin model library with the current sentence content. If there is no matching cabin model, the library is represented by the feedback buffer. If multiple matching cabin models are found, the middle window is extended downward, the corresponding native cabin model is displayed in the extension, the user is selected to select one; or only one matching cabin module is found; the cabin code of the same record is given , native language module and source language module. The mother tongue module is highlighted in the extension of the middle window, and the source sentence cabin content is nested in the source language module, the note is displayed below the native module of the window, and the cabin model code is read into the world text buffer. . However, the subsequent take-up of the eye 908 step. From the left to the right of the native language model, the current cabin eye is marked one by one on the native language module; the current cabin eye mark (the original cabin eye mark + FFD0H on the cabin model) is stored in the world text buffer; For the cabin eye content, perform step word meaning determination 909. Read the simple sentence box of the source language or a word string in the cabin eye from left to right, and query the source language string field of the meaning group library. If there is no identical word string, save the current source word string in the feedback buffer. If multiple identical word strings are found, the contents of their same-recorded native language string fields are respectively taken out, backed up in the regret selection cache, and displayed in the lower part of the expanded middle window, and the receiving user selects one of them; or only one identical word is found. The string, the execution step gives a native string 909. The content of the native language string field of the current record is taken into the current parent sentence box or the current native language capsule, and the intention group code is stored in the world text buffer; the word meaning determination 909 is continued until the current simple sentence cabin or the current cabin operation is completed. Personality loss compensation 911 for the current sentence or cabin is based on the information of the personality loss table. Then correct the current sentence or 911 of the native sentence word order 911 according to the information of the native language word list. And query →, a move word order button. When the → button is clicked, the word string clicked by the current sentence or the cabin user is moved after the next word string; when a button is clicked, the word string clicked by the current sentence cabin or the cabin user is forwarded to the previous word string. Before; At the same time, add the word order after the move to the original word list for use. Subsequent judgment 912, if the current sentence compartment is still unprocessed, perform the step of taking the cabin eye 908. The subsequent judgment sentence ends 913. If no, and the current sentence has a sentence box that has not been processed, the execution step takes the sentence cabin 905. If all the sentences in the current sentence are processed, query the feedback buffer and the command button: when the feedback buffer is not empty, the information of the feedback buffer plus the source language, native language, current source sentence and other information is used as email feedback. Go to the support site; clear the feedback buffer and save the "feedback" flag in the world cache. Repentance command , ,

And make relevant modifications. When receiving the save exit command, save the contents of the world cache to World, file name = source file name .sjw; If the source text is not finished, the source offset is recorded in the file header. If the repentance operation and the save exit command button are not clicked, the execution step source statement is read into 901. The following steps are further illustrated by an example: Start preparing for some interfaces. The source statement reads 901, reads a sentence from the source language such as "Children not Allowed!" as the current sentence is displayed in the middle window, the native text display shows the content of the processed preceding sentence, and the source text displays the content of the current sentence. Judging the small idiom 902, querying the source idiom field of the idiom library with the current sentence. Yes, give a small idiom 903, take out the "Children are not allowed to enter!" in the native idiom field of the same record, this native idiom is displayed in the middle window, and the idiom code of the same record (for easy reading to the code name) And the low-word decimal number, the same as below; For example: "^ Read into the world buffer, and then execute the step source statement to read in 901. If the source statement read is "The doctor told his patient that he would prescribe him Some patent medicine on condition that he strictly follows his instructions.,, as the current sentence is displayed in the middle window, the native text display shows the last sentence of the processed sentence "Children are not allowed inside!", the source text display minus the current sentence content. The minor idiom 902 is used to query the source idiom field of the idiom library with the current sentence. None, not found, continue to call the sentence-type subroutine 904. The source sentence type field of the query sentence library is queried with the current sentence, then only Find a matching sentence pattern, give the sentence pattern of the same record, the parent statement type, and the source statement type, and tell the parent sentence type "{1} to his {2}, if Can {4}, it can be {3}." Emphasis is placed on the upper part of the middle window, and the source sentence is nested in the source statement "the {1} told his {2} that {3} on condition that {4 } .,,, Note is displayed under the mother tongue "the 1 {doctor} told his 2 {patient} that 3 {he would prescribe him some patent medicine} on condition that 4 {he strictly follow his instructions} . And the sentence code "sentence code 001061" is read into the world text buffer. Then execute the step 905 step, from left to right in the middle window to mark the current sentence in the parent sentence, and save the current sentence number in The world text buffer area. At the same time, the corresponding sentence cabin content of the source language is marked and taken out as the current sentence cabin content, and it is judged whether the current sentence cabin content belongs to the simple sentence cabin 906. If it is a simple sentence cabin, the step word meaning is determined 909. For example: Sentence 1, sentence 2; save sentence box number"

01", "sentence box number 02"; take out the corresponding sentence content of the source language "d _OC tor", "patient" judgment is a simple sentence cabin (a word string does not expand; there are three strings to expand again), the meaning of the word is determined The steps are "Doctor" and "Patient" respectively. The execution steps are given to the native language string 909. The content of the native language string field of the current record is taken into the current parent sentence compartment 1, the sentence cabin 2, becomes "1{Doctor} tells him 2{ Patient}, if it can {4}, it can be {3}. "take the sentence cabin, from left to right in the middle window to indicate the current sentence cabin in the parent sentence type, now should handle the sentence cabin 4. Continue to take the sentence cabin 905; take out the corresponding sentence content of the source language" he strictly follow his instructions: Simple sentence cabin 906, no, perform a step to match the cabin mold 907. Give the same recorded cabin code "condensation", the native language module "[1] + his +[2]" and the source language module "he+[l]+his+[2]". The source sentence cabin content is nested in the source language module as "he 1 [strictly follow] his 2 [instructions]". The mother tongue after the sentence 1-2 treatment is "the doctor told his patient that if he can 4{1 [definitely execute] + his +2 [doctor]}, then he can {3}." ,

Him some patent medicine ", judge the simple sentence cabin 906, no, perform the procedure to check the cabin model 907. Query the source language module field of the cabin model library with the current sentence content. Only find a matching cabin model "he would prescribe Him [ 1 ] "; gives the same recorded cabin model code "condensed i3⁄43⁄43⁄4WT, mother tongue module "open [1] to him" and source language module "he would prescribe him [1]". The mother tongue module is highlighted in the extension of the middle window, and the source sentence cabin content is nested in the source language module "he would prescribe him 1 [some patent medicine], and the note is displayed below the native font of the window, and Read the cabin model code "" into the essay buffer. However, follow the steps of 902. From the left to the right, the native cabin is marked one by one on the native cabin model; The world text buffer; at the same time labeling and extracting the corresponding cabin content of the source language, "some patent medicine execution step word meaning determination 909. Read a word string in the cabin eye of the source language from left to right, query the source language of the meaning group library String field. If multiple identical word strings are found, the contents of their same-recorded native language string fields are taken out, backed up in the regret selection cache, and displayed in the lower part of the extended middle window, and the receiving user selects one of them. For example: some several - ifcb quite a few patent patent license effects understand the medical internal medicine user selected as "some special effects". 909. Take out the contents of the current record of the native language string and fill in the current native language capsule to "open 1 [some special effects] to him", take out the group code "meaning group code 008264", "Italian group code 017655", "Italian Group code 005484" is stored in the world text buffer; the current cabin operation is completed. (There are "loss compensation, native language order 911" because it is not involved, to be added below.) The mother sentence is "The doctor told his patient, if If you actually execute his doctor's advice, you can give him {{1} some special effects}. "There is also a cabin eye 912, no. Continued judgment sentence ends 913, yes. At this point, all the sentence boxes of the current sentence are all processed, query feedback buffer and command button: When the feedback buffer is not empty, the feedback buffer will be The information plus the source language, the native language, the current source sentence and other information is made into an email feedback to the support website; the feedback buffer is cleared, and the "feedback sentence" flag is stored in the world text buffer. When the regret operation command button is clicked, according to The user clicks on the confession string, and extracts the corresponding content in the confession selection cache to allow the user to re-select the word string and make related modifications. This is clear enough and complete, and can be implemented by those skilled in the art. When selecting one of a plurality of meanings, the class, the directory, and the plurality of terms belonging to the time are all included in a "repentance operation table", and the repentance operation is in the parent sentence clicked by the user, according to the sentence cabin in which the user is located. , the cabin eye and the word string give the contents of the same batch table, let the user re-elect and modify accordingly. The regret operation table has a regret batch, a sentence cabin number, a cabin number, a string of words, Fields such as group code. The cabin number and the eye number are the same, and the source string is different. The content of the world text buffer, as in the case of the above example, is "^"00064; Sentence pattern 001061; sentence cabin No. 1; Yiqun code 002131; sentence cabin No. 2; Yiqun code 006386; sentence cabin No. 4; cabin model code 00207; cabin eye No.1; Yiqun code 016841; 017951; cabin eye 2; Yiqun code 019882; Save the file as World, file name = source file name. SJW; If the source text is not finished, write down the source text offset in the file header. However, the entire process step is terminated. If the repentance operation and the save exit command button are not clicked, the execution step source statement is read into 901. Supplement (the above is "loss compensation, native language order 911", etc., here): According to the information of the personality loss table, the current sentence or cabin eye loss compensation 911. The reason for the loss of personality loss is that the above-mentioned simple sentence cabin and cabin eye are the upper and lower concept, but the size is the same, except that they are less than or equal to three meaning clusters except the word. Not imaginary words such as Chinese quantifiers, English articles, etc. These lost words are compensated when read in their native language. The information is derived from the personality loss table, and the personality loss table contains fields such as associated strings and compensation strings. The native language word order sometimes needs to be adjusted. The reason is also from the sentence cabin and the cabin eye. They are the same size, and they are less than or equal to three Italian clusters except for the unspoken words. There is no requirement for the order of the three clusters in the machine, so the various applications based on the components of this statement. There may be cases where the native language word order sometimes needs to be adjusted. The adjustment is simple and convenient. First, a native language word list can be used. The native language word list contains the first word string, the read string, and the adjustment string field. The read string is the serial sequence when read in the native language; the first string is the first string of the read string; the adjusted string is the word order that should be adjusted. After the current sentence box or the current cabin eye is determined by all the words, the table is checked and the match is automatically adjusted. Then, let the system interpret the "one button" and "one button" buttons. If clicked by the user, adjust it according to the user's intention. When the "→ button" is clicked, the word string clicked by the current sentence or the cabin user is moved after the next word string; when the "one button" is clicked, the word string clicked by the current sentence cabin or the cabin user is forwarded to Before the previous word string; at the same time, the moved word order is added to the native language word list for use. One situation that does not require intervention is that the user's readers do not have to intervene when they ask for speed and do not mind. Therefore, the functions of loss compensation, native language order, etc. are made available to the user. About the support website: The above mentioned emails are sent to the support website. When the support website receives the feedback email from the user, after the expert processes it in real time, the new component is added to the corresponding component library, and the new component and related information is fed back to the user in real time, and the original "feedback sentence" flag is replaced by the user's participation. . This is one of the 1 user support; users can have any opinions, suggestions, etc. to communicate and support in this way. In addition, for the 2 version upgrade, social testing and social accumulation can be obtained. 3 Guide multilingual development together, the application of the present invention provides a platform for worldwide cross-lingual communication. It also ended the history of the natural evolution and development of natural language in their respective independent systems; the beginning of a multi-lingual common rapid development process. For example, it is necessary to correct, eliminate or add a string of meaning texts; promote new terms and the like can directly suggest and recommend to the author through the application of the present invention; and publicize the explanation to the reader. As mentioned above, in the process of reading a foreign language by the user in the mother tongue, the contents of the world text buffer are saved to generate the world text. As long as one foreign language article has been read by one person, thousands of people in the back can read the world, and reading the world is faster and easier than the native language to read the foreign language, without intervention, accurate semantics, read the user choice, world language Multilingual reading ,

Group code 002131; sentence cabin 2; Yiqun code 006386; sentence cabin 4; cabin model code 00207; cabin eye 1; Yiqun code 016841; Yiqun code 017951; cabin eye 2; Yiqun code 019882; Cabin No. 3; cabin model code 00206; cabin eye No. 1; Yiqun code 008260; Yiqun code 017655; Yiqun code 005484; ": If read in Chinese, perform steps: 1 Remove the code one by one; 2 Use the switch statement Classify the code separately; 3 if it is a sentence number, a cabin number, to indicate the current sentence or the current cabin; 4 to decompose the sentence code, idiom code, cabin model code, and group code into a library A record number, which gives the contents of a certain read field of a certain library or a record. If it is an intention group code, it is given to the sentence cabin or the cabin eye according to the instruction; 5 continues execution until the end of the text. For example:

^ - fetch - idiom library 64 recorded Chinese idiom field content "Children are not allowed inside!". Sentence pattern 1061 Chinese sentence pattern recorded:

"{1} tells his {2}, if {4}, it can be {3}.,, sentence 1 indicates that the Chinese string "doctor" recorded by the W-fetch-sense group library 2131 is filled in according to the reference. The sentence cabin 1 becomes: "1{Doctor} tells his {2}, if {4}, it can be {3}. ,, sentence cabin No. 2 - 亓, 卢^^W^ - take - meaning group string library 6386 recorded Chinese string "patient" according to the instructions filled in the sentence cabin 2 becomes: "Doctor told his 2{patient}, if Can {4}, it can be {3}." The Chinese cabin model recorded in the sentence cabin No. 4 H亓, cabin model code 207 "[1] + his + [2]" is filled in the sentence cabin 4 as follows:

"The doctor told his patient that if he could 4{[1] his [2]}, he could {3}.,, the cabin eye 1 indicates that the Chinese string of the 16841 record of the meaning group was "really". Code 017951- -^ 17951 The Chinese string "execution" recorded is filled in the cabin eye 1 as follows:

"The doctor told his patient that if he could [4] perform [2], he could {3}.,,,,,,,,,,,,,,,,,,,,,,,,,,,,, According to the instructions, fill in the cabin eye 2 , , sentence cabin No. 3 - 亓, 凝 - take - cabin model library 206 recorded Chinese cabin model "open [1] to him" According to the sentence filled into the sentence cabin 3 becomes:

"The doctor told his patient that if he can actually execute his doctor's advice, he can 3 [open [1] to him}." Cabin 1 indicates the meaning of the group code - the Chinese string "some" recorded by the group of letters 8260 ; Lu W 7^5- fetch - meaning group string library 17655 recorded Chinese string "special effect"; take - meaning group string library 5484 recorded Chinese string "medicine"; according to the reference into the cabin eye 1 becomes:

"The doctor told his patient that if he can actually execute his doctor's advice, he can give him a "some special medicine"." The native language reading method is one of the applications based on statement components. Referring to the decoding step in which the four sentence component libraries are used for the encoding process of the current sentence and the decoding process of the world sentence, a plurality of sentence component-based application systems can be generated: a method system based on sentence component-based world text generation. Used to convert traditional text into world, then multilingual readout. A text transformation method based on statement components. Used to convert a source text into a textual text, or to convert it into multiple languages. A machine translation method based on statement components. Used to translate a source language into a language, or to translate into multiple languages. The software system resulting from the implementation of the present invention can be implemented on existing medium, small, micro, supercomputers, notebook computers, PDAs, and the like, or on separate or connected computers. Implementations can be run on a variety of computer networks, particularly on the Internet. The implementation can also be run on devices such as "Personal Digital Assistant", PDA (Personal Digital Assistant). The product after the implementation of the invention can be applied to work, study, leisure, travel, etc., which need to communicate with people in other languages; it can be used in homes, institutions, schools, and various fields involving foreign languages.

Claims

 WO 2009/103208 Rights Requirements PCT/CN2008/072593

A statement component device comprising a CPU, an input, an output, and a related index table storing a response query, further comprising: a sentence component storage unit (101), the statement component storage unit including a statement component formed in the form of electronic data Statement component library: sentence pattern library (300), used to store sentence structure components, has a sentence pattern, a multi-text sentence type field, which contains at least one record, which has the same semantic meaning as different sentence patterns in the record; a cabin model library (400) for storing cabin modules, a cabin model code, a multi-lingual cabin model field, which contains at least one record, which has the same semantic meaning as a different language cabin model in the record; 500, 502), for storing the meaning group component, the intention group code, the multi-language string field, and the at least one record, having the same semantic meaning as the different language group strings in the record; the idiom library (600), For storing small idiom components, there are idiom codes, multi-language idiom fields, which contain at least one record, which has the same semantic meaning as different idioms in different records; Yitong Code Compilation Department (103), Sentence storage section member (101) is connected to each of the recording preparation to a statement of intended library member pass code.

2. The statement component device according to claim 1, wherein the statement component of the statement component library is a further component for assembling a language sentence or a standard component for encoding a sentence, and the statement component includes the following four types:

A sentence structure component (201, 301), which is used to form the basic structural framework of a sentence, represents the basic semantic category of the sentence, and also determines the order and number of sentence boxes contained in the sentence, and covers the sentence. Complex grammatical phenomena of sentence-like sentences;

2 cabin module (202, 401), which is used to form the basic structural framework of the complex sentence cabin, represents the basic semantics of the sentence cabin, and also determines the number and number of cabins contained in the sentence cabin. And the complex grammatical phenomena of this type of sentence cabin;

3 means group members (501, 503), which are members of the group of intentions, for filling the components of the simple sentence cabin (203~204) or the cabin eyes (205~207);

4 small idiom components (601), sentences that are too short to be used to separate sentence patterns and sentence cabins serve as small idiom components, which are used to directly form short sentences.

3. The statement component apparatus according to claim 1, wherein the multi-text component field of the statement component library is set in a language, and one text corresponds to one text component field; ^W P. ² 9/1?3208 Four clusters of meaning strings or idioms can be extracted from them

The two fields of the modulo, the string or the idiom constitute a corpus, a linguistic library, a first language library or a second language library for language translation and text conversion.

4. The statement component device according to claim 1, wherein the Italian code preparation unit adds the current library representative number to the current library record only when a new record occurs in any one of the four libraries. The number synthesizes the code, and fills in the code field of the new record of the current library, as the unified double-word fixed-length hexadecimal code of the statement component. The Italian code uniquely represents the language components in the current record of the current library. The same semantics.

5. A method for reading a foreign language in a native language based on a statement component and generating a world document, comprising the following steps:

54. Interface, the user specifies the native language and the source language are each of the documents contained in the component library;

55. The sentence is read in, a sentence is read into the source language as the current sentence is displayed in the middle window, the native language text is displayed at the end of the processed previous sentence, and the source text is displayed minus the current sentence content;

56. Using the four-statement component library to generate the Italian-language code by looking up the current sentence, giving the content of the native-language field of the same record, filling the relevant data into the feedback buffer area, the world text buffer area, and the regret selection cache;

57. Judgment processing, if the entire sentence of the current sentence is processed, query the feedback buffer and the command button: When the feedback buffer is not empty, add the feedback buffer information to the source language, native language, current source sentence, etc. E-mail feedback to the support website, clear the feedback buffer, and store the "feedback sentence" flag in the world text cache; when the repent command button is clicked, according to the user's click on the remorse string, the corresponding in the regret selection cache is taken out. The content allows the user to re-select the string and make related modifications; when receiving the save exit command, save the contents of the world text buffer as world text, if the source text is not finished, the source text offset is recorded in the file header; When the save exit command button is not clicked, step S5 is performed.

 6 . The statement component-based native language reading foreign language method according to claim 5, wherein the four sentence component library uses the four sentence component library to obtain an Italian code by looking up the current sentence, and simultaneously gives the same recorded native language. The step S6 of the field content further includes the following steps:

5601. Judging a small idiom, querying the source idiom field of the idiom library with the current sentence, if not, not finding, performing step S602, if yes, taking out the native idiom in the native idiom field of the same record, displaying In the middle window, and the recorded idiom code is read into the celestial buffer, and then step S5 is performed;

5602. Calling the sentence-type subroutine, querying the source sentence type field of the sentence pattern library with the current sentence, if a matching sentence pattern is found, performing step S603, if multiple matching sentence patterns are found, displaying corresponding in the lower part of the middle window Native language The sentence ^, ^ ^^ (^ is then executed in step S603, if a matching sentence pattern is also no τ (^3⁄4 ^ number in the feedback buffer;

5603. Give the sentence code of the same record, the parent sentence type and the source sentence type. The mother sentence type is mainly displayed in the upper part of the middle window, and the source sentence pair is nested in the source sentence type, and the note type is displayed in the native language of the window. Below, and read the sentence code into the world buffer;

5604. Take the sentence cabin, from left to right, in the middle window, mark the current sentence cabin in the parent sentence type, deposit it into the current sentence compartment bowing the 'world' buffer area, and mark and extract the corresponding sentence content of the source language as the current sentence. The cabin content, determining whether the current sentence cabin content belongs to a simple sentence cabin, if not, executing step S605, if step S608 is performed;

 5605. Check the matching mode, and query the source language mode field of the cabin model library with the current sentence content. If a matching cabin model is found, go to step S606. If multiple matching cabin modules are found, extend the middle window downward. Displaying the corresponding native module in the extension, accepting the user selection and then performing step S606, if there is no matching cabin model, the storage module represents the number of feedback buffer; giving the same recorded cabin code, The native language module and the source language module, the mother tongue model is highlighted in the extension of the middle window, and the source sentence cabin content is nested in the source language module, and the note is displayed below the native language module of the window, and The cabin model code is read into the world text buffer;

5607. The eyes are taken from the left to the right in the native mode, and the current cabin eye is marked one by one on the native language module. The current cabin eye is marked in the world space buffer, and the corresponding cabin contents are marked and extracted. Go to step S608;

 5608. Determine the meaning of the word, read the simple sentence box of the source language or a word string in the cabin eye from left to right, and query the source language string field of the meaning group library. If only one identical word string is found, step S609 is performed, if Find a plurality of identical word strings, respectively take out the contents of their same-recorded native language string fields, back them up in the regret selection cache, and display them in the lower part of the extended middle window. After receiving the user selection, perform step S609, if one and the same word There is no string, and the current source word string is stored in the feedback buffer area;

5609. Take out the content of the currently recorded native language string field and fill in the current parent sentence box or the current native language cabin, and take out the intention group code and store it in the world text buffer; continue to step S608 until the current simple sentence cabin or the current cabin operation is completed. ; according to the information of the personality loss table, the current sentence or cabin eye loss compensation operation; then correct the current sentence or the cabin's native word order according to the information of the native language word list; finally query the forward and backward shift word order button, the current enter button After being clicked to move the word string clicked by the current sentence or cabin user to the next word string, when the back button is clicked, the word string clicked by the current sentence cabin or the cabin user is moved forward before the previous word string, Adding the removed word order to the native language word list for backup, and subsequently performing step S610;

5610. Judging, if the current sentence compartment is still unprocessed, step S607 is performed. If no, the current sentence has a sentence compartment unprocessed, and step S604 is performed. If all the sentence compartments of the current sentence are all processed, the subsequent step S7 is performed.

9. ² ^/^ Requirement ₅ is a sentence-based native language read foreign language and f^^ 8ί> ⁵ , ⁹³ is characterized by: the interface as a native language to read foreign languages, can be dedicated to a screen or pop-up An active window or the interface is made as a float below the cursor or the user can move, let the user specify which language and source language are included in the component library or select a fixed version, can be a screen or a window Divided into upper, middle, lower or front, middle and back. The middle part is used to display the current sentence and related information in the operation. The lower part or the back part is used to display the source language to be read. The upper part or the front part is used to display the read. The native language text, in addition, in their lower or prompt line, the repentance operation, the save exit command button, and the forward and backward shift word order buttons; the email feedback to the support website, when the support website receives the feedback message from the user After being processed by the website in real time, the new component is added to the corresponding component library, and the new component and related information is fed back to the user in real time, and the original "feedback sentence" flag is replaced by the user's participation.

8. A method for reading a foreign language based on a sentence component according to claims 5 to 6, and generating a world document, characterized by the generated world text: a sentence pattern, an idiom code, a cabin model code, an ensemble code The component code, and the sentence box number and the cabin eye label are composed of elements; the world language reading language is selected by the user to select any one of the sentence component libraries; the world text reading process is to use the decoding of the four component libraries. The process, the steps are:

1 sequentially remove the code one by one;

2 use the switch statement to classify the code separately;

3 where if it is a sentence number and a cabin eye label, it is used to indicate the current sentence compartment or the current cabin eye;

4 Decompose the sentence pattern code, idiom code, cabin model code, and meaning group code into a certain record number of a certain library, and give the contents of a certain read field of a certain library and a record, wherein the meaning group code is given according to the instruction. To the current sentence compartment or the current cabin eye;

5 Continue to execute 1 until the end of the text.

9. A source and text text conversion method based on a statement component, comprising the following steps: s5. reading a sentence, reading a sentence of the source language as a current sentence; s6. using the four sentence component library to the current sentence Check the table matching, give the content of the target component field of the same record, fill in the relevant data into the feedback buffer, and repent the selection cache; s7. Judgment processing, when the entire sentence of the current sentence is processed, query the feedback buffer and the command button: When the feedback buffer is not empty, the feedback cache information plus the source language, the language type, the current sentence and the like are made into an email feedback to the support website, and the feedback buffer is cleared;

? ⁸ ^? ^ The content responds to the user click and makes relevant modifications; when receiving the save exit command, save the given target component as the target text. If the source file is not finished, the source offset is recorded in the file header; When the save exit command button is not clicked, the current sentence is determined, and step s5 is performed.

10. The method according to claim 9, wherein the converting operation step s6 further comprises the following steps: s601. Judging a small idiom, querying an idiom with a current sentence The source language idiom field of the library, if not, is not found, step s602 is performed, if yes, a small idiom in the idiom field of the same record is given, and then step s5 is performed;

S602. Calling the sentence-type subroutine, querying the source sentence type field of the sentence pattern library with the current sentence, if a matching sentence pattern is found, performing step s603, if multiple matching sentence patterns are found, responding to the user selection Step s603 is performed, if there is no matching sentence pattern, the stored sentence library represents the number in the feedback buffer area; s603. gives the same statement type of the statement, and inserts the current sentence pair into the source sentence type; s604. Cabin, from left to right, the current sentence cabin in the target sentence type is marked, and the corresponding sentence cabin content of the source language is taken out as the current sentence cabin content, and it is judged whether the current sentence cabin content belongs to a simple sentence cabin. If not, step s605 is performed, if Perform step s608 _;

S605. Check the cabin module, and query the source language module field of the cabin model library with the current sentence content. If a matching cabin model is found, go to step s606. If multiple matching cabin modules are found, respond to the user selection. Then, step s606 is performed, if a matching cabin model is not available, the storage model library represents the number of feedback buffers; s606. gives the same recorded language module and source language module, and sets the current sentence content to the ground cover. In the source language module; s607. take the cabin eye, from left to right on the target cabin model one by one to mark the current cabin eye, remove the source language corresponding cabin eye content, perform step s608 _;

S608. The meaning of the word is determined, reading a simple sentence box of the source language or a word string in the cabin eye from left to right, querying the source string field of the meaning group library, if only one identical word string is found, step s609 is performed, if Find a plurality of identical word strings, respectively take out the contents of their same record target string fields, back up the regret selection cache, and then perform step s609 after receiving the user selection. If there is no identical word string, save the current source string. In the feedback buffer area; s609. Take out the content of the source string field of the current record and fill in the current target statement cabin or the current target cabin eye, and continue to step s608 until the current simple sentence cabin or the current cabin operation is completed; The information of the table carries out the personality loss compensation operation of the current sentence cabin or cabin; and then corrects the word order of the current sentence cabin or cabin eye according to the information of the target word list; ^w , back shift word order button, the current enter button is clicked, the current sentence ^si^^卩! i ⁸ / ^{Q 2} , ⁵ ^ After the word string is moved after the next word string, when the back button is clicked, the current sentence box is clicked. Or the word string clicked by the cabin user advances before the previous word string, and the changed word order is added to the original word list for subsequent use, and the subsequent step s610 is performed; s610. Judgment, if the current sentence cabin is still unprocessed, the execution Step s607, if no, the current sentence has a sentence box unprocessed, and step s604 is performed. If all the sentence cases of the current sentence are all processed, the subsequent step s7.